Coordinating distributed work

Coordinating distributed work Exploring situated coordination with gaming-simulation

Coordinating distributed work Exploring situated coordination with gaming-simulation

PROEFSCHRIFT ter verkrijging van de graad van doctor aan de Technische Universiteit Delft, op gezag van de Rector Magnificus Prof. dr. ir. J.T. Fokkema, voorzitter van het college voor promoties, in het openbaar te verdedigen op donderdag 4 december 2003 om 15.30 uur door Joeri VAN LAERE ingenieur in de technische bedrijfskunde geboren te Terneuzen

Dit proefschrift is goedgekeurd door de promotor: Prof. dr. H.G. Sol

Samenstelling promotiecommissie: Rector Magnificus, Prof. dr. H.G. Sol, Prof. dr. ir. J.E. van Aken, Prof. mr. dr. J.A. de Bruijn, Prof. dr. L.I.A. de Caluwe, Prof. dr. J.L.A. Geurts, Prof. dr. ir. G.J. de Vreede, Prof. dr. C.M.J. van Woerkum,

voorzitter Technische Universiteit Delft, promotor Technische Universiteit Eindhoven Technische Universiteit Delft Vrije Universiteit Amsterdam Universiteit van Tilburg University of Nebraska at Omaha, USA Wageningen Universiteit

Luctor et emergo Ik worstel en kom boven

COLOFON Published and distributed by: Joeri van Laere van musschenbroekstraat 3 6533 NS Nijmegen The Netherlands Phone: +31 24 3510539 Mobile: +31 6 25248272 E-mail: [email protected]

Delft University of Technology Faculty of Technology, Policy and Management Jaffalaan 5, P.O. Box 5015, 2600 GA Delft The Netherlands Phone: +31 15 2782318 Fax: +31 15 2783429

English editor: Eric van Laere Cover photos: Eric-Jan Pater, Regiopolitie Amsterdam-Amstelland Cover design and printing: Universal Press, Veenendaal

Laere, J. van Coordinating distributed work Doctoral dissertation, Delft University of Technology ISBN: 90-9017526-1 Key words: Coordination, distributed work, distributed groups, virtual groups, gaming-simulation, gaming, computer supported cooperative work, business engineering, organization science.

© Copyright 2003 Joeri van Laere All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the author.

Preface and acknowledgements In more and more organizations nowadays employees are dispersed over multiple sites or they perform parts of their tasks at the customer’s, at home or on the road. For these organizations coordination of distributed work is of vital importance. However, many organizations struggle with integrating new organizational structures, new work practices and new information and communication technologies. In this dissertation we blend practical examples of coordination at the Amsterdam Police Force with a variety of theoretical perspectives from disciplines like organization science, information technology research, group dynamics and social psychology. This results in the development of a design approach that should guide organizations in improving the coordination of their distributed work processes. This research has been a wonderful opportunity to meet a lot of inspiring people. I am very grateful for their valuable contributions to my distributed work. In Delft I met Henk Sol. I thank him for his guidance and feedback in this research project, and for the freedom to choose the directions in the research that I preferred. A second tower of strength throughout my research was Gert-Jan de Vreede. Thanks for the numerous comments, the GDR experiences in Holland, India and South Africa, and the supportive social talks. Furthermore I would like to thank all my colleagues of the systems engineering group for the wonderful time. I’ll remember the “aio-weekends”! In Amsterdam Bernard Welten welcomed me. I would like to thank him and Mariette Christophe for the opportunity to conduct research in the Amsterdam Police Force for over 4 years. This experience has left behind an indelible expression. I also thank Paul Gademan for coaching and supervising my research from the police perspective during these four years. There are many more policemen to be thanked because they contributed to one of the case studies: the 20 regional coordinators, members of the regional project Youth in District 6, members of the regional project Fire Arms, all policemen who participated in the simulation games. A particular word of thanks I owe to Henk Bindt because he dragged me down to the bottom level of the Amsterdam Police organization to show me how work was really done there. I also thank Reinder Doelemen who taught me about intranets, and who arranged network facilities to be able to use the police intranet and e-mail in the

games. Furthermore I thank Cassandra Vrolijk, Patrick van der Blom and Jeroen Korf for dedicating their master thesis project to one of my case studies and I thank Stephan Schwagermann for the instant translation of my functional requirements into Access forms and buttons to be used in the game. Distributed over the Netherlands I thank the members of the Ph.D.-network on Knowledge and learning in organizations for the inspiring discussions once every two or three months. In Wageningen I found a new challenge at the chairs of communication management and communication and innovation studies. I thank my colleagues for the nice working atmosphere and the possibilities to finish my dissertation. In Nijmegen and Gouda I thank Miriam and Helga for watching over Lies and Jens to enable me to complete this research and this book. In Westdorpe my parents watch over my struggle to integrate being a father, being a husband, working, finishing a dissertation and having a social life. Thanks for always supporting me in my choices, and daddy, thanks for checking every inch of this book on incorrect English. At home every day ends with Ilse, Lies and Jens. Thanks for the highly necessary distraction, thanks for your support, thanks for your patience when tele-working daddy was behind the computer all day. It is over now!!

Joeri van Laere Nijmegen, October 2003

Contents 1 OPPORTUNITIES AND CHALLENGES FOR DISTRIBUTED WORK .......................................1 1.1 Distributed work....................................................................................................... 2 1.2 Coordination of organizational work........................................................................ 4 1.3 Groupware classifications ........................................................................................ 7 1.4 Benefits and opportunities in coordinating distributed work ................................... 8 1.5 Drawbacks and challenges in coordinating distributed work ................................... 9 1.6 Research issues....................................................................................................... 15 1.7 Research question and objective ............................................................................ 18

2 RESEARCH DESIGN ........................................................................................................23 2.1 Research philosophy and strategy .......................................................................... 23 2.3 Research instruments.............................................................................................. 26 2.4 Research site: The Amsterdam Police Force.......................................................... 29 2.5 Research outline ..................................................................................................... 31

3 THEORETICAL VIEWS ON COORDINATION ....................................................................35 3.1 Overview of selected theories and approaches....................................................... 35 3.2 Work arrangements ................................................................................................ 41 3.3 Dependencies ......................................................................................................... 45 3.4 Coordination mechanisms ...................................................................................... 47 3.5 Communication and decision making .................................................................... 51 3.6 Improving coordination.......................................................................................... 55 3.7 Conclusions ............................................................................................................ 60

COORDINATING DISTRIBUTED WORK 4 COORDINATION AT THE AMSTERDAM POLICE FORCE ................................................ 65 4.1 Case study set-up ....................................................................................................65 4.2 The work group Youth Annoyance.........................................................................66 4.2.1 Work process.................................................................................................... 67 4.2.2 Coordination mechanisms................................................................................ 70 4.2.3 Coordination problems .................................................................................... 71 4.2.4 ICT opportunities ............................................................................................. 74 4.2.5 Factors hampering utilization of ICT opportunities ........................................ 75 4.3 The Regional Project (Fire)Arms............................................................................76 4.3.1 Work process.................................................................................................... 77 4.3.2 Coordination mechanisms................................................................................ 79 4.3.3 Coordination problems .................................................................................... 80 4.3.4 ICT opportunities ............................................................................................. 81 4.3.5 Factors hampering utilization of ICT opportunities ........................................ 82 4.4 Reflection................................................................................................................83

5 AN INTEGRATED VIEW ON COORDINATION OF DISTRIBUTED WORK ........................... 95 5.1 Coordination choices ..............................................................................................95 5.2 Composition and competence .................................................................................99 5.3 Commitment and cohesiveness.............................................................................102 5.4 Contact and connection.........................................................................................105 5.5 Content and context ..............................................................................................107 5.6 The coordination framework as a whole ...............................................................109

6 A DESIGN APPROACH FOR COORDINATION OF DISTRIBUTED WORK ......................... 115 6.1 Way of thinking: Framework and guidelines........................................................115 6.2 Way of modeling: Task-actor models and simulation-games ...............................126 6.3 Way of working: Problem solving and appreciative inquiry ................................130 6.4 Way of controlling: Adaptive and incremental .....................................................133

CONTENTS 7 FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE ................................137 7.1 Action research set-up.......................................................................................... 137 7.2 Collaborative problem solving at the neighborhood teams .................................. 139 7.2.1 Work processes and coordination problems ..................................................139 7.2.2 Opportunities for coordination of distributed work .......................................141 7.2.3 Design of coordination for distributed work processes..................................142 7.2.4 Game design...................................................................................................146 7.2.5 Game results...................................................................................................152 7.3 Electronic interaction in the regional projects...................................................... 158 7.3.1 Work processes and coordination problems ..................................................158 7.3.2 Opportunities for distributed coordination ....................................................160 7.3.3 Design of coordination for distributed work processes..................................162 7.3.4 Game design...................................................................................................166 7.3.5 Game results...................................................................................................172

8 REFLECTION ................................................................................................................179 8.1 Reflection on the application of the design approach........................................... 179 8.2 Reflection on coordination in the Amsterdam Police Force................................. 185 8.3 Value of the design approach for service organizations....................................... 189

9 EPILOGUE ....................................................................................................................197 9.1 Research findings ................................................................................................. 197 9.2 Research approach................................................................................................ 204 9.3 Future research ..................................................................................................... 205

REFERENCES ...................................................................................................................209 SUMMARY .......................................................................................................................223 SAMENVATTING ..............................................................................................................233 CURRICULUM VITAE ......................................................................................................243

1 Opportunities and challenges for distributed work Organizing (distributed) work and coordinating the efforts of (distributed) organizational members has always been one of the key issues in organization science (Malone and Crowston, 1994; Mintzberg, 1983). Moreover it has been a dilemma in society since its origin. Aken (1994) refers to challenges of organizing and coordinating work in the Stone Age, in the ancient empires of Mesopotamia, China and Egypt, in the Roman Empire, in trade organizations like the Dutch V.O.C. and in armed forces through the ages. Due to the growing size of organizations and their continuing quest for resources or potential ‘markets and customers’ in far-away areas the organization members and the activities they executed became more and more dispersed. Consequently the coordination of their efforts (adjusting the subdivided tasks in such a way that they produce a meaningful whole) became more and more complicated. The increasing rate of development of information- and communication technologies (ICT) in the past 20 years has raised new questions and answers around this old issue. For instance: Is it easier to coordinate and collaborate with distributed colleagues when technologies like e-mail, intranet, internet, mobile phones and mobile computers are available? Is computer supported coordination productive, or even more productive than face-to-face coordination? Does organizational performance increase when more coordination is supported by ICT? What work can or should be conducted in a distributed setting? What tools or technologies are needed for what kind of work? Can everybody perform distributed work or are specific skills needed? The aim of this research is to shed more light on the heavily studied but still largely unexplored field of coordinating distributed work. At first sight there seem to be numerous opportunities to solve well-known coordination problems more easily. By adopting these new technologies coordination, collaboration and the accomplishment of work itself can become independent of time and place. This independence leads to more flexibility for individual group members in accomplishing their tasks and may thereby improve individual performance, work group performance as well as organizational performance. Malone and Crowston (1994) predict a shift to more coordination-intensive organizational structures, like for instance adhocracies (Mintzberg, 1979a; Toffler, 1970), network organizations (Van Alstyne, 1997) and (inter) organizational networks as a result of the decreasing costs of coordination. 1

COORDINATING DISTRIBUTED WORK In practice it proves to be quite difficult to utilize these ICT opportunities. Most organizations fail to fully utilize the opportunities for improving their coordination. They tend to stick to their usual ways of working and rely on traditional communication media. Mediated collaboration and coordination in distributed work settings is perceived to be more complex and less effective than coordinating or collaborating face-to-face (Abel, 1990; Finholt et al., 1990; Rice, 1994; Strauss and McGrath, 1994). Furthermore numerous examples of problems and failures when trying to change traditional ways of coordination have been reported and discussed in literature (Bardram, 1997b; Grudin, 1988; Grudin, 1994; Holmstrom, 1999; Howcroft and Mitev, 1999; Markus and Connolly, 1990; Orlikowski, 1992; Swan et al., 2000). Shulman (1996) even states that the major consequence of the introduction of new information technologies has not been better communication but faster misunderstandings. This introductory chapter elaborates on these different, sometimes opposite, arguments around coordination of distributed work. First the issues of ‘distributed work’, ‘coordination’ and some classifications of the ‘ICT that support it’ are described and defined. Subsequently the promising opportunities of distributed work are contrasted with coordination problems and implementation difficulties that may be experienced. From that discussion some research issues are identified that show the need for an integrated design approach to improve the coordination of distributed work. That need is finally translated into a research question and a research objective.

1.1 Distributed work Traditionally, members of work groups consisted of members of the same organization who often had their workplace at physically the same location. Although there have always been organizations with units dispersed over several sites, coordination and communication involved primarily adjusting activities with local colleagues. This is understandable given the fact that coordination crossing physical distances relied on traditional post or later telephone, which were slow or costly alternatives. In that situation a work group with members from several sites would not really have been an asset for an organization. Therefore the interdependencies between distributed units, groups and individuals were kept to a minimum. And if there were any interdependencies between distributed actors, they were kept as weak as possible, so they could be coordinated by standardization (Galbraith, 1973; Mintzberg, 1983). At present, distributed work groups and strong interdependencies between distributed individuals, groups, units and even organizations are increasingly common. Work group members often work on different locations and each member is contributing to more than one work group or project. Kinney and Panko (1996) show that over half the projects in their study has at least one member from another site, and many several, so project teams are already characterized by a good deal of distributed work.

2

CHAPTER 1: OPPORTUNITIES AND CHALLENGES FOR DISTRIBUTED WORK Factors that have contributed to this situation include, among others (Drucker, 1988; Haywood, 1998; Nijhuis, 1996): • the increasing speed, at significantly lower cost, of human and data transport as a result of the development of ICT; • the increasing size and the globalization of organizations; • the growing tendency of organizations to cooperate with each other; • the growing popularity of telecommuting and telework; • and the continuous need for improving organizational competitiveness. Considering what we exactly mean with ‘distributed work’ we observe a wide variety of definitions for an even so wide variety of closely related terms as for instance telework, distance work, dispersed work, telecommuting, telehomework, telecooperation et cetera. The most straightforward definitions emphasize the use of ICT, and the independence of time and space constraints, for instance: • telework is time- and location independent work supported by information- and communication technology (Benschop, 1998); • telework is a term which includes notions of independent, alternating, mobile and generally cooperative forms of work and denotes media supported, dispersed task accomplishment utilizing a multitude of expressive forms and types of workplaces, or more shortly: telework is media supported distributed work (Wigand et al., 1997). These definitions emphasize that the main difference between dispersed versus co-located work groups may be that the key ingredients in effective functioning ‘coordination, communication and decision making’ are partly or totally executed at a distance. For distributed groups who never meet face-to-face this may be true. However, as long as distributed groups have the opportunity to communicate face-to-face, the question remains to what extent they will choose for computer supported coordination. Also, colocated groups may take advantage of computer supported means of coordination, communication and decision making. A definition based on media use is not distinctive. Therefore we choose for the definition below: Definition 1: Distributed work Distributed work is work that is jointly executed by a group of interdependent actors while some of them are (part of the time or all the time) working at different sites From definition 1 it follows that the degree of distribution of the work can vary: • from one of the actors being on a different location, to all actors working apart on several different locations; • from small distances between the several locations, to globally dispersed groups; • from part of their work being executed while dispersed, to all of their work being done while on dispersed locations So even two colleagues working on a different floor in the same building (small distance), who should share experiences (weak interdependency), can be characterized as carrying out distributed work (with a low degree of distribution).

3

COORDINATING DISTRIBUTED WORK

1.2 Coordination of organizational work This section briefly introduces the concepts of ‘organizational work’ and ‘coordination’. The descriptions and definitions presented here serve as a starting point for discussion. A systems perspective on organizational work In this research organizational work is viewed from a systems perspective by adopting an input-process-output framework. According to such a systems perspective an organization can be seen as a purposeful or even purposive system: a structured set of interrelated actors and activities that perform certain organizational functions in order to fulfill demands of its environment (Leeuw, 1990; Sol, 1982; Vreede, 1995). Organizational processes can be seen as structured sets of subsequent transformations or activities. These activities are carried out by a set of related actors (organizational members), that have their own goals and their own perception of their working environment (Leeuw, 1990). When executing activities, actors use technology, information, and other resources to create value for internal or external customers. Finally, the activities are related in time and place, have a beginning and an end and have inputs and outputs (Alter, 1996). The following main elements of a work process can be distinguished: 1. input (raw material or information); 2. activities or tasks which have to be accomplished within the work process; 3. actors (people) who carry out these tasks; 4. supporting (information- and communication) technologies; 5. output (products as well as information); 6. control signals (form managers or control systems) to control the accomplishment of the tasks (like task assignments, priorities or work protocols); 7. organizational context in which the tasks are carried out. The relation between these elements is visualized in figure 1:

control (6)

input (1)

task execution (2)

actors (3)

Figure 1: Work as an organizational process

4

technologies (4)

organizational context (7)

output (5)

CHAPTER 1: OPPORTUNITIES AND CHALLENGES FOR DISTRIBUTED WORK These work processes can take place at various levels within the organization. Bots and Sol (1988) define a micro, meso and macro level. The micro level considers tasks carried out by individuals, focusing on individual performance. The meso level deals with processes within an organization and has to do with group as well as organizational performance. Finally, the macro level deals with processes carried out by different organizations. As we study internal organizational coordination the focus in this thesis is on the micro and meso level. We distinguish between work and coordination on an organizational, group and individual level. Specialization and coordination: the paradox of designing organizations Some definitions of coordination (Hengst-Bruggeling, 1999; Malone and Crowston, 1994; Vreede, 1995) are summed up in Table 1. Table 1: Various definitions of coordination Definitions of coordination 1. to bring (parts, movements etc.) into proper relation, cause to function together or in proper order (Oxford dictionary, 1982) 2. the act of working together harmoniously (American heritage dictionary, 1981) 3. the act of managing interdependencies between tasks carried out to achieve a goal (Malone and Crowston, 1990) 4. the way in which interdependent organizational activities of two ore more actors mutually influence each other in order to attain a certain objective (Vreede, 1995) 5. the integration and harmonious adjustment of individual work efforts toward the accomplishment of a larger goal (Ellis et al., 1991) 6. the collection of activities required to maintain consistency within a work product or to manage dependencies within the work flow (Curtis, 1989) 7. the timely and deliberate adjustment of decisions in different operational units with respect to certain variables pertaining to either of the characteristics time, place and physical object specification of the process of delivering a product or a service (Sheombar, 1995) 8. the communication of requests, promises and reports of the status of the commitment (Winograd and Flores, 1986) Although we all have an intuitive sense what the word coordination means, the diversity of these definitions illustrates the difficulty of defining coordination (Malone and Crowston, 1990). From this quick overview we can conclude that coordination is related to the division of work and task interdependency (3, 4, 5, 6), to decision making (7) as well as to communication (8). Various authors agree that the division of work and coordination are the two main principles of organizing (Aken, 1994; Bosman 1995, Eijck, 1996; Gulick and Urwick, 1937; Kastelein, 1990; Mintzberg, 1979a). First work has to be decomposed into tasks (specialization), and these tasks have to be allocated, as the resources of one man are 5

COORDINATING DISTRIBUTED WORK limited. Subsequently this division of work into tasks gives rise to the adjustment and integration of these tasks (coordination). At first sight, coordination follows after specialization. However, by decomposing tasks and the organizational workforce one creates dependencies between tasks, organizational units and actors. As Aken (1994) illustrates it is not that difficult to develop an allocation of tasks. The challenge for organizations is to decompose tasks in such a way that they can solve the created coordination problem easily and effectively. Kastelein (1990) describes this as the fundamental paradox of organizing. Specialization leads to a rise in effectiveness, efficiency and quality of work within each sub task. Coordination creates synergy by adjusting the sub-tasks and combining their outputs into useful and integrated products or services. However, both specialization and coordination also have negative consequences for the organization. Specialization leads to a loss of overview. The whole has become so large, differentiated and complex that it is impossible for one actor or unit to grasp every relevant detail and or all dependencies. Furthermore coordination increases heteronomy: one is more and more coordinated by others as one has to give up direction and control powers to other organizational members in order to let the organization as a whole function properly. Loss of overview and heteronomy together result in delineation: one does not recognize the work-, life- or world situation as one’s own. So the conflicting results of specialization and coordination lead to growing productivity on the one hand, but delineation on the other hand. One way to solve this dilemma is to return to easier forms of specialization and coordination, such as small, local, selfsupporting communities. However, this is not a realistic option as one does not want to abandon high level of productivity one is used to (Kastelein, 1990). The alternative is to reduce delineation, which is one of the important themes in research on the specialization - coordination paradox. Examples of contributions are: • creating awareness with the help of ICT: showing what other organizational members are doing currently, or what their contributions are to the larger whole helps to reduce the loss of overview (Steinfield et al., 1999); • defining self-managed work teams (Kuipers and Amelsvoort, 1992; Sitter, 1982; Trist, 1951) helps to reduce heteronomy. Summarizing the specialization and coordination of tasks are interrelated and have to be executed concurrently. Managing dependencies is only meaningful if one also addresses changing those dependencies in order to simplify the coordination problems. Therefore, in practice, people refer to task allocation as ‘coordinating’. In this research the following definitions are used to distinguish between organizational design, organizing work, decomposition, specialization, task allocation, coordination and other related terms. Definition 2: Organizational design Organizational design includes both the specialization and the coordination of organizational tasks. 6

CHAPTER 1: OPPORTUNITIES AND CHALLENGES FOR DISTRIBUTED WORK Definition 3: Specialization Specialization is the division of the ultimate organizational task into specialized sub-tasks and the allocation of these tasks among organizational actors. Definition 4: Coordination Coordination is the way in which dependencies between specialized and interdependent organizational tasks and actors are managed in order to attain a certain objective. Coordination can occur both with and without the support of ICT. We speak of media supported or computer supported coordination when a computer or other media are used for managing the dependencies between actors or tasks. Examples are media supported communication between actors for coordination purposes, as well as the use of MS Project instead of a planning board.

1.3 Groupware classifications The ICT that support coordination, communication and decision making are often called Groupware: “information technologies that mediate electronic interpersonal interaction” (Coleman, 1997). They can also be defined as “computer based systems that support groups of people engaged in a common task or goal and that provide an interface to a shared environment” (Ellis et al., 1991). Ellis et al. see the ‘common task’ and ‘shared environment’ as characteristics or dimensions. Thus a technology can be more or less ‘common task oriented’ and more or less ‘shared environment oriented’ which leads to a groupware spectrum of technologies. There are various ways to distinguish between groupware applications. McGrath and Hollingshead (1994) make a classification based on the function a group support system (GSS) has for the group (communication, information retrieval or task support), see Table 2 for examples. Table 2: Classification of GSS based on functionality Function Communication (Group Communication Support Systems) Information retrieval (Group Information Support Systems) Task support (Group Performance Support Systems)

Explanation Supports a meeting or individual communication Handle all information relevant to a group Designed for a specific task (may include communication or information function)

Examples E-mail, computer conferences Audio conferences Video conferences Shared drive on network Lotus Notes, Intranet, World Wide Web Co-authoring system Group scheduling tools Work-flow system Group decision support system

7

COORDINATING DISTRIBUTED WORK Another way to distinguish between available media and technologies is the framework of DeSanctis and Gallupe (1987). They take into account which time- and place constraints could be valid (Table 3). This framework is often referred to for illustrating the opportunities for flexibility that are being raised by ICT: work can be executed any time and any place. Table 3: Classification of media based on time and place constraints Same time Different time

Same place Conversation, meeting, flip-over, slide projector shared room, billboard, reading corner

Different place telephone, video conferencing, document sharing, chatting letter, fax, e-mail, voice mail, computer conferencing, shared database

One can also classify groupware with respect to the amount of senders versus receivers involved (one-to-one, one-to-many, many-to-many), or considering the organizational level the supporting systems aim at (organizational, group or individual). However, these distinctions depend on how the technologies are applied. For instance, e-mail or telephone can both be applied as a one-to-one and many-to-many medium, and computer conferencing can be a group or organizational level medium. The conclusion thus far is that there is a wide variety of technologies that can be used for coordination, communication and decision-making in distributed as well as face-to-face settings. We will now proceed discussing the main benefits of these technologies and the opportunities they create for improving coordination and organizational performance.

1.4 Benefits and opportunities in coordinating distributed work The obvious benefits of distributed work result from the growing independence of time and place. Examples are a rise of flexibility for both the employer and the employees. Employers mention a larger pool of possible job candidates, greater access to scarce experts, greater utilization of the workforce, more opportunity to work at the customers site and possibly a round the clock workforce. Employees mention increased independence, larger pool of jobs to choose from, and greater flexibility (Haywood, 1998). Furthermore advocates speak of cost savings (reduce of travel and office space), productivity rises, more employee satisfaction and environmental benefits due to reduction of traveling (Nijhuis, 1996). Others point at the opportunities to change and improve coordination in organizations. Malone and Crowston (1994) observe: “One obvious way of generating new coordination processes is by considering alternative ways of communication.” Furthermore Hiltz and Turoff (1998) state: “It is assumed that the best way to do something is the way it was originally done face-to-face. The real issue is: How do we use the opportunity of asynchronous communications to create a group process that is actually better than faceto-face group communications?” 8

CHAPTER 1: OPPORTUNITIES AND CHALLENGES FOR DISTRIBUTED WORK Vreede (1995) concludes that the developments of ICT have both been enabling as well as required for organizations to improve their coordination: • In the early years organizations have long focused on supporting the individual needs of their workers by offering stand alone support such as word processors, spreadsheets and database applications. Nowadays the use of ICT is increasingly shifting from supporting pure computational activities and information processing towards supporting coordination activities like mediation of interpersonal and group communication (Malone and Rockart, 1991). ICT thereby offers the ability to initiate the restructuring of organizational processes and the coordination of these processes. It may change the patterns of organizational communication and coordination, whether anticipated or not (Culnan and Markus, 1987; Flores et al., 1988; Lewin an Stephens, 1993). • On the other hand, new complex and organic organizational forms like matrix-, project- network-, and virtual organizations need ICT to coordinate their work processes productively (Fulk and DeSanctis, 1995). Organizations that adopt these structures have recognized that flexibility and adaptiveness are necessary conditions, that work is more and more becoming a group effort, and that well-coordinated interdependencies between their actors are crucial. To deal with these challenges, they require ICT that are capable of disseminating information faster, more cheaply and more selectively. These benefits and opportunities make ICT an important asset to complex and organic organizations. Utilizing these opportunities seems to have a profound impact on the effectiveness and efficiency of their performance. However, the next section shows that organizations may encounter severe difficulties in utilizing these opportunities.

1.5 Drawbacks and challenges in coordinating distributed work Contrary to the promising benefits and opportunities there are also a multitude of drawbacks and challenges to working in distributed settings. The major perceived backlogs and difficulties reported in a survey amongst 514 managers and some seminar sessions are (Haywood, 1998): • communication problems; • difficulties with team building and cultural issues; • difficulties with managing team process and work flow; • cost and complexity of supporting technology; • the difficulty of controlling employees at a distance; • threat of becoming isolated, lack of recognition; • the distraction on the various work locations, especially at home.

9

COORDINATING DISTRIBUTED WORK Furthermore research in the field of CSCW (Computer Supported Cooperative Work) has revealed possible causes why computer mediated communication may be less productive than face-to-face communication in particular situations: • A lack of media richness may result in a decrease of the quality of interaction. • Contacts may become less frequent. This lowers the amount of ‘group awareness’: the up to the minute knowledge of each other’s activities. • Contacts may become more formalized. This results in too much focus on efficient task completion and too little attention for group development and socialization. This may decrease trust and cohesiveness in the group and thereby influence performance negatively. • Especially coordination in the asynchronous mode asks for different skills and points of attention. Lack of experience may thus decrease performance. • Redesign, implementation and change are in itself problematic. The multitude of aspects and actors involved complicate these processes. These issues are discussed below. Media richness, social presence, communication richness and media choice Daft and Lengel (1984; 1986) developed a distinction between supporting communication technologies on the basis of ‘media richness’: the extent to which a medium is capable of sending rich information, i.e. providing several kinds of information like text, pictures, smell, noise etc. and providing immediate feedback. A closely related concept is ‘social presence’: the experienced psychological nearness of the person(s) one communicates with (Short et al., 1976). As one moves from the rich face-to-face situation, via two-way video, audio (telephone), to e-mail or written memos, the number and type of sensory channels available for information exchange are reduced and people are assumed to feel a greater social distance (Andriessen, 2002). The main predictions are that for coordinating more complex tasks, richer media need to be used, and that when this ‘media match’ between complexity and media-richness is violated, the performance (decision time, amount of consensus) decreases. McGrath and Hollingshead (1994) enriched this picture of media match theory by demonstrating that not only performing a complex task with a lean medium is ineffective (medium too constrained), but that also performing a simple task with a rich medium is inefficient (information too rich). In addition they pose that the fit between medium and task is not a one-to-one relation but should fall within a ‘wide band of good fit’. A marginal misfit can be compensated with more or less mental effort and adaptation processes. Although media match theory appears to be quite plausible and popular, the principle has also met criticism as empirical support is rather limited (Andriessen, 2002). Andriessen gives some reasons for this limited support. First, researchers do not always distinguish adequately between the objective channel capacity and the subjective social presence. Social presence may vary substantially between users and may depend on the experience the user has with the medium (Fulk and Ryu, 1990). Secondly, media choice may be influenced by other factors like accessibility and freedom of choice (Culnan, 1985; Rice and Shook, 1988), geographical distance (Trevino et al., 1987; Webster, 1998), urgency (Steinfiled and Fulk, 1986), symbolic value of the medium (Trevino et al., 1987) and

10

CHAPTER 1: OPPORTUNITIES AND CHALLENGES FOR DISTRIBUTED WORK social influence (Fulk et al., 1987; Fulk, 1993; Salancik and Pfeffer, 1978). This implies that people may choose a tool that is too lean or too rich when these other factors provide good reasons for it. Thirdly, simple tasks are often not as simple as they seem to be, but consist of sub-tasks, each with its own media requirements. Menneke et al. (2000) describe how a group had to go through a complex stage of finding out what information partners had and to define an adequate strategy for solving the problem, before they could start a simple problem-solving task. Following these critiques some alternative theories have been posed. Ngwenyama and Lee (1997) introduce the concept of communication richness as opposed to media richness. They view the communication process from a critical social theory perspective. Besides the technology characteristics the communication richness perspective also includes the organizational context (norms, power, authority and status of actors), the content of what people said in their interactions and the subjective meaning it has for the actors involved. From this study it follows that not the medium in itself, but the way in which actors use the medium is the most relevant predictor for their performance. Kydd and Ferry (1992; 1995) also conclude that there are subjective factors involved in determining the richness of a medium in addition to the objective technology characteristics of media richness theory. They developed an instrument that can capture these subjective factors in order to explain why for instance e-mail is perceived to be richer than dictated by the definition of media richness. Dennis et al. (1998) describe media synchronicity theory. According to this theory all tasks consist of two fundamental communication processes: conveyance of information and convergence of opinions. Media synchronicity is ‘the extent to which a communication environment encourages individuals to work together on the same activity, with the same information at the same time’. They argue that conveyance processes require low synchronicity (few feedback activities, strong concurrency, intensive fast and multiple information exchange). In contrast convergence processes require high synchronicity (direct feedback, low concurrency, discussion focuses on single issues). In conclusion, media choice proves to be an important concept to prevent situations where the interaction quality is too low. However, a rich medium does not guarantee interaction quality and with a poor medium good interaction quality can be reached under certain conditions. Knowing and understanding these surrounding conditions is thus a requirement for successful media choice. Group awareness and informal communication The effort of researchers trying to work on the quality of communication and coordination has focused primarily on non-verbal cues, i.e. ‘making the medium richer’. However, opportunistic connections, informal interactions (Bandow, 1997) and tacit coordination (Nijhuis, 1996) could be more important aspects of for instance video conferencing. Although the quality (media richness) of video conferencing is perceived to be much lower than face-to face communication and is even suggested to be as low as audio only, video-conferencing might support another function. It could facilitate group awareness: ‘the up to the minute knowledge of other people’s activities that is required for an individual to coordinate and complete their part of a group task’. 11

COORDINATING DISTRIBUTED WORK In distributed settings group members and their status are frequently unknown to each other, which complicates coordination (Finholt et al., 1990). Rice (1994) concludes that informal communication is not likely to occur and therefore everything is more ambiguous, complex and uncertain and thus less effective. Providing for awareness is thus an important aspect of facilitating distributed work. Steinfield et al. (1999) describe an awareness mechanism as a mechanism focusing on gathering and delivering awareness data and possibly suggesting actions on particular conditions that are sensed. They discern activity awareness, availability awareness, process awareness, perspective awareness and environmental awareness as different types of awareness data. Furthermore they describe how delivery can take place either passive or active, differentiated or undifferentiated, customized or fixed, with a focal or peripheral focus, within a single or across different applications and accessible anywhere or on a particular place. Finally gathering can be organized explicit or embedded. Awareness features can ease coordination for groups that have an interest to work together. Clearly there are lots of ways to provide awareness. Still, the group or the individual user need to configure the awareness data that support their performance best, in order to prevent information overload. Group development and socialization Chidambaran and Bostrom (1997a) distinguish between sequential and non-sequential models of group development. A well-known example of a unitary model of group development is the model of Tuckman (Tuckman, 1965; Tuckman and Jensen, 1977), see Table 4. Unitary models assume that all groups go through a certain series of predefined stages. In contrast non-sequential models recognize the uniqueness of each group and consequently reject the idea of a single, predetermined series of stages. They focus on the underlying factors, like deadlines or externally imposed structures, that cause shifts in group development paths that can vary for different groups. Table 4: Development stages of groups Stage Forming Storming Norming Performing Adjourning

Main tasks Coming together, first orientation Struggling over the goal and task to do Developing common ideas and norms how to do the task Executing the tasks Evaluating and dismantling the team

Chidambaran and Bostrom (1997a) argue that sequential and non-sequantial models can be viewed as being complementary. The first show what is the development pattern of groups, while the latter show how they develop. Both agree on the existence of certain critical periods, temporal milestones, in a group’s developmental path. Clearly, first meetings, crisis points and final meetings require dedicated support.

12

CHAPTER 1: OPPORTUNITIES AND CHALLENGES FOR DISTRIBUTED WORK Working together for a long time does not mean that a group is well-developed (Chidambaran and Bostrom, 1997a; Weick and Roberts, 1993). A well-developed group is characterized by: being cohesive, managing conflict effectively, balancing socioemotional and task needs, communicating effectively and being involved actively in group activities (Chidambaran and Bostrom, 1997a). Improved group processes are a necessary but not sufficient condition for improved group performance. Trust and cohesiveness are believed to be crucial for effective functioning of a group or team (Andriessen, 2002; Druskat and Wolff, 2001). Without proper trust building the group cannot grow from the second and third stage into the later stages. Lack of trust will thus have a negative effect on group performance (Lane and Bachman, 1998). Handy (1995) defines trust as confidence in someone’s competence and his or her commitment to a goal. Cohesiveness is the power to bind the group and relate members emotionally to each other and to a common task. On the other hand team members do not have to be best friends: mutual trust should be combined with respect of differences and productive controversy (Nahavandi and Aranda, 1994). Open and clear communication are essential to build and maintain trust, which also follows from some issues O’Brien (1995) has defined: “commitment and follow through, openness, team confidentiality, directness, honesty about abilities, honesty about knowledge and expertise, willingness to share, separate the person from the problem”. Both the possible decrease in quality in communication (media richness) and the possible decrease in frequency of communication (awareness) can hamper this need for open and clear communication and active participation. Furthermore studies have found that in computer mediated coordination there tends to be too much focus on efficient task completion and that group development and socialization are not emphasized. Some researchers conclude that ICT cannot completely substitute for face-to-face contacts (Abel 1990; Finholt et al., 1990). They argue that groups require face-to-face contact initially (forming, storming, norming) and future relationships (performing) can be supported by ICT (Haereni and Sigholt, 1995). However, Chidambaran and Bostrom (1997b) give examples how ICT can have a positive impact on group development. For instance, in GSS simultaneous information exchange supports task needs and active participation, while anonymity supports socio-emotional needs, open and clear communication, trust and cohesiveness. Unique characteristics of the asynchronous mode Research on asynchronous group support (Hiltz and Turoff, 1996) reveals the following: • participants change their behavior (tendency towards much longer entries); • unique coordination problems occur (participants that lag behind the group); • there is a need for facilities for meta-communication (communication about how and when what tools will be used); • and furthermore it is suggested that asynchronous groups outperform face-to-face groups in creativity and brainstorming activities, but on the other hand, have more problems in reaching a consensus without a structured decision process.

13

COORDINATING DISTRIBUTED WORK As noted earlier, asynchronous communication creates the opportunity to do things differently and may result in a work process that is better than face-to-face communications (Hiltz and Turoff, 1998). But the above findings show that asynchronous communication is quite different from synchronous interaction and thus requires specific skills. Examples are notification of absence and feedback issues (message received, message read, when will one take action upon it?). If people are not used to communicate asynchronously the benefits may not be utilized and communication may appear inefficient and disappointing. For example, Hooff (1997) demonstrates that users in larger organizations start using e-mail for simple information exchange. Only when they get accustomed to its use, more and richer tasks may also be performed with this medium. Redesign and implementation of computer supported coordination Grudin (1988; 1994) identifies some factors why groupware implementations often fail, for instance: lack of critical mass, the disparity between who does the work and who gets the benefit and the ignorance of social and political factors. Vreede (1995) observes that studies into the impact of ICT on organizations and organizational coordination have found equivocal and sometimes conflicting results. It is clear that ICT can bring improvements, but the real issue is how to design and implement it effectively. Recent case studies and theoretical contributions stress the importance of taking into account organizational context and social factors (Robertson et al., 2000): organizations should be hospitable for technology (Ciborra, 1996). Furthermore GSS needs to be flexible and responsive to the needs of the group (Fjermestadt and Hiltz, 2000), and GSS should be useful (Liao and Landry, 2000). However, good process design proves to be important as well: GSS can improve good processes, but can make badly designed processes worse (Qureshi and Vogel, 1999; Vreede et al., 2002). Organizational coordination and ICT have to be designed and implemented in an integral way. ICT will not impact an organization’s performance unless corresponding investments are made in the people, tasks and structure of the organization (Vreede, 1995; Morton, 1992; Leavitt, 1965). But given the involvement of technological, work and social aspects, supporting and improving distributed coordination and collaboration proves to be a complex issue (Qureshi and Vogel, 1999). Okamura et al. (1995) give an overview of some famous interventions that may help technology adoption: technology champions, training during introduction and ongoing training, facilitators, expert users, local gurus and experienced users. Cole (1995) identifies leader’s participation, degree of supporting execution of critical tasks, benefit for all group members (not only management), and consistency with group norms as important factors. Finally, Vreede (1995; 1997) discusses the importance of user’s participation during the design of organizational coordination mechanisms. Cole (1995) and Okamura et al. (1995) emphasize the importance of an ongoing learning process. The integral design methods should be applied in a more continuous manner: resulting in structural gradual changes instead of a radical change once in a while. Design methods should more and more become a part of the implementation and change process. Although these processes are already seen as cyclical instead of sequential, redesign is 14

CHAPTER 1: OPPORTUNITIES AND CHALLENGES FOR DISTRIBUTED WORK often still an instantaneous process. Adaptable new ICT can enable redesign approaches that get near to continuous redesign, continuous change and continuous learning. Okamura et al. (1995) point at the importance of ‘technology-use mediation’. This is defined as the deliberate, organizationally-sanctioned intervention within the context of use which helps to adapt a new (general purpose) communication technology to that context, modifies the context as appropriate to accommodate use of the technology, and facilitates the ongoing usefulness of the technology over time. CSCW mediators carry out this intervention. It is a continuous process that carries on after design and introduction of the technology in the organization.

1.6 Research issues The main reason for starting this research has been our amazement about the fact that the benefits and opportunities of ICT for the coordination of distributed work seem so obvious, but are so hard and problematic to realize in practice. Qureshi and Zigurs (2001) also wonder: “Why is reality so far from the ideal?” ICT brings benefits as well as new problems and the added results may be unclear. Researchers as well as practitioners struggle with integrating ICT, tasks, structures and human behavior in organizations. Four issues, which can be seen as critiques on contemporary research and understanding, are complicating these attempts: • lack of integration of the findings of various research disciplines; • limited conceptions of the fundamental notions of coordination and communication; • domination of experimental studies while group work proves to be socially, organizationally and temporarily embedded; • human aspects are marginally addressed in the majority of the design approaches. These issues are discussed below. Integration of understanding of various research disciplines Most IT design literature does not address the organizational context of ICT and much of the organizational design literature does not address ICT to support organizational coordination (Culnan and Markus, 1987; Holt, 1991; Huber and McDaniel, 1986; Orlikowski and Barley, 2001; Vreede, 1995). Orlikowski and Barley (2001) illustrate this with the example of telecommuting. IT literature focuses on technological trends and opportunities and is strongly materialistic and optimistic. Organizational Science (OS) literature focuses on institutional and cultural inertia and adopts a more pessimistic view. IT literature insists on interpreting current conditions against future possibilities, while the OS literature interprets the present in terms of the past. Orlikowski and Barley argue for the development of integrated understanding by focusing on an empirical description of what is actually going on in practice today. She suggests that the multi-disciplinary research community known as CSCW could serve as an example. It emerged out of human computer interaction research in de mid-1980s when computer scientists, software designers and social scientists were interested in developing technologies to support collaborative work (Greif, 1988). They aim at integrating 15

COORDINATING DISTRIBUTED WORK ethnography and systems design. Artman and Waern (1995) observe that bridging the gap between descriptive science and design science may be impossible, as designers want to change work-tools and work-context in something better, whereas studying current practice risks conserving nonfunctional procedures. They conclude that visions about better informing each other should never die and that the more approaches are applied the richer the possibilities for developing increasingly more acceptable and usable tools for people. The challenge is to combine the local and the general, the practical and theoretical. Observations of particulars in practice (IT, CSCW) are important for theory building, and theory (OS) is important for making sense of specifics. Interplay between them is crucial (Orlikowski and Barley, 2001). Conceptions of communication and coordination Both Shulman (1996) and Putnam et al. (1996) state that researchers have primarily treated communication as a tool, a conduit or a lens. These conceptions or metaphors limit the complexity and completeness of communication. Researchers have applied a technical means of knowing to human problems. This has led to false and exaggerated expectations. Putnam, Phillips and Chapman propose that discourse or symbol metaphors are more suited because they take into account meaning-generation, relational aspects, persuasive appeals and social and political context. Shulman argues that researchers should distinguish between the human and technological aspects of the infrastructure of communication by making a critical conceptual distinction between information and communication. Orlikowski and Barley (2001) also stress the importance of bridging the gap between the physical and the social in theories of technological change. Both Shulman and Orlikowski and Barley point at the promising efforts of social constructionists in addressing this issue. A similar line of reasoning can be followed for conceptions of coordination. Smagt (1997) argues that the information processing perspective (Galbraith, 1973; 1974) has too long dominated organizational science. This approach takes the execution of a coordination process for granted. It is assumed that interdependencies between work processes are clear and that the best coordination mechanism can be selected beforehand. The situated action perspective (Suchman, 1987; 1994) sees coordination mechanisms and decision rules as resources instead of descriptions of prescribed behavior. It asks for more interactional coordination because of the circumstantially contingent character of meaning and intention. The information that needs to be processed and the actors that need to be involved change because of situational circumstances. Orlikowski (2001) points at the importance of the research area of situated coordination for new theories of organizing. Concepts that emerged of the portraying of techno-social interaction are promising because they often refer to the entwining of human activity and technology. Case studies and action research can complement experimental research There is a growing agreement that work-group performance depends on social, organizational and temporal contexts (McGrath, 1990; 1991; Shulman, 1996). McGrath points out that groups are engaged in a messy array of projects, tasks and steps operating simultaneously. A group is playing several games at once, with different agendas for

16

CHAPTER 1: OPPORTUNITIES AND CHALLENGES FOR DISTRIBUTED WORK different audiences and different functions. This may explain the lack of consistency of findings across work group studies as different investigators may be focusing on different functions, at different levels, at different times (Shulman, 1996). Davison et al. (2000) and Fjermestadt and Hiltz (2000) state that experimental work has dominated and continues to dominate the discipline of GSS research. Much experimental research involves artificial problems being tackled by artificial decision-makers, typically students. Few have studied in situ applications of GSS in organizations. Clearly field studies, case studies or action research studies can elicit issues that are important to organizations when adopting and appropriating new technologies. For example, a lot of experimental studies in CSCW research focus on the support of one task or on one of the four dimensions in the different time and place matrix. Several authors (Nijhuis and Vreede, 1996; Sakamoto and Kuwana, 1993) have argued for more action research studies that deal with combining various synchronous, asynchronous, same place and different place media for the support of a group task. This variety of media would better match the variety of tasks a group normally performs and would take into account the impact of social, organizational and temporal contexts. Addressing the human aspects of distributed work There is a growing confidence that successful virtual collaboration or distributed work does not depend on the degree of technological sophistication. It is “how the tools are appropriated and used” that matters (Qureshi and Zigurs, 2001). However, research on virtual teams and knowledge sharing among dispersed professionals mainly addresses issues like technology design, media choice, goal formulation and alignment, task design and marginally focuses on skills and capabilities of the people involved. More insight from human resource management, the social sciences and from the area of group dynamics is needed to complete the picture on coordinating distributed work. The importance of human aspects is also reflected in the discussions on conceptions of coordination and communication. In design approaches for virtual and distributed work there is also marginal attention for the human side of change. Although the complexity and importance is recognized (Hills (1997): the people part is hard), it is often treated as a part of change management after design. Examples like participative design, training and coaching by facilitators or mediators all focus at adoption of the technology but do not necessarily incorporate human skills and capabilities as a part of the socio-technical system to be redesigned. CSCW mediators come closer in the kind of interventions they could do, but still their interventions only take place in the implementation phase. Vreede (1995), who used dynamic modelling and GSS as support for participative design, suggests that gamingsimulation techniques may be a valuable design instrument to let organizational members experience change alternatives within their future organizational context. Gamingsimulation could in that way provide socio-technical prototypes of possible future work situations. It could enable a rich discussion on human, technological and organizational issues in an early phase of the design process.

17

COORDINATING DISTRIBUTED WORK

1.7 Research question and objective The main conclusions from the preceding discussion are that: 1. Although groupware applications offer obvious opportunities to improve coordination of distributed work processes the full potential is seldom completely capitalized: ICT bring benefits as well as new problems and the added results may be unclear. 2. A major reason for this could be that rational views are overemphasized in current research and understanding. Research in the areas of organization science, information technology, social sciences, CSCW and group dynamics all discuss aspects of the phenomenon of coordinating distributed work, without developing an integrated understanding. We argue that there is a need for more integrative views and approaches to the problem of coordinating distributed work. 3. In developing such an integrative approach the conception of coordination itself and the impact of contextual and human aspects need particular attention. 4. In addition, gaming-simulation could be a useful design instrument to understand and communicate these various aspects of coordinating distributed work processes in one integrated perception. These observations are the starting point for our research and result in the following general research question to be answered in this study: GENERAL RESEARCH QUESTION

How can actors in large service organizations improve the coordination of their distributed work processes, given the opportunities and potential drawbacks of information- and communication technologies? This question reflects our belief that somehow distributed actors could benefit more from ICT opportunities than they do currently. The choice for large service organizations as a research object is grounded in the expectation that especially these organizations can take advantage of the flexibility opportunities offered by modern ICT. First it is easier to work with distributed actors on an information-product (service) than on a physical product. Secondly, large organizations have to deal with multi-dimensional coordination needs that inherently lead to complex dependencies between distributed actors in the organization. Moreover, their bureaucratic history may create barriers on the behavioral and contextual side of the change process. Consequently many kinds of human and contextual barriers could be observed in such organizations to make our picture as complete as possible. To answer this question we need to deal with issues like (Laere, 1998b; 1999b): • thoroughly conceptualizing coordination in distributed work settings; • understanding what factors influence coordination of distributed work; • formulating guidelines that indicate how coordination of distributed work can be improved; • and developing a support tool that can help organizational members improve the coordination of their distributed work processes. These issues are incorporated in four more detailed questions that have guided and focused our research activities. 18

CHAPTER 1: OPPORTUNITIES AND CHALLENGES FOR DISTRIBUTED WORK RESEARCH QUESTION 1 How can coordination in distributed settings be conceptualized in a way that takes into account organizational, technological, social and human aspects of distributed work? The current discussion on coordination of distributed work lacks a common definition and understanding of coordination that incorporates views from various research disciplines. Before addressing the question which way of coordination is better or worse in a certain situation we need to have an impression of the kind of coordination mechanisms that are applicable in a distributed setting. RESEARCH QUESTION 2 Why are ICT opportunities marginally utilized for improving coordination of distributed work, and what factors can promote and hamper ICT utilization? Given a sound understanding of coordination, a second step is to look for the impact of ICT on the effectiveness and efficiency of coordination of distributed work. Although some first benefits and drawbacks have been discussed in sections 1.4 and 1.5 we need to conduct a more extensive search for reasons and factors that promote or hamper the utilization of ICT opportunities for coordination or distributed work. Only if these opportunities and challenges are known and understood we are able to guide distributed workers in dealing with them. RESEARCH QUESTION 3 How can distributed groups and actors be guided in improving the coordination of their distributed work processes? We aim at developing some guidelines that either guide distributed actors in selecting appropriate coordination mechanisms given situational circumstances or show how they can influence organizational, technological, social and human aspects of distributed work. RESEARCH QUESTION 4 How can we design prototypes of distributed work processes that give integrated insight in the effects of (re)designed distributed work processes, supporting ICT and applied coordination skills within an organizational context before actual implementation of these new work processes? We observe that (lacking) coordination skills of distributed workers can have profound effects on the outcomes of distributed work processes. Furthermore we observe that most design methods poorly address the combined effects of organizational design, task design, information systems design and human skills development. Technology-use mediation (Okamura et al., 1995) approaches this kind of integration. However, that approach is only applied in the implementation phase and is still mainly technology centered. A support tool that can provide integrated insight might enable us to explore and improve possible future coordination scenarios already in the design phase. Gaming-simulation is suggested as a promising design instrument that could answer these needs (Vreede, 1995).

19

COORDINATING DISTRIBUTED WORK In answering these research questions we pursue the following research objective. RESEARCH OBJECTIVE

Formulate a design approach for improving the coordination of distributed work that incorporates an integrated view on distributed work processes, coordination mechanisms, supporting ICT and skills for distributed coordination within an organizational context The ultimate goal of this research is to develop a design approach that guides actors in large service organizations how to improve the coordination of their distributed work processes, based on an integrated view on coordination (Laere, 1998b). Such a design approach can be seen as an inquiry system as defined by Sol (1982, 1990): A structured set of instruments, embedded in a way of thinking, a way of working, a way of modelling and a way of controlling, which can be used as a context in problem-solving activities. It expresses a methodology in view of a problem area. Consequently our design approach should (Vreede, 1995): • make a worldview explicit with respect to improving coordination of distributed work; • outline the steps, or working method, necessary to analyze, diagnose and change distributed work situations in which coordination problems occur; • be capable of modeling distributed work processes and their coordination in current and future situations; • offer guidelines and techniques to manage and control the application of the modeling techniques according to the working method. Our design approach can also be seen as a field-tested and grounded technological rule (Aken, 2003). Aken uses the term “technological rule” to designate the knowledge one may use to design an intervention (or series of interventions or artefact) to produce in a given setting a certain desired outcome. The relevance of our research can be illustrated from a societal as well as a scientific point of view. Our approach can enable organizations to better understand and improve the coordination of their distributed work processes. Thereby organizations can better utilize the possibilities of the numerous technologies available and embed the use of general technologies in specific situations. Furthermore it can lead to more deliberate and more explicit coordination in organizations and to a better understanding of organizational coordination as a whole.

20

CHAPTER 1: OPPORTUNITIES AND CHALLENGES FOR DISTRIBUTED WORK

The scientific relevance consists of contributions to various research themes that are currently developing, namely: • the role of ICT in coordination intensive organizations and the discovery of new coordination mechanisms and processes in coordination theory (Malone and Crowston, 1994; Orlikowski and Barley, 2001); • the importance of the situated action perspective and human skills in redesigning organizational coordination (Suchman, 1987; Weick, 2001); • understanding the effectiveness of different forms of meta-structuring and thereby a better understanding of the adoption of groupware and organizational change (Orlikowski, 1995). In the next chapter we present a research approach for answering these research questions and fulfilling our research objective.

21

2 Research design In this chapter we respectively present our research philosophy, strategy and research instruments. Subsequently we introduce our research area, the Amsterdam Police Force, and discuss how it fits within our research design. The chapter is concluded by presenting the research outline and by clarifying the further structure of this dissertation.

2.1 Research philosophy and strategy The research philosophy and strategy define the basic principles of the way in which the research question is addressed: • A research philosophy determines what kind of knowledge can be obtained and what the limits to that knowledge are (Hengst-Bruggeling, 1999; Janssen, 2001; Vreede 1995). The philosophy guides the choice of research instruments for collecting and analyzing data on the phenomenon studied (Hengst-Bruggeling, 1999; Vreede 1995). • A research strategy consists of a rough overall plan for conducting a research project (Meel, 1994). The strategy outlines the sequence of data acquisition and analysis articulated in steps to be undertaken (Vreede, 1995). In general 4 different research philosophies can be discerned (Galliers, 1991; HengstBruggeling, 1999; Janssen, 2001; Meel, 1994; Vreede, 1995): • The ‘hard’ positivist research tradition: relying on the researcher’s objective observation of organizational phenomena, using “hard” research instruments such as laboratory experiments, field experiments and surveys and focusing on repeatability, experimental control and analytical rigor. • The ‘soft’ interpretive research tradition: relying on the researcher’s subjective interpretations and understanding of organizational phenomena, using research instruments such as action research, field studies, role playings and reviews, focusing on shared norms and practices and common languages between actors. • The ‘critical’ emancipatory research tradition: relying on the researcher’s normative interest in political organizational phenomena, using research instruments such as critical heuristics, stakeholder analysis, dialectical analysis and context analysis, focusing on coercion, power relationships, historical contexts, conflicts and ideal speech situations.

23

COORDINATING DISTRIBUTED WORK •

The ‘counterbalancing’ pluralistic research tradition: relying on the researcher’s awareness for possible biases in research approaches used when studying organizational phenomena, using multiple research instruments that are complementary and can provide valuable feedback for one another, focusing on combining different avenues to significant research outcomes for the same problem area under study.

Although ‘hard’ scientific research has long been the norm when studying organizational phenomena, a growing interest for soft, more interpretivist research can be found in both organizational and information systems literature. Also critical and pluralistic approaches are advocated (Meel, 1994). Furthermore Wicks and Freeman (1998) describe a pragmatic approach that stresses that research needs to have practical value and should focus on serving human purposes. Research strategies can roughly be categorized by making a distinction between theory building and theory testing (Galliers, 1991; Hengst-Bruggeling, 1999; Meel, 1994). Research strategies for theory building or exploration are often based on inductive reasoning and use qualitative research instruments for data collection (case studies, action research). Research strategies for theory testing are often based on deductive reasoning and use quantative research instruments (laboratory experiments, field experiments). The choice of a research strategy is based on the nature of the research problem (well, moderately or ill-structured) and the status of theory development (theory building, testing or expanding) in the research field (Vreede, 1995). We argue that our research problem is ill-structured (Janssen, 2001; Sol, 1982). The analysis in chapter 1 shows that there are many opposing perspectives on coordinating distributed work. Also there are many alternatives for improving the coordination of distributed work and the effectiveness and/or efficiency of the various courses of action cannot be evaluated using only numerical data. Furthermore we do not know of any coordination theories that integrate organizational design, information systems design and human skills development. Although numerous studies on IT adoption in organizations are known, few of them include organizational, technological, social and human factors. Besides, lots of them are descriptive and do not aim at designing and implementing new coordination mechanisms. On the other hand existing theories like for instance contingency approaches (Mintzberg, 1983) cannot be neglected. Given the expected impact of organizational context and human behavior on the issue of coordinating distributed work we argue that it is advisory to study the coordination of distributed work in its natural environment. Furthermore we argue that our research question, aiming at the integration of organizational design, information systems design and human skills development, cannot be answered in a purely deductive way. Consequently our research approach needs to follow an interpretive philosophy and inductive reasoning. 24

CHAPTER 2: RESEARCH DESIGN The inductive-hypothetical research strategy (Sol, 1982), based on Churchman’s Singerian inquiring system (Churchman, 1971), proves to be a suitable research strategy to support theory building and to study ill-structured problems (Hengst-Bruggeling, 1999; Janssen, 2001; Lohman, 1999; Meel, 1994; Sol, 1982; Vreede, 1995). The Singerian inquiring system aims at expanding scientific knowledge by adapting it endlessly, inductively and multidisciplinary based on new observations. The main benefits of the inquiring system that also hold for the inductive-hypothetical research strategy are (Sol, 1982; Vreede, 1995): • it stresses the inductive specification, testing and expanding of a theory; • it offers possibilities for interdisciplinary research; • it makes space for the generation of various solutions for a problem situation; • it emphasizes learning by considering analysis and synthesis as interdependent activities. The inductive-hypothetical research strategy consists of five steps (Churchman, 1971; Meel, 1994; Sol, 1982; Vreede, 1995), see also Figure 2: 1. Initiation: using a number of rudimentary theories some empirical situations are described; 2. Abstraction: the essential aspects are abstracted into a conceptual model; 3. Theory formulation: from the descriptive model a prescriptive conceptual model is derived in the form of a theory; 4. Implementation: in order to test the theory the model is implemented in one or more prescriptive empirical situations; 5. Evaluation: the results of the prescriptive empirical situations are evaluated.

1. Initiation

Descriptive empirical model

5. Evaluation

2. Abstraction

Descriptive conceptual model

Prescriptive empirical model

4. Implementation

3. Theory formulation

Prescriptive conceptual model

Figure 2: The inductive-hypothetical research strategy

25

COORDINATING DISTRIBUTED WORK The exact interpretation of the inductive-hypothetical strategy for our research is illustrated in Table 5 (page 32) where we combine the research questions, the research strategy and the research instruments into a research outline.

2.3 Research instruments Although the research strategy clearly describes the order and interdependence of the research steps, it does not offer guidelines how to carry out individual steps. The instruments of a given research concern the means with which data on the phenomenon studied is collected and subsequently analyzed. They define how the steps from the research strategy are carried out (Vreede, 1995). The choice for a set of research instruments depends on the research philosophy used, the current level of knowledge, the nature of the research problem, the aspects of the research problem the researcher wants to focus on, the control the researcher has over the actual behavioral events and the availability of resources (Benbasat et al., 1987; Hengst-Bruggeling, 1999; Vreede, 1995;Yin, 1989). In our research we use case studies as primary research instruments. They are conducted during the initiation and implementation phases of our inductive-hypothetical research strategy. The abstraction, theory formulating and evaluation phases are individual creative processes inspired by additional literature search and discussion of ideas with fellow researchers and case study participants. In the remainder of this section we elaborate on the value and design of case studies in our research approach. A case study is an empirical inquiry that investigates a contemporary phenomenon within its real-life context, when the boundaries between the phenomenon and its context are unclear, and in which multiple sources of evidence are used (Yin, 1989). Depending on the degree of participation and intervention of the researcher, three kinds of case studies can be discerned (Benbasat et al., 1987; Bockstael-Blok, 2001; Galliers, 1991; Vreede, 1995): • observatory research, where the researcher maintains a greater distance to the research subject; • participatory research, where the researcher interacts socially with the organizational subjects, but does not intervene; • action research, where the researcher intervenes in the phenomenon studied in order to apply a theory to practice and evaluate its worth. Case studies as research instruments fit our inductive-hypothetical research strategy very well (Sol 1982; Vreede, 1995).

26

CHAPTER 2: RESEARCH DESIGN We give three arguments to employ case study research (Benbasat et al., 1987; Vreede, 1995; Yin, 1989): 1. The phenomenon needs to be studied in its natural setting where the boundaries between the phenomenon and its context are unclear. We argue that given the importance of the organizational context and human factors it is necessary to study coordination of distributed work in its natural setting 2. The focus should be on process rather than outcomes; on how and why questions rather than what, who, where, how many and how much questions. For this study, the goal is to explore how distributed actors actually coordinate, how this can be improved, and why certain opportunities and drawbacks outweigh others. Hence, it is more appropriate to explore the process of coordination than to explore the resulting products. 3. There is a lack of previous studies, and we aim at elaborating theoretical understanding. Few previous studies have been done holding an integrated (organizational, social and technological) view on coordination of distributed work. Besides, lots of them are descriptive and do not aim at designing and implementing new coordination mechanisms. The major strengths of case study research is that it involves the most direct form of observation, that it captures reality in greater detail and that subjects forget that they are indeed the subject of research (Checkland, 1981; Galliers, 1991; Vreede, 1995). The main critique on case study research is that it offers little basis for scientific generalization (as one cannot generalize from a single case). Also researchers may have little control over experimental conditions and they may rely too much on subjective interpretations of collected data, which results in a lack of rigor. Furthermore case studies may need long periods of time in the field and result in massive unreadable documents (Miles, 1979; Yin, 1989). To overcome these critiques Yin (1989) advocates exercising great care in designing and doing case studies. Matters that should receive attention are (Meel, 1994; Yin, 1989): • design of the case studies (character, object, goal, single/multiple units of analysis and single/multiple cases); • site selection; • initial theoretical framework; • sources of evidence; • method of analysis. Below is described how we deal with these issues in our research. Design of the descriptive case studies For the descriptive phase of our research participatory research is preferred over observatory research. Participation of the researcher provides easy access to sources and a possibility to get a feel of the issues from inside, which is necessary to cope with the complexity of our research subject. For instance, in addition to observing coordination executed by organizational members we can also discuss with them why they make particular choices.

27

COORDINATING DISTRIBUTED WORK The object and goal of the case studies in the descriptive phase follow from our first two research questions. The object of study is coordination of distributed work processes in a large service organization. The goal is to increase understanding of what coordination means and to identify what the experienced opportunities and challenges of ICT in supporting coordination of distributed work processes are. We choose to conduct two participatory case studies in the same organizational context. One is focusing on an ad hoc work team, the other is focusing on a long term social network of professionals. In both studies we conduct observations on the micro (individual), meso (group or process) and macro (organization) level (Bots and Sol, 1988). Yin (1989) typifies this combination of multiple units of analysis and multiple cases as multiple embedded case studies. The main reason to study two different cases (two different distributed work processes) is to create a diverse and thereby complete overview of existing coordination mechanisms, ICT opportunities and challenges. The choice to limit the research in this phase to only two case studies in only one organization is grounded in the belief that an in-depth study of the issues around coordinating distributed work is more valuable than a broad study. In section 4.1 (page 65) the design of the descriptive case studies is explained in more detail. Design of the prescriptive case studies For the two case studies in the prescriptive implementation phase we choose for action research in order to obtain the dual outcomes of action (change) and research (understanding). The aim of these prescriptive action research studies is to show and evaluate the value of our design approach as a ‘proof by demonstration’ and thereby answering our third and fourth research question. Again we conduct two action research studies each aiming at the improvement of coordination in an existing distributed work process. For similar reasons as above the studied work processes are different in nature, but within one organizational context. These studies also focused at multiple units of analysis such as: general processes and single activities, teams or social networks and individuals and the different elements of the design approach. Consequently these are again multiple embedded case studies. In section 7.1 (page 137) the design of the prescriptive case studies is explained in more detail. Site selection We choose to conduct all 4 case studies in one organization because it is important to be long enough in an organization to be able to understand what is going on (Meel, 1994; Mintzberg, 1979b). This organization is the Amsterdam Police Force. In the subsequent section (2.4) we elaborate on the suitability of this organization for our research design.

28

CHAPTER 2: RESEARCH DESIGN Initial theory An inductive research strategy and the use of qualitative research instruments do not release the researcher from formulating a conceptual framework before starting a research project (Meel, 1994). Such an initial intellectual framework (Checkland, 1985) offers a rough outline of the theories with which we enter our descriptive case studies. Its main purpose is to facilitate and guide the data collection process, to confine the possible level of generalizability and to avoid ending up with story telling rather than theory building or theory testing (Vreede, 1995). Our initial theoretical framework consists of existing theoretical perspectives on coordination (see chapter 3). These different perspectives on coordination guide our observations in the two descriptive case studies. Sources of evidence During all case studies both quantitative and qualitative data from various sources are collected to enable a rich representation of the phenomena under investigation and to permit comparison and contrast of the collected data (multiple sources of evidence). Examples are semi-structured and open interviews, extensive study of organizational documents and minutes of meetings, attendance at meetings, observation of work methods, journals of communicative actions, questionnaires, evaluative group sessions and quantitative performance indicators in the games. The precise use of these sources in the various case studies is explained in sections 4.1 (page 65) and 7.1 (page 137). Method of analysis For each case study the interpretation and analysis of the data from our data sources is documented in a case study report (Laere, 1998a; 1999a; 2000; 2002a). To improve validity and to counterbalance the researcher’s subjective interpretations, all four case study reports are discussed with members from the participating organization and additional researchers are involved in the two prescriptive case studies (Blom and Korf, 2001; Vrolijk, 2000).

2.4 Research site: The Amsterdam Police Force In this section we introduce the Amsterdam Police Force, the case organization where we conduct all our case studies. First we explain the organizational situation in 1998 when we entered the police force and the general (coordination) problem they were facing at that time. Then we highlight those features that make this organization an appropriate site for conducting our research. The duty of the Amsterdam Police Force is to ensure safety, livability and societal integrity and to deliver a reliable service to the inhabitants of the city. In order to fulfill this complex overall task the Police Force has defined a tremendous amount of sub tasks to be executed by approximately 5500 organizational members. The core organization is hierarchically subdivided in geographical areas. This hierarchy consists of 3 levels: the region, 8 districts and 32 neighborhood teams.

29

COORDINATING DISTRIBUTED WORK Each neighborhood team consists of approximately 60 to 80 policemen. Currently a fourth level is introduced by making certain individual members of the neighborhood teams responsible for smaller areas in their neighborhood (community policing). Besides the core organization, some centralized staff departments (management support, human resources, facilities, ICT department, Criminal Investigations Department) exist. These staff departments are not included in the research analysis and are treated as external parties. This structure is a typical example of the divisional form. The major advantage of this organizational structure is that the neighborhood teams (and later on the area managers) are close to the citizens and can adjust their service to local problems. This market driven approach was the core of the decentralization process during the reorganization in 19921994. In the ideal situation, neighborhood teams should be able to function autonomously in the divisionalized organization. However, two major dependencies among the neighborhood teams complicate coordination within the police force: • Criminals tend not to restrict their activities to one geographic area. The same criminal may for instance rob people or burgle houses in different neighborhoods. So neighborhood teams need to exchange information on crime reports and coordinate their criminal investigations on the district or regional level. Also known effects of active prosecution or prevention in one neighborhood may create a shift of trouble to an adjacent neighborhood. • Another typical backlog is that the formerly centralized “product” expertise (for example car criminality, burglary, drug addiction and youth criminality) is spread over all the neighborhood teams. The underlying danger is that these teams operate too autonomously, communicate their experience too little and start to ‘reinvent the wheel’ over and over again. For reasons of manageability and limited research resources we restrict our research to the second problem: sharing of expertise and experience between neighborhood teams. To avoid problems like ‘reinventing the wheel’ professionals for each expertise were appointed in each neighborhood team shortly after the decentralization. These professionals have the task to communicate with their colleague experts from other neighborhood teams. They share their experiences with ‘the region’ and they collect relevant information and knowledge from the other colleague experts for their own neighborhood team. Sometimes they form regional work groups to tackle complex new problems that more neighborhood teams are facing at the same time. At the regional level a regional coordinator (integration manager) is assigned for each topic to make sure that the coordination between the local experts works out smoothly. These social networks of professionals on one expertise field are named ‘regional projects’. However, their actual behavior comes closer to the concept of communities of practice (Wenger, 1998) than to what generally is meant by projects.

30

CHAPTER 2: RESEARCH DESIGN A danger of the introduction of these specialized professionals is that their existence again stimulates the effect of ‘pigeon holing’, which was the most important reason to abolish their predecessors: the centralized functional departments. Pigeon holing means that the worldview of a professional is too much influenced by his own expertise and experience (Mintzberg, 1983). He starts to see aspects of his expertise in each problem he faces and he only solves that aspect of the problem, and not the whole problem, which leads to sub optimization. In contrast, the intention is that professionals from different topics collaboratively solve problems in a neighborhood team and thereby produce integral solutions. The organizational concept created is named the matrix organization (although it is not a matrix organization in its purest form). The matrix organization of the Amsterdam Police Force tries to integrate two extremes: • geographical specialization (because of a diversity in local market needs) • functional specialization (because of a growing complexity of tasks). The coordination and control burden in this kind of organizations is enormous (Mintzberg, 1983). Policemen have to cross the geographical borders and share their expertise, but at the same time have to have an integral vision and have to look over the walls of their ‘pigeon holes’. After about 5 years of developing this matrix organization results are both promising as well as disappointing. This is the moment when we enter the organization and start investigating current coordination problems and possible future ways of coordination. The results of the case studies conducted can be found in chapter 4 and 7. The following features make the Amsterdam Police Force an interesting organization to conduct our research: • the Amsterdam Police Force is a large service organization with a bureaucratic history that has to cope with multi-dimensional coordination needs as articulated in our research question; • they are currently facing problems in coordinating the efforts of distributed organizational members carrying out their work processes; • ICT opportunities are seen as potential avenues to solve these problems; • as the distances are quite small there are opportunities to carry out coordination both face-to-face and media supported, so we expect to encounter various ways of coordinating distributed work; • they are willing to cooperate in our research.

2.5 Research outline This section provides an overview of the main issues discussed in this chapter. Table 5 shows the interrelation between research questions, phases in our inductive-hypothetical research strategy, the research instruments and the chapters in this dissertation.

31

COORDINATING DISTRIBUTED WORK Table 5: Research outline Research Questions

Phases in cycle

Instruments

Ch.

Initial theory Initiation Abstraction

Literature study Participatory case studies Literature study

4.4

Theory formulation

Literature study

5

Theory formulation

Literature study

6

Implementation

Action research studies Literature study

7

DESCRIPTIVE PHASE

1. Conceptualizing coordination 2. Promoting/hampering factors 3. Guidelines for improvement

3 4

PRESCRIPTIVE PHASE

1. Conceptualizing coordination 2. Promoting/hampering factors 3. Guidelines for improvement 4. Gaming as design instrument 1. Conceptualizing coordination 2. Promoting/hampering factors 3. Guidelines for improvement 4. Gaming as design instrument

Evaluation

8

The descriptive phase of the research focuses at exploring and understanding the topic of improving coordination of distributed work in practice by looking for conceptions of coordination, factors that influence improvement of coordination and guidelines that help to improve coordination. We start with a literature exploration that results in a rough initial theoretical framework based on existing theories of coordination (chapter 3). This initial theory guides our observations in the participatory case studies (chapter 4). In the abstraction phase the case study observations and the results of some additional literature search result in some requirements for our integrated view on coordination and for our design approach (section 4.4). The second part of the research focuses on the development of a design approach for improving the coordination of distributed work. In chapter 5 we develop an integrated view on coordination of distributed work that incorporates the factors that promote and hamper improvement of coordination of distributed work. In chapter 6 the design approach is completed by describing the way of thinking, modeling, working and controlling with a particular emphasis on guidelines for improving coordination and on gaming-simulation as a support instrument. Next the approach is applied in two action research studies (chapter 7). Its value is evaluated by looking at the added value of the elements of the design approach, by comparing the action research outcomes with the original situation at the Amsterdam Police Force and by showing the value of our approach for other service organizations described in cases in literature. These evaluation results are presented in chapter 8. Chapter 9 concludes the dissertation with summarizing the main results and suggesting some directions for future research.

32

CHAPTER 2: RESEARCH DESIGN

33

3 Theoretical views on coordination The aim of the literature study on coordination presented in this chapter is to develop an initial understanding of what coordination is in order to be able to recognize and identify aspects of coordination when we study it in our participatory case studies. Section 3.1 introduces some approaches to coordination from the disciplines of organizational science, information technology, group dynamics and social psychology. In the remaining sections the various interpretations that researchers from these disciplines give to coordination and related subjects are described. This discussion is structured by first elaborating on the definitions given in section 1.2. We recall that organizational design encompasses specialization and coordination of the dependencies created by specialization. Section 3.2 discusses specialization in detail by looking at different kinds of work arrangements (ways of subdividing the ultimate organizational task). Section 3.3 presents different views on dependencies (relations between specialized tasks) and section 3.4 elaborates on coordination mechanisms (ways of coordinating these dependencies). In section 3.5 the related subjects of communication and decision making are discussed and in section 3.6 particular attention is paid to the question how coordination can be improved and how the degree of improvement can be determined and measured. Finally section 3.7 summarizes those elements from the various perspectives that we select to guide observations in our descriptive case studies.

3.1 Overview of selected theories and approaches Malone and Crowston (1994) summarize views and studies on coordination from economical, computational, biological, sociological and other perspectives (see Table 6).

35

COORDINATING DISTRIBUTED WORK Table 6: Examples of coordination theories (Malone and Crowston, 1994) Discipline

Theory / study object

Key coordination issues

Economics

Transaction cost theory Agency theory Team theory

Hierarchy versus market Incentives, limited observation Information exchange, interdependent decisions, same ultimate goals Incentives, conflicting goals Exclusive or concurrent access Task completion Segmenting and assigning tasks

Computer Science

Operations research Biology

Mechanism design theory Managing shared resources Managing unreliable actors Dividing parallel programs into units Managing information flows Scheduling and queuing policies and techniques Human physiology Ecology Animal group behavior

Linguistics, philosophy

Insect group behavior Speech acts theory

Share information with and without recognition Resource allocation Coordination of activities of different parts of a human body Coordination of activities of different plants and animals Interaction between simple rules and environment Simple rules for task allocation Requests and commitments

It is beyond the scope of this thesis to give a detailed overview of all the insights of these different approaches. We restrict our further discussion to a limited set of approaches from various research disciplines. These are: • The information processing paradigm (Galbraith, 1973), contingency theories (Mintzberg, 1983), managing task- and actor dependencies (Malone and Crowston, 1994; Crowston, 1994; Ven et al., 1976) and sociotechnical systems theory (Trist, 1951; Sitter, 1982, Kuipers and Amelsvoort, 1992) from the discipline of organization science. • Busines Process Redesign approaches (Drucker, 1988; Davenport and Short, 1990; Davenport, 1993; Hammer and Champy, 1993; Womack and Jones, 1994) from the discipline of IT research. • Theories and approaches on group performance and group design in face-to-face settings (Campion et al., 1993; Cummings, 1981; Goodman et al., 1987; Hackman, 1987; McGrath, 1984; Shaw, 1981) and computer supported settings (Pinsonneault and Kraemer, 1989; Nunamaker et al., 1991; McGrath and Hollingshead, 1993; 1994) from the discipline of group dynamics. • The theories of enactment and sensemaking (Weick, 1969; 2001) and situated action or situated coordination (Suchman, 1987; Schmidt and Bannon, 1992; Artman and Wearn, 1995) that have roots in the disciplines of cognitive psychology, human computer interaction, and social psychology.

36

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION The selection of these approaches is based on the observation made in section 1.6 that our design approach needs to pay attention to human and contextual aspects of coordination and to the impact of ICT opportunities. As is shown in Table 7, the selected theories and approaches differ in their attention for these topics. Some address IT issues, some address human issues. When addressing context some approaches refer to situational context, some refer to organizational context and some refer to environmental context of the organization. We will elaborate on that distinction in the introduction of the selected theories and approaches below. Table 7: Attention of theories and approaches for different aspects Theory / approach Information processing Contingency theories Socio technical systems Business process redesign Group dynamics (FTF) Group dynamics (CSCW) Sensemaking Situated action

IT aspects X

-

Human aspects X

X

-

-

X

X

X

-

X

X

X

Context aspects Organizational Environmental Environmental Organizational Organizational Situational Environmental Sitiuational

Organization science perspective Aken (1994) gives an overview of schools in organizational science (see Table 8). The classical schools (especially Webers bureaucracy) and the schools that focus on structure of organizations are the schools that explicitly address coordination. Well known is Mintzbergs profound synthesis of research on organization design in which he presents six organizational configurations and related coordination mechanisms (Mintzberg, 1979a; 1983). Furthermore the work of Galbraith (1973, 1993, 1994) provides insight in design and coordination of complex lateral organizations. He adopts an information processing view and focuses on concepts like uncertainty and horizontal communication. Malone and Crowston (1994) and Ven et al. (1976) elaborate on Thompson’s conception (Thompson, 1967) of dependencies between tasks and actors that need to be managed. Finally we choose to make use of the findings of sociotechnical systems theory (STS). This theory originates from field studies of the Tavistock Institute of Human Relations (Trist, 1951) and is a reaction to the unilateral emphasis placed on technical and human aspects in, respectively scientific management and human relations. Their field studies showed that the social and the technological systems in organizations are mutually influencing each other and need joint optimization. Considering coordination, especially Sitter (1982) focuses on structure and division of labor in his elaboration of the systemtheoretical basis of the sociotechnical design theory.

37

COORDINATING DISTRIBUTED WORK Table 8: Schools in organization science Classical Scientific management Management principles Bureaucracy People in organizations Human relations Revisionism Leadership of organizations Management process Decision making Strategic management Organizational change Structure of organizations Analysis and design of structure Structure and environment

Key authors Taylor, de Gilbreth’s, Emerson Fayol, Urwick and Gulick, Barnard Weber Key authors Mayo, Roethlisberger & Dickson Maslow, Mc Gregor, Blake & Mouton, Likert Key authors Drucker, Mintzberg, Kotter Simon, Kepner and Tregoe, March, Lindblom Ansoff, Andrews, Mintzberg, Quinn, Porter Lewin, Bennis, Tichy, Kanter Key authors Pugh c.s., Galbraith, Mintzberg Burns and Stalker, Lawrence and Lorsch, Miles and Snow Woodward, Perrow, Emery and Trist Chandler

Structure and technology Structure and strategy Functional schools Marketing management, operational management, management of innovations, administrative information science, knowledge management

Formal schools (Sociotechnical-) systems theory, cybernetics, management science / operational research Galbraith presents some universal guidelines and does not explicitly address human or contextual aspects of coordination. Mintzberg pays attention to context by defining certain contingencies (features of the organization itself and its environment) that determine how organizational structure and coordination should be organized. Both Mintzberg and Galbraith concentrate on coordination on the organizational level. Sociotechnical systems theory addresses human aspects by seeing the quality of workinglife as an important factor that can increase organizational performance. Furthermore it addresses the environmental context of the organization in a similar way as Mintzberg uses contingency relations. The early work (Trist, 1951; 1981) focuses on individual task design and on coordination on the group level. Later on design guidelines for the organizational level are developed (Amelsvoort, 1992; Leavitt, 1965; Sitter, 1982). Information technology perspective Business process re-engineering (BPR) or business engineering (BE) is organizational transformation focusing on the integral design of both IT and organizational processes and structures (Davenport and Short, 1990; Meel, 1994; Sol, 1992). It can be defined as: “the fundamental rethinking and radical redesign of business processes to achieve

38

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION dramatic improvements in critical contemporary measures of performance such as cost, quality, service and speed” (Hammer and Champy, 1993). Key principles of BPR are process modeling, radical change, cross-functional self-managed teams, empowerment of workers and IT as a process support (Munkvold, 2000). Although embraced as a radically new perspective, breaking away from existing paradigms, several authors have pointed out that BPR actually is a synthesis of existing paradigms, such as scientific management, the total quality movement and sociotechnical systems theory (Mumford, 1995; Munkvold, 2000). Although BE has received less theoretical and methodological support than sociotechnical systems theory (Meel, 1994; Mumford, 1995), BE is much more aware of the possibilities of using IT to enable and support new organizational structures (Meel, 1994). Meel (1994) argues that IT is poorly addressed in sociotechnical systems theory because ‘technical change’ often refers to changing work procedures rather than changing the supporting technologies. Although BE combines organizational, process and task engineering it clearly hinges on the process dimension (Hammer and Champy, 1993; Meel, 1994). Furthermore it combines the information processing paradigm, a flat and work flow oriented structure and cross functional, self-managed teams in its view on coordination (Meel, 1994; Munkvold, 2000). In comparison with STS we conclude that BE relies more on universal guidelines and pays less attention to contextual and human aspects. Group dynamics perspective The fields of social psychology, group psychology and organizational behavior provide theoretical and empirical evidence about the functioning of groups with respect to group design, group processes, group context and group performance (Hinssen, 1998). Compared to the approaches discussed above researchers in these fields pay more attention to ‘soft’ factors, like group cohesion, group culture, group norms and group attitudes (Hackman, 1987; Hinssen, 1998). Few researchers in the CSCW field (for example McGrath and Hollingshead, 1993; 1994; Nunamaker et al., 1991; Pinsonneault and Kramer, 1989) have used the findings and theories on small group research as a basis for studying the impact of ICT use on group functioning and performance (Hinssen, 1998). Furthermore Shulman (1996) blames both researchers on traditional and computer mediated groups for often ignoring the social, organizational and temporal contexts and multiple functions of these groups. For instance, the histories and expected futures that group members share and their contributions and costs to the embedding system are neglected when only one meeting is studied. Group dynamics mainly focus on coordination within groups, thereby offering possibilities to enrich our understanding of coordination on the meso or process / group level. Studies in the discipline of group dynamics pay more attention to human aspects of coordination, but on the other hand they seem to neglect organizational contexts. Contributions from the CSCW field address IT issues.

39

COORDINATING DISTRIBUTED WORK Social psychological perspective Weick (1969, 2001) formulates a theory called ‘organizational sense making’ that stresses the importance of collective mind in creating organizations and the subjective origins of ‘realities’. Organizations are much more a product of subjective thought and feeling, rather than an objective reality that can be observed differently (Banner and Gagne, 1995). Weick (1969, 2001) argues that we do not construct reality out of theory, but out of experience. A person has to wade into a chaotic set of environmental stimuli and somehow, through action, attempt to ‘unrandomize’ these data and impose order. This is called enactment. Together with enactment three other processes play a role in sensemaking. Selection is the retrospective interpretation of enacted cues. Retention is the preservation of plausible meanings of enactment articulated in cause effect assumptions, cognitive maps and organizational memory. And remembering is both believing and doubting what you know: guidance from the past needs to be combined with alertness to the non-routine in the present. Weick (2001) stresses that these four processes are heavily intermingled: doing produces knowing, but doing may also be informed by prior knowing. Sensemaking is about continuation, journeys rather than destinations. Weick’s conception of coordination includes notions as loose and tight coupling, coping with ambiguity and heedful interrelating. A class of theoretical approaches that can be typified as Situated Action (Artman and Wearn, 1995; Lave and Wenger, 1991; Resnick, 1991; Suchman, 1987) employs a view of reasoning that corresponds with Weick’s view on organizations. In addition they discuss the issue of coordination in more detail. Proponents of Situated Action argue that cognition becomes only manifest in acts (including speech acts) and that cognition is situated and can only be studied meaningfully when taking context and environment into account. Situated action takes interaction as the main unit of analysis. Focus is shifted from direct stimuli-response (behaviorist tradition) or cognitive architecture and schemata (information processing tradition) to reciprocal interaction with (physical) tools and social interaction in situ. Common features among the heterogeneous approaches within situated action are (Artman and Wearn, 1995): • Through interaction and communication we acquire and accomplish knowledge, meaning and intelligence; • knowledge and meaning develop within a situation and cannot be accounted for outside that situation; • acknowledging how people act and think is described by resources, which can be both material objects as well as external representational artifacts and immaterial artifacts such as language, symbols signs and theories. Both the social psychological and situated action view on organizational functioning emphasize the enactment of organizations and environments and the reciprocal relationship between structures of action and cognitive resources and constraints

40

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION embedded in a situated context. Both approaches are thus addressing human as well as contextual aspects. Weick’s approach focuses both on the organizational and group level, but does only marginally address the role of IT. Situated action pays more attention to the impacts of IT, but concentrates on the group- and individual level. In the remaining sections we show how these theories and approaches picture coordination and how they determine and guide improvement of coordination.

3.2 Work arrangements Depending on how an organization is subdivided and structured different kinds of work arrangements on the organizational, group, and individual level can be discerned. The organizational level Several authors in organizational science have described the various ways how components of the organization can be grouped. Combining the overviews of Gulick and Urwick (1937) and Mintzberg (1983) the categories shown in Table 9 can be discerned. Table 9: Categories for grouping or organizational units Product/market oriented categories Functional oriented categories Common purpose or contribution Common skills and knowledge Particular output Common work process or function Particular customers Particular geographic area Particular time (shift work) Mintzberg noted that these categories could be reduced to two main dimensions, namely a product/market oriented dimension focusing on external outputs and a functional oriented dimension focusing on internal production methods. Sometimes these dimensions can overlap (Mintzberg, 1983). In the case of a professional workforce (surgeon, ophthalmologist) their function or skills resemble the characteristics of their clientele. These general dimensions have resulted in a set of different basic organizational structures on the organizational level in the last century (Mintzberg, 1983; Galbraith, 1994; Banner and Gagne, 1995): • the simple structure (small firms); • the machine bureaucracy (highly specialized, routine operating tasks); • the professional bureaucracy (complex but standardized tasks); • the divisionalized form or sector structure (market diversity); • the matrix- and project organization (combining functional and output orientation); • the adhocracy (organic organizations, innovative work); • the missionairy form (ideological organization); • the network- and virtual organization (cooperation between multiple organizations).

41

COORDINATING DISTRIBUTED WORK It is important to note that these forms are theoretical extremes. One will hardly find any organization that resembles one of these forms exactly. However, often a dominant perspective can be discerned. Mintzberg (1983) and Schoemaker (1998) also identify hybrid organizations that are in between two or more extremes and use characteristics of each, or where parts of the organization resemble one form and others resemble a different one. The group or work process level On the group level these structures result in different groupings of tasks and actors. One can discern between units, processes and groups. Traditionally the main elements are units like functional departments, business units or geographical offices. They can be found in the hierarchical structures (all bureaucracies and the divisionalized or sector structures). In the more flexible and organic forms the corner stone of the organization are groups or teams. When groups (two or more people working together) establish common and shared goals they become teams (Nijhuis and Vreede, 1996). A team is a small number of people (5-25) with complementary skills who are committed to a common purpose, performance goals and approach for which they hold themselves mutually accountable (Katzenbach and Smith, 1993). We will elaborate on the concept of groups and teams after discussing business processes. Eijck (1996) argues that organizational design traditionally focuses on elements of organizational structure: “the establishment and interrelation of organizatonal units departments and the people that perform tasks have been more important than the actual design of tasks and processes themselves”. Literature on the topic of business reengineering (Hammer, 1990; Davenport and Short, 1990; Drucker, 1988; Hammer and Champy, 1993) shows a shift of focus towards the design of organizational processes as opposed to organizational structure. Galbraith (1973), once concentrating on flexible organizational forms like the matrix structure, also made a move towards a process view in recent work by discussing lateral organizational capability (Galbraith, 1993; 1994). A business process relates several individuals or groups who perform tasks that are interdependent (just like units and groups). Kuipers and Amelsvoort (1992) give examples of different kinds of business processes (see Table 10). Table 10: Types of business processes Process type job shop product line decoupled processes parallel processes continual processes ad hoc processes 42

Description job jumps from machine to machine or unit to unit in functional structure all products follow same schedule, assembly line, product oriented structure buffers, stocks in pure product line to cope with product variance group technology, different product lines for each product family process industry, automated manufacturing maximum product variance, custom made goods and services

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION Besides the growing interest for processes as opposed to structure the importance of teams is growing. Although teams can be found in traditional and modern structures, their role in modern structures as matrix, network and virtual organizations is larger. In traditional organizations the detailed labor division recorded in work instructions can be seen as the foundation of the organization. In modern organizations self-managed work teams that make decisions autonomously are the foundation (Schoemaker, 1998). In general one can distinguish between groups according to their degree of formality and permanence (Moorhead and Griffin, 1989): • command groups (formal and permanent); • task groups (formal and temporary); • friendship groups (informal and permanent); • interest groups (informal and temporarily). Dumaine (1994) discerns between five kinds of teams in organizations. • management teams (coordinate work of other teams); • work teams (perform daily work, sometimes ‘self managed: empowered to make decisions how the work gets done’, permanent character); • project teams (best people for a project, sometimes ‘virtual: regardless of location’, temporary duration form one or two weeks to several years); • quality circles (workers and supervisors who meet intermittently address workplace problems concerning quality and productivity, permanent, also: quality improvement teams, productivity teams, employee involvement teams, communities of interest); • special-purpose problem solving teams (members possessing knowledge and expertise related to a specific problem, temporarily: disbanding once the problem has been resolved). The individual level On the individual level the design issue is to define the degree of horizontal and vertical specialization (Mintzberg, 1983). Horizontal specialization refers to the predominant form of division of labor, the breadth or scope of a job. The question is how many different tasks are contained in each job and how broad or narrow each of these tasks is. Vertical specialization refers to the depth of a job, to the amount of control over the work. Vertical job specialization separates the performance of the work from the administration of it. After the discovery of the advantages of specialization by Adam Smith (1910) horizontal and vertical specialization have been exploited to extremes in scientific management (Taylor, 1947; Weber, 1947). In response to the resulting machinelike jobs where all possible brain work had been removed the human relations school (Herzberg, Mausner and Snyderman, 1959; Herzberg, 1968) introduced the terms job enlargement (horizontal job enlargement) and job enrichment (vertical job enlargement). They propagated a wider variety of tasks and more control over them in order to increase the quality of the working life and the motivation of the workers. Mintzberg (1983) concludes that the design of jobs is an art of balancing the amount of horizontal and vertical specialization: a job can be too large and too narrow. Furthermore some people prefer highly specialized work whereas others prefer enlarged jobs. 43

COORDINATING DISTRIBUTED WORK Organizational design and retrospective interpretation It is important to stress that job design on the individual level is specialization in its purest form. Assigning organizational members to units, groups or business processes is in fact a first step in coordination rather than specialization. It favors intra-group coordination over inter-group coordination. Mintzberg (1983) explains the relation between organizational design from an organizational, group and individual perspective as follows: “Given the overall organizational needs (goals to be achieved, mission to be accomplished, as well as a technical system to accomplish them) the designer delineates all the tasks that must be done. This is essentially a ‘top-down’ procedure, from general to specific tasks. He or she then combines these tasks into positions according to the degree of specialization desired, …, and builds the superstructure, first by determining what types and how many positions should be grouped into the first-order units, and then what types and how many units should be grouped into ever-more-comprehensive units, until the hierarchy is complete. This last step is of course a ‘bottom-up’ procedure, form specific tasks to the overall hierarchy. This is the procedure in principle. In practice, the organizational designer may take many shortcuts, reversing the top-down or bottom up procedures. For example, he typically starts with knowledge of specific structures and so can often move from missions to units directly. … Organizational design is much less common than organizational REdesign: incremental shifts from existing structures. In practice, as goals and missions change, structural redesign is initiated from top-down; as the technical system of the operating core changes, it proceeds from the bottom up” (Mintzberg, 1983, pages 45-46). This explanation shows again that task specialization and coordination are highly intertwined activities in practice. From Weick’s point of view (Weick, 2001) it is not the rational organizational design that guides action in the organization, but action is the source for post hoc rationalizing, sensemaking or retrospective interpretation. The work arrangements described before are examples of mental maps. They picture the complex organization in an understandable way in order to reduce ambiguity. However, Weick (1985a) states that these perceptions of reality are much too rational and exaggerate the orderliness in organizations. Hindsight bias (overestimation of the predictability of past events) fails to incorporate the experiments, false starts and corrections that enabled people to learn and improve in adaptive action. When failing to compensate for this perceptual bias, observers conclude that rational work arrangements work and that organizations are tight-coupled systems. Weick (1985a) argues that organizations should be seen as loosely coupled systems. Within those complex organizations small, stable segments exist (like short noninterruptible tasks, few stable social relations per person, and brief writings). Connections among those segments have variable strength, which makes that they are loosely coupled (see also section 3.3). Although this model of organization is more chaotic, it does not claim that order is completely absent. Weick (1985a) states: “Organizations may be anarchies, but they are organized anarchies; they may be loosely coupled, but they are loosely coupled systems”. 44

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION Although these views of Mintzberg and Weick seem quite polarized, contributions from the situated action perspective give clues how they can be combined. For instance Suchman (1987) argues that one should differentiate between work and representations of work. Work or action is essentially situated and consists of ad hoc improvisations. Plans and process models are then a rational anticipation before the act or a post hoc reconstruction afterwards. Plans do exist and may be valuable in guiding work but should be understood as an artifact of our reasoning about action, and not as the generative mechanisms for action (Suchman, 1987). We conclude this subsection with a definition of a work arrangement and some examples of types of work arrangements (see Table 11). Definition 5: Work arrangement A work arrangement is (a perception of) a particular division of tasks amongst actors. Table 11: Work arrangements on different levels Organizational level Organizational Group Individual

Work arrangements Structures Units, groups, business processes Job breadth and job depth

3.3 Dependencies Whatever work arrangement is chosen, it will always result in the creation of some dependencies that need to be coordinated. Mintzberg (1983) distinguishes between 4 kinds of interdependencies on the organizational level: • work flow interdependencies (product assembly, service integration); • process interdependencies (knowledge exchange among specialists); • scale interdependencies (economies of scale); • social interdependencies (need for social relationships and getting along). On the group level Thompson (1967) distinguishes between pooled, sequential, and reciprocal interdependencies. Pooled processes share a limited resource such as time, money, machine capacity or workforce capacity or each give a discrete contribution to the overall outcome. In the case of sequential interdependence each task delivers the input for the next in a pre-determined sequence. When interdependence is reciprocal, each sub-task influences the other, and therefore actors continuously have to discuss how to carry on with their activities. Ven et al. (1976) add a fourth dependency, namely team interdependency. Here tasks are worked on jointly and simultaneously (rather than being passed back and forth). Thompson argues that more complex interdependencies (complexity increases from pooled to sequential to reciprocal) ask for different ways of coordination, and thus other coordination mechanisms.

45

COORDINATING DISTRIBUTED WORK Malone and Crowston (1994) define four types of interdependencies, and state that there may be more: • shared resources (resembles pooled, input interdependency); • producer - consumer relationships (resembles sequential interdependency); • simultaneity constraints (resembles reciprocal interdependency); • task - sub task relations (resembles pooled, output interdependency). Weick (1985a) does not speak about dependencies, but describes connections that exist between segments in loosely coupled systems. These connections have variable strength. They are activated suddenly (rather than continuously), occasionally (rather than constantly), negligibly (rather than significantly), indirectly (rather than directly) and eventually (rather than immediately). Connections of variable strength produce ambiguity because variability makes it difficult to anticipate, plan, implement, coordinate and control. Variability makes it hard to be rational, but on the other hand it enables people to stumble onto adaptive actions they had not thought of. A loosely coupled system is a problem in causal interference: the prediction and activation of cause-effect relations is made much more difficult. However, it is not a flawed system. It is a social and cognitive solution to constant environmental change, the impossibility of knowing another mind, and to limited information processing capabilities. Because ambiguity cannot be fully removed the organization needs to cope with ambiguity as well as to reduce it to tolerable levels. Both dependencies in the rational view from organization science and connections in the social psychological view of sensemaking are examples of relations between certain elements. In the rational view these relations are taken for granted and more or less prescribe how coordination should be achieved. In the social psychological view these relations are variable and our perceptions of them are fallible. In fact coordination (managing the dependencies) influences the strength of the connection as action preludes sensemaking and limits the possible meanings of the situation under consideration. Whether we perceive the relation as a strong dependency is thus both a result of coordination and a starting point for coordination. We conclude this section with a definition of a dependency and some examples. Definition 6: Dependency A dependency is (a perception of) a relation between two or more tasks, units or actors that makes them unable to reach their goals autonomously Table 12: Dependencies on different levels Organizational level Organizational Group Individual

46

Dependencies Work flow, process, scale, social Input-pooled, output-pooled, sequential, reciprocal

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION

3.4 Coordination mechanisms Numerous coordination mechanisms have been identified to manage dependencies. They all aim at integrating the various tasks, units, processes and actors in order to achieve a certain objective. Table 13 gives an overview of a set of basic coordination mechanisms for the organizational level that are described in organization science. In this overview the value of the work of Mintzberg (1983) is prominent. He summarizes most of the earlier contributions on the topic of organizational coordination. In addition he provides a framework that relates the various kinds of basic coordination mechanisms by picturing them as a set of basic pulls on the organization. The strengths of the various pulls determine the actual structure of the organization (Mintzberg, 1983). As observed earlier (in sections 1.2 and 3.2) division of work is both a step preceding coordination in organizational design, as well as a first means to group tasks, units or actors and thus to coordinate. Table 13: Coordination mechanisms on the organizational level Mechanism Division of work (Simon, 1945) Direct supervision (Mintzberg, 1983) Standardization of work (Mintzberg, 1983) Standardization of skills (Mintzberg, 1983) Standardization of output (Mintzberg, 1983)

Standardization of norms (Mintzberg, 1983) Mutual adjustment (Mintzberg, 1983)

Related notions Creation of slack resources, creation of self contained tasks (Galbraith, 1973), unit grouping and unit size, power of decision making and decentralization (Mintzberg, 1983) Downward transmission of authority and influence (Simon, 1945), hierarchy (Galbraith, 1973) Establishment of standards (Simon, 1945), rules (Thompson, 1967), rules and procedures (Galbraith, 1973), behavior formalization (Mintzberg, 1983) Training and indoctrination of members of the organization (Simon, 1945), training and indoctrination (Mintzberg, 1983) Establishment of standards (Simon, 1945), standards (Thompson, 1967), plans (Thompson, 1967), programs (Galbraith, 1973), targeting or goal setting (Galbraith, 1973), investment in vertical information systems (Galbraith, 1973), planning and control systems (Mintzberg, 1983) Shared values (Mintzberg, 1983) Providing channels of communication (Simon, 1945), mutual adjustment (Thompson, 1967), creation of lateral relations (Galbraith, 1973), liaison devices (Mintzberg, 1983)

Coordination mechanisms on the group level have been described in the area of group dynamics. In the overview provided by Moorhead and Griffin (1989) the amount of similarities with coordination mechanisms on the organizational level is striking (see Table 14), especially when one considers that organization science and group dynamics have developed their view on coordination rather independently. The absence of 47

COORDINATING DISTRIBUTED WORK standardization of work and skills on the group level can be explained by the relatively short lifetime of groups, which makes the standardization investments only feasible on the organizational level. Furthermore this comparison shows clearly that researchers on the group level pay more attention to human aspects. In fact, standardization of norms does not really cover all the human issues raised on the group level. Table 14: Coordination mechanisms on organizational and group level Organizational level Coordination mechanisms (organization science) Division of work Direct supervision Standardization of work Standardization of skills Standardization of output Standardization of norms

Mutual adjustment

Group level Coordination mechanisms (group dynamics) Group composition Task assignment Leadership Goals Decision making Planning Interpersonal attraction Social affiliation Trust Cohesiveness Norms Information sharing and communication Feedback and corrective actions

BPR approaches heavily rely on the information processing paradigm of Galbraith (1973). Their perception of coordination resembles the organization science perspective with a strong emphasis on standardization of work and (standardization of) information sharing between tasks and groups in workflows and between primary process and management. In contrast the situated action perspective relies heavily on plans as the key coordination mechanism. Bardram (1997a) explains the value of plans in anticipatory reflection on recurrent actions (Weick’s guidance from the past) and defines situated planning as the building, altering, sharing, executing and monitoring of plans within cooperative work activities. Also in Bardram’s case description of coordination in a hospital context the main issue proves to be the utilization of past experience in the form of plans, schedules and check lists together with the management of the necessary deviations from those anticipated plans. This is similar to Weick’s description of ambivalence as the optimal compromise (Weick, 1998). Note that in organization science and BPR the coordination mechanism is chosen in the design phase and is standardized and fixed during execution, whereas in situated action the coordination mechanism is still negotiable during execution of work.

48

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION Vreede (1995) notes that several authors distinguish between coordination mechanisms on the process level on a temporal basis (Ching et al., 1992; Kieser and Kubicek, 1977; March and Simon, 1958). Anticipatory coordination or coordination by plan occurs before execution of activities, adaptive coordination or coordination by feedback takes place during execution of activities, and review coordination is done after execution of activities. Malone and Crowston (1994) note that a special feature of coordination in human systems is that incentives, motivations and emotions play a major role. Weick (1985a; 2001) does not discuss coordination explicitly. However, in the previous section we concluded that the connections between loosely coupled segments (Weick, 1985a) resemble dependencies. And because coordination is the management of dependencies, the way in which organizations deal with these connections may include notions of coordination. We recall that connections produce ambiguity in organizations because they are of variable strength. And because ambiguity cannot be fully removed the organization needs to cope with ambiguity as well as to reduce it to tolerable levels. The obvious way to reduce ambiguity is to strive for connections of constant strength. Weick (1985a) gives the following examples how the variability of connections can be decreased: • Rules: When there are more rules, more severe rules, clearer rules and rules with less latitude for deviation connections become tighter and less variable. • Agreement on rules: The more agreement there is on the content of rules, the nature of violations, and how violations will be handled, the tighter the coupling between loosely coupled segments in organizations. • Feedback: the sooner people learn about the effects of their actions, the tighter the coupling. • Attention: As attention becomes more constant, connections become more stable (when attention shifts due to changes in salience or need, connections become more variable). • Labels: Words and labels carry implications for action. They serve to focus attention, shrink the number of possibilities of what is going on and suggest appropriate action. • Action: Action itself can also simplify environments. Before action takes place the meaning of a situation is limitless. The situation takes only distinct form and meaning when action is inserted into it and when people reflect on their actions. • Interaction: By turning to another person and in common build some idea of what is occurring one can produce consensual definitions that may lead to coordinated action. Coping with ambiguity means finding a balance between high and low ambiguity by simultaneous loose and tight coupling and handling other related dilemmas (Weick, 1985a; 1998). This will be discussed in more detail in section 3.6 (page 55) when we discuss the improvement of coordination. Besides reducing and coping with ambiguity Weick and Roberts (1993) describe three crucial activities of the interrelating processes in a social system. When acting in a social system people construct their actions (contributing), while envisaging a social system of

49

COORDINATING DISTRIBUTED WORK joint actions (representing), and interrelate that constructed action with the system that is envisaged (subordinating). In this way interrelations are constructed and reconstructed continually by individuals. Weick and Roberts (1993) describe how collective mind develops when people can interrelate their social activities more heedfully. We will come back to the notion of heed in section 3.6 (page 55) when we discuss improvement of coordination. We conclude this section with a general definition of a coordination mechanism. Simon (1945) speaks of “organizational influence mechanisms that are meant to bring about coordination in the activities of the organizations members”. Schmidt and Simone (1996) define coordination mechanisms as consisting of protocols and artifacts. A protocol is a set of agreed on procedures and conventions, and an artifact is a distinct and persistent symbolic construct in which the state of the execution of the protocol is dynamically represented. A protocol can be more or less strict (acting as a script or a map). In situations where the protocol is beyond its bounds, ad hoc, informal coordination is necessary. The protocol may address issues like tasks to be executed, actors involved and their responsibilities, what information to be shared by whom et cetera. For our research we use the following definition. Definition 7: Coordination mechanism A coordination mechanism is an organization influence mechanism intended to manage dependencies between tasks or actors, and consisting of a protocol (a set of agreed on procedures and conventions) and an artifact (a symbolic construct that represents the state of execution of the protocol) Finally, we summarize the various coordination mechanisms identified in Table 15. Table 15: Coordination mechanisms on different levels Level

Coordination mechanisms

Organizational

Division of work Direct supervision Standardization of work Standardization of skills Standardization of output Standardization of norms Mutual adjustment Composition Anticipatory Interpersonal attraction Adaptive Leadership Review Decision making Task allocation Planning Communication Feedback Time management, personal effectiveness

Group

Individual

50

Rules Agreement on rules Feedback Attention Labels Action Interaction Contributing Representing Subordinating

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION Compared to the earlier discussions on work arrangements and dependencies, this overview of coordination mechanisms shows less fit between the different perspectives. The reasons for these differences become clear in the next two sections by addressing the nature of two basic elements of coordination (namely communication and decision making) and by answering the question how to improve coordination.

3.5 Communication and decision making In section 1.2 we already observed that communication and decision making play a major role in coordinating work in organizations: • Without communication organizational workers cannot influence the behavior of other workers and therefore cannot coordinate their activities (Malone and Crowston, 1994; Vreede, 1995). • Without decision making organizational workers cannot change their future behavior and therefore cannot attune their courses of action to that of other workers (Malone and Crowston, 1994). Communication Communication can be defined as the process where two or more parties exchange information and share meaning (Moorhead and Griffin, 1989). It has been studied both from a technical point of view (Shannon and Weaver, 1949) and from an organizational and social point of view (Allen, 1977; Putnam et al., 1996; Rogers and Agarwala-Rogers, 1976; Weick, 1969). Most of the time communication between individuals, groups or organizations is viewed as a transmission process. In such a process a source sends a message by encoding it into symbols and transmits it through a medium to a receiver who decodes the symbols and responds with feedback (an attempt to verify the meaning of the original message). Noise, anything that distorts or interrupts the communication, may interfere in any stage of the process (Shannon and Weaver, 1949; Moorhead and Griffin, 1989). Patterns of communication between persons or groups emerge as information flows and can result in relationships and communication networks. Group communication networks can be described in terms of the density of communication, the distance between members, the relative freedom of a member to use different paths to communicate with others and the position of members. Typical small group communication networks include the wheel, chain, circle and all channel networks. Roles that people can play in inter-group and organizational communication networks include gatekeeper, liaison, cosmopolite and isolate. The three primary methods of communication are written, oral, and nonverbal communication (Moorhead and Griffin, 1989). Effective organizational communication provides the right information to the right person at the right time and in the right form (Moorhead and Griffin, 1989). However, various problems can lead to impede the flow of information and give rise to misunderstandings between a sender and a receiver. Table 16 (adapted from Moorhead and Griffin, 1989) lists an overview of such problems. 51

COORDINATING DISTRIBUTED WORK Table 16: Problems in different phases of the communication process Phase

Communication problems

Source Encoding, transmission and decoding Receiver

Filtering (withhold information assuming the receiver does not need it) Lack of common experience (results in differences in encoding and decoding) Semantics – jargon (different meanings for same words) Medium (should be appropriate to message as well as intended receiver) Selective attention (only read or listen to selective parts) Value judgments (degree of reinforcement or change of personal beliefs) Lack of source credibility (influences attention for and value of the message) Overload (vital information can be overlooked) Omission (reception and understanding of message is not verified) Noise (high degree of irrelevant information) Status differences (focus on information from colleagues of the same status) Time pressures (misunderstanding or ignorance) Overload (misunderstanding or ignorance) Communication structure (standardized channels and messages may not be able to cope with situational circumstances)

Feedback Organizational factors

Shulman (1996) points out that the belief that communication is a tool, an instrument for getting your message across, has led to false and exaggerated expectations what information technology can do. His key point is that researchers should distinguish between the human and mechanical infrastructure of communication and should not apply a technical means of knowing to the human problem of communication. Similarly, Woerkum and Meegeren (1999) criticize the transmission view and point at the active role of the receiver, who actively constructs meaning rather than passively receives signals. In accordance with these critiques Putnam et al. (1996) give an overview of metaphors of communication that could serve as an alternative for the mechanical transmission view: • Conduit: The conduit view resembles the traditional transmission view as a conduit is a channel through which something is conveyed. • Lens: Communication is seen as a filtering process like searching, retrieving, routing information and deals with issues like distortion, blocking, gatekeeping and accessibility. • Linkage: Communication is a connector that links people together in networks of relationships. • Performance: Communication is seen as dynamic processes of interconnected exchanges in social interaction (performance refers to activity rather than output). • Symbol: In this metaphor communication is a process of representation and interpretation through the creation, maintenance and transformation of meanings. • Voice: Communication appears as the expression and suppression of voices of organizational members. • Discourse: Communication is seen as a conversation and focuses on language as the nexus for untangling the relationships among meaning, context and practice. Communication is the co-production of language in situated practice where text and context are intertwined with action and meaning. 52

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION Although both the rational perspective (dominant in organization science and information technology approaches) and the social-psychological perspective (enactment, situated action) see communication as a means to achieve coordinated action, they differ in their approach to communication. The rational perspective adopts the mechanical view of communication (conduit, lens and linkage), whereas the social-psychological perspective relies on views where interaction and meaning are central issues (performance, symbol, voice and discourse). Communication is also a means to reveal actual use of coordination mechanisms. It is only by looking at communication processes themselves that the actual patterns and mechanisms of coordination, both formal and informal, within organizations can be identified (Monge and Eisenberg, 1987; Qureshi, 1995; Vreede, 1995). Decision making Organizations are defined as purposeful systems (Ackof and Emery, 1972; Leeuw, 1990; Sol, 1982). From a centralized decision making perspective the concept of coordinating behavior (in order to reach the purposes of the organization) can be seen as an issue of control (Kickert, 1979; Parker Follet, 1937). The ultimate goal of this control is that the system (the organization) will display the behavior that is desired by the controller. From a more decentralized decision making perspective Simon (1945) states that (groups of) individuals need to be coordinated because they are not capable of applying the right strategies for their actions and decisions, due to a lack of insight about the strategies of other individuals. He defines coordination as providing everybody with insight into everybody else’s decision-making behavior, so that every individual is capable of making the right decisions. Simon (1960) describes three stages in the overall process of decision making: • Intelligence (finding occasions calling for a decision); • Design (inventing, developing and analyzing possible courses of action); • Choice (selecting a particular course of action from those available). Leeuw (1990), drawing upon the work of Simon, defines decision making as choosing from alternatives aiming at the goals to be attained. Simon (1960) discerns between a rational and behavioral model of decision making. In the early idealistic view of the homo-economicus the decision maker knows his goals exactly, knows all relationships between actions and results exactly, can oversee all alternatives and has unlimited information processing capacity. The emphasis is on analytic activities that can reveal the one best solution. Problems need to be decomposed into sub problems until these sub problems can be solved optimally. Simon (1960) introduces the concept of bounded rationality to deal with the imperfection of human decision making. Bounded rationality refers to the fact that decision-makers are not perfectly informed and do not have unbounded information processing capacity. They are satisficers instead of maximizers: as soon as a satisfying decision is found the search for more alternatives is stopped. He also observes a continuum from programmed decisions at the one end to nonprogrammed decisions at the other. Decisions are programmed to the extent that they are

53

COORDINATING DISTRIBUTED WORK repetitive and routine or a definite procedure has been worked out to deal with them. They are unprogrammed to the extent that they are new and unstructured or where there is no cut-and-dried method for handling the problem. Unprogrammed decisions require problem solving (Moorhead and Griffin, 1989; Pugh and Hickson, 1989). Cyert and March (1963) elaborate on the behavioral model by recognizing the importance of seeing organizations as a shifting multiple goal coalition. The composition of the firm is not given, but negotiated; the goals are not given but bargained. People in the organization do not behave rationally but have erratic preferences. They change or ignore their preferences or state them in an ambiguous way. Furthermore organizations have properties of organized anarchies. Goals are discovered rather than defined, processes are not completely understood resulting in trial and error behavior and there is fluid participation (Pugh and Hickson, 1989). Consequently decision making is not rational and following a specific set of sequential steps, but more like a garbage can model or organizational choice (Cohen et al., 1972; March and Olsen, 1976). The opportunity or need to arrive at a decision, to make a choice, can be seen as a garbage can into which various kinds of problems and solutions are dumped by participants as they are generated. A decision is an outcome of the interplay between problems, solutions, participants and choices, all of which arrive rather independently of each other and change continuously. An organization can be seen as a dynamic collection choices looking for problems, issues and feelings looking for decision situations in which they might be aired, solutions looking for issues to which they might be the answer and decision makers looking for work. A decision happens when participants who have the time and energy to do it attach solutions to problems and problems to choices (Pugh and Hickson, 1989). The latter leads to the question how to come to effective decision making in a context with such an unstructured nature. Bruijn et al. (1998) argue that a rational and reductionist approach focusing on content does not have much value in these kinds of situations because it neglects changes in the problem situation, new relevant information becoming available, and different perceptions and opinions of participants involved. Bruijn et al. (1998) provide an approach called process management for decision making in networks of actors that have different interests but depend on one another to solve their problems. This approach emphasizes a process orientation (a focus on how decisions are made). Major underlying principles of the approach are that an authoritative solution can only be found when all relevant actors are properly involved in the decision making process, and that the involvement of them can be planned and managed by an independent process facilitator. Key elements of the approach and attention points for the facilitator are (Bruijn et al., 1998; Bockstael Blok, 2001): • the need for a sense of urgency; • integrity and openness; • protection of central interests and core values; • stimuli for progress; • guaranteeing sufficient content related output.

54

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION In addition Bruijn et al. (1998) discuss some elements of effective behavior for participants in decision-making processes in a network context. Examples are the art of coupling (awareness for available problems and solutions among different actors, awareness for potential linkages between them, and awareness for the right momentum to create the linkage) and re-framing (redefining the perceptions of problems and solutions). Also having redundant and abundant relations with other actors yields important information and increases the likelihood of linkages to occur. We conclude that rational approaches take communication and decision-making for granted. They assume that the situation under study is analyzable and that the best or a satisfying solution with regard to ‘how and what to communicate’ and ‘how and what to decide’ can be defined. In that view communication and decision making are unproblematic elements of coordination. In contrast behavioral approaches emphasize the problematic nature of communication and decision making. They argue that these processes underlying coordination can be sources of coordination problems. It is selfevident that these different viewpoints have a profound impact on how each of them deals with improving coordination.

3.6 Improving coordination The key question that remains to be answered is how the various theories and approaches view the improvement of coordination: How can we distinguish between ‘good’ and ‘bad’ coordination? During our discussion in this chapter the differences arise between the views of rational and social psychological approaches. Therefore we present the way in which they answer this question separately here. Rational view (organization science, information technology) In organization science the choice for particular work arrangements and accompanying coordination mechanisms follows from a detailed analysis of the organizational mission, the environment and the work itself. In answering the question which kind of grouping serves the purposes of the organization and the demands of the environment we recall that Mintzberg (1983) distinguishes between 4 basic criteria that need to be balanced against each other: work flow interdependencies, process interdependencies, scale interdependencies and social interdependencies. Thompson (1967) argues that the problem with structuring organizations lies in the fact that components of complex organizations are not unidirectional. For example, an organization that groups its components according to geographic location will have problems to coordinate common processes. The design question to be answered is not ‘what criterion to use?’ but ‘what criterion deserves the highest priority?’. Thompson continues stating that organizations will group their components in such a way that minimizes coordination costs. Grouping tasks and people into work arrangements creates dependencies between them. These dependencies can either be pooled, sequential 55

COORDINATING DISTRIBUTED WORK or reciprocal. Since reciprocal interdependence is the most complex and hence most costly, followed by sequential, Thompson (1967, page 59) concludes: “The basic units are formed to handle reciprocal interdepence, if any. If there is none, the basic units are shaped according to sequential interdepence, if any. If neither of the complicated types of interdependencies exists, the basic units are shaped according to common processes (to facilitate the handling of pooled interdepence).” Considering the handling of multiple dependencies Thompson (1967) explains that the lowest level groups contain the major reciprocal interdependencies; higher order groups are then formed to handle the remaining sequential interdependencies, and the final groups, if necessary are formed to handle any remaining pooled interdepencies. Mintzberg (1983, page 62) concludes correspondingly: “By choosing the market basis for grouping the organization opts for work-flow coordination at the expense of process and scale specialization. Thus, if the work-flow interdependencies are the significant ones and if they cannot easily be contained by standardization, the organization should try to contain them in a market based grouping to facilitate direct supervision and mutual adjustment. If the workflow is irregular (as in a job shop), if standardization can easily contain work-flow interdependencies, or if the process or scale interdependencies are the significant ones (as in the case or organizations with sophisticated machinery), then the organization should seek the advantages of specialization and choose the functional basis for grouping instead.” Which of the general criteria deserves most priority in structuring the organization can be answered by looking at several situational or contingency factors. Mintzberg (1983) states that effective structuring requires internal consistency amongst design parameters as well as consistency between design parameters and contingency factors. Examples of contingency factors are (Mintzberg, 1983; Leeuw, 1994; Vreede, 1995): stability of environment, complexity of environment, market diversity, hostility of environment, technical system, age, size, power, fashion, task uncertainty and strategy. Leeuw (1994) illustrates that environment, technology (primary process) and strategy influence each other and together affect the structure of the organization. This confirms Mintzberg’s demand for fit between all those external and internal features. Although most contributions from the contingency perspective are on the organizational level and quite abstract, socio-technical design theory proves that the contingency framework can also be used on the group level (Leeuw, 1994). Kuipers and van Amelsvoort (1992) specified uncertainty by splitting it in the complexity and unpredictability of the order flow. Complexity refers to the amount of different types of orders and the magnitude of the differences between the various types. Creating parallel flows of more homogeneous order families lowers complexity. Unpredictability refers to the degree of uncertainty about the size and mix of the order flow at a certain point in time. If orders are extremely unique or the size of a flow is too small, it is not possible to create a separate order flow and then an ad hoc project structure is needed. Otherwise there needs to be flexibility to move people and equipment from one order flow to another. 56

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION Also in group dynamics influence of the contingency approach can be found (Moorhead and Griffin, 1989; Donaldson, 1996). Considering composition simple, sequential tasks that require cooperation and speed are best performed in homogeneous groups, whereas complex, collective tasks that require creativity are best performed by heterogeneous groups. Application of other coordination mechanisms depends on the effects desired. Larger group size provides more resources available, but may complicate internal coordination and reaching agreement. Strong group norms and cohesiveness make groups more efficient and productive, but may have negative consequences like groupthink. Furthermore frequent and open communication is often mentioned as an indicator of good work group functioning, but one may argue that excessive communication may lead to inefficiency. Adequate outputs, goal accomplishment and congruence of group and organizational goals are generally accepted as the major indicators for good work group performance. However, Shulman (1996) criticizes researchers that there is a lack of standards for good work group performance. The danger of goal achievement as a standard is illustrated by the observation of Galbraith (1973) that an obvious way to improve actual performance is to lower the level of required performance. He argues that the strategy of reduced performance standards is chosen if no other strategy is executed. Finally BPR approaches emphasize general principles such as process and activity simplification, bureaucracy elimination, standardization and technology utilization. Examples of guidelines are “capture information at the source”, “eliminate duplication of information”, “reduce information flow”, “reduce contact points”, “execute activities concurrently”, and “break complex processes into simpler ones” (Davenport and Short, 1990; Hammer and Champy, 1993; Kock, 1999; 2001). Note that some of these guidelines are contrastive to some from group dynamics and socio-technical systems theory that pay more attention to human issues. In summary the following guidelines how to improve coordination can be derived from the rational perspective: • Organizing is aimed at complexity reduction (Leeuw, 1994; Galbraith, 1973; Thompson, 1969). So first of all one should aim at work arrangements with low interdependencies and decentralize coordination mechanisms as much as possible. Both decrease coordination costs. • It should be considered that only variety can control variety (Ashby, 1969; Weick, 1969; Mintzberg, 1983). So when the uncertainty (variety and unpredictability) of the environment, the technology or the strategy increase, the work arrangements and coordination mechanisms should become more flexible and sophisticated to cope with this uncertainty. • There should be a fit between environmental, technological, strategical and structural features (Mintzberg, 1983; Leeuw, 1994). Otherwise there are avoidable intraorganizational pressures that cost a lot of energy. Leeuw (1994) summarizes that the extent to which the environment, technology and strategy are divisible determines what should be coordinated (where dependencies between tasks and actors are inevitable). Subsequently, the predictability of environment, technology and strategy determines how the dependency should be coordinated. High and 57

COORDINATING DISTRIBUTED WORK easy predictable dependencies can be coordinated with simple rules. High but hard to predict dependencies correspond with complex, bureaucratic, tight and inflexible coordination mechanisms. Finally low predictability asks for adaptable and flexible coordination mechanisms or buffers. The social psychological view (sensemaking, situated action) Now we understand how rational approaches tackle the improvement of coordination, we continue our discussion with the view of social-psychological approaches. As discussed before, Weick (1985a) argues that connections of variable strength between loosely coupled systems produce ambiguity. Starting point for coordination is the observation the organization needs to cope with ambiguity as well as to reduce it to tolerable levels (as ambiguity cannot be removed completely). In section 3.4 we conclude that Weick provides some clues for good coordination when he elaborates on how one should cope with ambiguity (Weick, 1985a; 1998) and when he describes collective mind and heedful interrelating (Weick and Roberts, 1993). In addition we discuss Weick’s warning (Weick, 1985b) that electronic information processing complicates the understanding of events (and thereby leads to bad coordination). Finally we address directions how to improve coordination apparent in the discipline of situated action. Coping with ambiguity is a central theme throughout all aspects of sensemaking. Weick (1998) argues that organization should not only reduce ambiguity, but also take advantage of some of the unique opportunities for change that occur when ambiguity increases. This is reflected in dilemmas like ‘simultaneous loose and tight coupling’, ‘ambivalence as the optimal compromise’ and ‘extensive deliberation and subtle action versus vigorous action and crude analyses’. Weick’s view how to handle these dilemmas corresponds with the way of thinking in the school of competing values (Quinn, 1988; Vaughan, 1996; Collins and Porras, 1997). Both sides of such a dilemma need to be combined: not by blending them into one gray indistinguishable whole, but by using them distinctly at the same time, all the time. As an example Weick (1998) explains that adaptability requires loosening and adaptation requires tightening. Highly differentiated organizations can both function when they are highly integrated by tight coupling between dispersed specialists as well as when they are loosely integrated because generalists make autonomous operating of units possible. Although loosely coupled systems may be inefficient, Perrow (1984) and Weick and Roberts (1993) show examples where tight-coupled systems may be dangerously ineffective because they cannot deal with sudden changes in the environment. Banner and Gagne (1995) define mature organizations as those who are able to combine bureaucratic and organic features whenever needed. Similar lines of reasoning can be put forward for usage of past experience and for the amount of deliberation before taking action. Ambivalence in using past experience is needed to balance between knowing and doubting. Knowledge should be treated as fallible and substantial. In solving problems this implicates that one should balance the use of past experience with alertness for the non-routine in the present. Likewise accurate sensing gives opportunity to learn about the situation as it originally existed, whereas strong action leaves more energy and commitment for getting the job done. 58

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION From studies of high reliability systems (where small events can escalate into catastrophic accidents) Weick and Roberts (1993) distill the importance of heed (to act with attentiveness, alertness and care), conduct (taking into account the expectations of others) and mind (integration of feeling, thinking and willing actualized in human behavior exhibiting qualities of intellect and character). Collective mind develops when people can interrelate their social activities more heedfully (Weick and Roberts, 1993). Interrelations are constructed and reconstructed continually by individuals through the ongoing activities of contributing, representing and subordinating. Heedful interrelating can be characterized by heedful contributing (loose coupling, diversity, strategic communication), heedful representing (mutual respect, coordination of action) and heedful subordinating (trust). These characteristics are quite different from developed groups (alignment of cognition, agreement, empathy, homogeneity, tight coupling and unrestricted candor). Weick and Roberts (1993) argue that developed group and developed mind are two different matters. An example of developed group and undeveloped mind is the phenomenon of groupthink. An example of developed mind and undeveloped group can be found in ad hoc project teams or groups in their forming stages. Weick and Roberts (1993) hypothesize that groups may be smartest in their early stages and loose mind when interrelating becomes more routine, casual and automatic. He also states that organic structures may have more capacity for mind than hierarchical and bureaucratic ones. Furthermore Weick and Roberts (1993) raise the possibility that technological tight coupling (Perrow, 1984; Leeuw, 1994) does not have to be dangerous in the presence of interactive complexity, as long as it is accompanied by socially tight coupling (by heedful interrelating). The accidents described by Perrow could be traced to flawed mind rather than flawed equipment. Several authors (Lave and Wenger, 1991; Orr, 1990; Sutton and Luis, 1987; Weick and Browning, 1986) stress the importance of narrative skills of insiders and the attentiveness of newcomers in ongoing social processes. When experienced insiders answer the questions of inexperienced newcomers not only the newcomers learn how to act heedfully, but also the insiders remind themselves how to act heedfully and how to talk about heedful action. Weick (1985b) warns for disruption of the sensemaking process when people spend a lot of time behind computer screens. The resulting loss of meaning hampers people in accurately diagnosing problems they are expected to solve. Weick (1985b) states that people using information technologies act less, compare less, socialize less, pause less and consolidate less, whereas all these activities are crucial to sense making. Weick (1985b) argues that many of these problems can be solved if people simply push back from their computers and walk around. Then they generate outcomes (feedback on actions), compare sources of information (validate cryptic data), meet people and discover what they think (socialize), slow down the pace of input (time for deliberation) and get a more global and contextual view of what is happening (consolidate). The electronic world makes sense only when people are able to reach outside that world for qualitatively different images that can flesh out cryptic representations. Reaching out should thus be stimulated. In addition Weick (1985b) draws some related conclusions. First the problem of information overload is not just too much information, but too much information of the same kind. Increasing the different types of information limits overload. Secondly, people need to listen more and talk less. Listening for the pattern, and by editing transmit the 59

COORDINATING DISTRIBUTED WORK pattern rather than the details, prevents from paralyzing themselves and others with too much detail. Finally corporate culture may become more important as it provides a framework within which cryptic data become meaningful. In contrast the situated action perspective sees ICT as useful instruments to support situated planning (Bardram, 1997a). The technological device is the artifact that represents the state of execution of a plan, and thereby creates a shared understanding of their work process for the actors involved. Although not explicitly termed so, the main indicator for good coordination is the malleability of the coordination mechanism during its use to the specific conditions of the work situation. This holds for both the protocol, the artifact and the actors involved. The actors involved need to actually sense the need for modification and effect it. This behavior resembles Weick’s view on ambivalent use of prior knowledge in current situations (Weick, 1998). The guidelines for improving coordination from this social psychological perspective are quite different from the ones summarized earlier: • Besides reducing complexity by information processing, one should pay attention to reducing and coping with ambiguity by interactive sensemaking. • There seems to be a shift in focus from limiting coordination costs (efficiency) to guarding coordination effectiveness (heedful interrelating). Too lean and efficient work arrangements where no slack resources are left can result in rising control efforts and declining performance (Leeuw, 1994). • Design of work arrangements and coordination mechanisms is not bounded and fixed in time, but is a continuous socially constructed process (Ciborra, 1996; Weick, 1993): An organization is a network of intersubjectively shared meanings that are sustained through the development and use of a common language and everyday social interactions (Walsh and Ungson, 1991). • Finally action (and thus also coordination) is situated, which means that one should balance experience from the past reflected in anticipated plans for recurrent actions with alertness for the non-routine. When every situation is a unique performance evaluation seems impossible because yardsticks are by definition not generally applicable and subjective for that unique situation.

3.7 Conclusions The aim of this final section is to illuminate the most important observations in the preceding discussion that are important for guiding observations in our participatory case studies. Therefore the following issues are discussed below: • How can we recognize and identify coordination in practice? • How can we recognize and identify coordination problems in practice? • How can we recognize and identify initiatives to improve coordination in practice?

60

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION Elements, levels and aspects of coordination In our literature study we describe that coordination starts with a division of labor that results in a certain work arrangement. This work arrangement creates dependencies between tasks and actors that need to be managed by coordination mechanisms. In addition we observe that coordination occurs on the organizational level, the group level and the individual level. Furthermore aspects of coordination that require attention are behavioral, IT and context (situational, organizational en environmental). We conclude that these elements, levels and aspects are useful anchors to guide our observations in the case studies. The overviews of possible work arrangements (Table 11, page 45), possible dependencies (Table 12, page 46) and possible coordination mechanisms (Table 15, page 50) are helpful to recognize elements of coordination, and to check whether we overlook particular ones. In these overviews the levels of coordination are already included. As illustrated in Table 7 (page 37) no theory or approach covers all identified aspects of coordination. In our problem statement in chapter 1 we hypothesize that all these aspects each have their value to arrive at a complete picture of coordination in organizations. Therefore we need to look in our case studies whether all aspects are identifiable and how they affect coordination. Sources of coordination problems: uncertainty and ambiguity During the discussion in this chapter we observe differences in the way how rational and social-psychological approaches to coordination picture coordination problems, coordination mechanisms and directions for improvement. Here we focus on the differences in defining coordination problems. We suggest that differences in understanding of reality explain different views on what is problematic in coordination. From the rational perspective there is one objective truth and some people are closer to that truth than others. Organizations should aim at controlling their environment and responding to its demands. Their main coordination problem is uncertainty about the environment and about the tasks to be performed. This is most prominent in the information processing perspective of Galbraith (1973): • Uncertainty is defined as the difference between the amount of information required to perform the task and the amount of information already possessed by the organization. • The greater the uncertainty, the greater the amount of information that needs to be processed among decision-makers during task execution in order to achieve a given level of performance. Consequently more information or more information processing capability will lead to a more effective solution in situations of uncertainty. Given this line of reasoning it is not surprising that Galbraith (1973) sees opportunities for ICT to improve information processing capability of the organization and thus an opportunity to increase performance in situations of high uncertainty. Similarly Simon (1960) has been enthusiastic about ICT for improving decision making.

61

COORDINATING DISTRIBUTED WORK According to Daft and Weick (1984) there is neither one objective environment nor one objectively observable organization carrying out its tasks. Instead of objective facts there are just subjective interpretations. Daft and Weick (1984) state that organizations have to build their environments before they can even have the luxury of controlling them. And strictly speaking they are then not controlling their environments, but controlling their own perceptions, their constructed environments. Consequently, coordination is not a problem of uncertainty but a problem of ambiguity. When people create different subjective realities the issue is to give meaning to each other’s reality and to reach a state of committed interpretation. Therefore Weick (1985b) and others who adopt this view (Shulman, 1996; Velde, 2002) are critical about the value of ICT. In their opinion there is no need for more and faster information processing, but a need for more interaction and time for interpretation and sensemaking. We conclude that coordination problems can result from uncertainty and ambiguity. In our exploratory case studies we need to pay attention to both sources of coordination problems. Furthermore we are particularly interested to find out whether these different coordination problems are solved with different improvement strategies. The difficulty of guiding and measuring improvement of coordination We conclude that the studied theories and approaches are rather vague about guiding and measuring improvement. This is more an observation than a blame. Our extensive discussion in section 3.6 shows that improving coordination is complex as a result of the complex nature of coordination itself. Furthermore it proves that the conception of coordination and coordination problems heavily influences the directions for improvement. This is an important observation to remember when we develop our own conception of coordination in chapter 5. Here we shortly summarize the difficulties of guiding and measuring improvement and the implications of that for observations in our case studies. Vreede (1995) summarizes several problems that arise when one tries to optimize coordination from a rational contingency perspective: 1. Several authors have tried to define a standard optimal coordination mechanism for specific organizational circumstances. However, the contingencies found influence one another, which leads to conflicting results. So there is no standard coordination mechanism for specific organizational structures. 2. Secondly, if there would be contingency relationships, the constant change and necessary adaptation of organizations would ask for a continuous adaptation of the coordination processes used. 3. Thirdly, decision making and communication are important, required and complicated processes in all instances of coordination (Malone and Crowston, 1994). The uncertainty about their outcomes and manageability is a complicating factor in analyzing coordination processes. 4. Finally, the extent to which an organization has improved its coordination cannot be measured directly, but should be measured in terms of indicators or evaluation criteria (Malone and Crowston, 1994). Examples are costs of coordination, delivery time, service level and throughput time (Eijck, 1996). 62

CHAPTER 3: THEORETICAL VIEWS ON COORDINATION The social psychological perspective answers to these problems by extensive attention for the openness and quality of communication processes (3), by making the continuous change in coordination mechanisms a central issue (2) and by adopting contingencies as possibly valuable past experience (1). The difficult issue remains how to measure improvement of coordination (4). Heedful interrelation and ambivalence are hard to put into practice and to specify in hard evaluation criteria. Banner and Gagne (1995) state that any successful attempt to measure organizational effectiveness should be multifaceted (drawing on aspects from different models, like we have presented in this chapter). For the observations we need to do during our case studies two issues are important: • First, it is interesting how organizational members picture ‘good’ and ‘bad’ coordination, and how they direct and monitor improvement of it • Secondly, its is interesting to see whether their improvement strategies resemble strategies advocated by rational approaches or social-psychological approaches or whether they combine strategies from both kinds of approaches. With this initial theoretical framework (consisting of elements, levels and aspects of coordination, different sources of coordination problems, and different approaches to improvement of coordination) in mind we enter our case studies in the next chapter.

63

4 Coordination at the Amsterdam Police Force This chapter presents the results of our first exploration of coordination in practice. Two different distributed work processes within the Amsterdam Police Force are analyzed in order to find out what types of coordination mechanisms, coordination problems, ICT opportunities and promoting and hampering factors can be discerned. Section 4.1 describes the case study designs by discussing the units of analysis, the selection of the two work processes and the data collection sources. The second and third section each concentrate on one of the two analyzed work processes, respectively the work group Youth Annoyance and the Regional Project (Fire)Arms. The final section reflects on the two case studies by discussing the differences and similarities between them. The outcomes of that reflection result in a set of requirements for our design approach, which is developed in chapter 5 and 6.

4.1 Case study set-up Given the findings of our literature studies in chapter 1 and 3 we look for the following units of analysis and data in our case studies: • (potentially) distributed work processes; • dependencies between tasks and actors involved in these work processes; • coordination mechanisms on organizational, group and individual level; • coordination problems (related to uncertainty or ambiguity); • strategies for improving coordination; • ICT opportunities; • promoting and hampering factors from organizational, technological, social and behavioral origin. The case study results in section 4.2 and 4.3 are structured following these units of analysis, which simplifies the transition to answering our research questions. Known contingency factors that affect the coordination mechanisms used are for example task variability and complexity, group size and group maturity. As our objective is to derive a (complete) set of coordination mechanisms for distributed work settings we aim to observe two work processes that are reasonably different in nature. The differences between the two processes selected are depicted in Table 17 below. The main reason for 65

COORDINATING DISTRIBUTED WORK not conducting research on multiple work groups and multiple regional projects is the limited availability of time and resources. Furthermore we assume that an in depth study of one work group and one regional project is of more value than a superficial study of several ones of each. Moreover some documents of earlier investigations can be used to discuss generalization of the results to similar groups and regional projects. Table 17: Differences between the descriptive case studies Factor TASK VARIABILITY TASK COMPLEXITY GROUP SIZE GROUP MATURITY

Youth Annoyance Ad hoc Multi-disciplinary Small Short term relation

(Fire)Arms Routine Mono-disciplinary Large Long term relation

The data sources used for each case study are shown in Table 18. Table 18: Data sources in the descriptive case studies Quantitative data source Journals of communicative actions Questionnaire Qualitative data source Organizational documents Organizational archives Interviews Observation of meetings Observation of work practices Evaluative group session

Youth Annoyance X

(Fire)Arms -

Youth Annoyance

(Fire)Arms

X

X

X

-

X

X

X

X

X

-

X

X

-

During the analysis and reflection stage our main points of reference are our research questions. In addition we use the analysis framework of adaptive structuration theory (DeSanctis and Poole, 1994), which discerns between task related, technology related, group related and organization related factors that affect technology adaptation. By using this well-known framework our observations become comparable to other studies.

4.2 The work group Youth Annoyance The work group Youth Annoyance is a work group consisting of six regional coordinators who need to give an advice how to deal with Youth Annoyance problems in District 6. This section starts with a description of their work process as this is the focal unit of analysis. The main work group activities and the activities of individual group members in the period from March until July 1998 are discussed chronologically. Successively the work process is revisited four times and reflected upon from different viewpoints (coordination mechanisms, coordination problems, ICT opportunities and factors hampering ICT utilization). 66

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE The researcher attends all group meetings and some related activities in District 6. Early in the collaboration process all work group members are interviewed to learn about their daily work practices, their backgrounds and their expectations considering the collaboration process. Also some policemen in District 6 are visited and interviewed to get an impression of their work process. The interviews in District 6 mainly focus on the Youth aspect. Furthermore the work group members are asked to keep a logbook on their interactions with each other outside the meetings and on their interactions with policemen in District 6. These logbooks are discussed individually with each work group member between the eighth and tenth group meeting. Finally an evaluative meeting is organized where the researcher and the work group members look back upon the collaboration process and outcomes. 4.2.1 Work process During the first two months of 1998 youth annoyance and criminality in District 6 (Amsterdam West) escalates. The following factors contribute to this escalation: • Youth rules the street (“policemen are not allowed to give someone a ticket here”, “trying out of new policemen”, “not only more, but also more severe criminal offenses”). • Escalation and massiveness of incidents (normal police action like moped controls, traffic accidents escalate because of provocation and drumming up colleagues at both sides). • Demographic trends (within 10 years 50% growth of Youth, 85% of total Youth will be Moroccan, and communication with Moroccan Youth is most problematic currently). • Social work organizations (efforts of different organizations are not coordinated mutually and with police strategies, promises are often not kept). • Work pressure (too high in the neighborhood teams). As a result of these developments the district head of District 6 requests assistance from the group of Regional Project Coordinators. He asks for two advices, namely: • What can his people do best, considering their work methods aiming at Youth problems, and how could regional project coordinators possibly assist? • How can he increase control over the activities of the social work organizations? A work group is formed that consists of 6 Regional Project Coordinators (Youth, Social Integration, Traffic, Robbery, Firearms and Car Criminality). Using their universal police experience and their specific expert knowledge they produce an advice for District 6. Besides the concrete assignment from the head of District 6 the group has a second (internal) goal. The Regional Project Coordinators feel that they operate too domain specific currently and want to gain experience in finding integral solutions for practical problems (instead of letting neighborhood teams merging the knowledge and work methods from several domains). From this point of view the collaboration process is also an experiment in finding out what kind of problems they can solve together and how they can work together productively despite their different backgrounds. 67

COORDINATING DISTRIBUTED WORK Our study encompasses the internal coordination within the work group as well as the individual and group interactions with (people from) District 6. Most of the internal coordination occurs in the 10 group meetings that are organized approximately once a week. Table 19 gives a rough overview of these group meetings (G) and important events in District 6 (D6). Next the individual actions of the group members are shortly discussed. A more elaborated discussion can be found in the case study report (Laere, 1998a). Table 19: Overview of group meetings (G) and events in District 6 (D6) Date and type

1

April 7th

G

April 13th

G

April 17th April 21st

D6 G

April 23rd

D61

April 28th

D6

May 7th

G

May 11th May 12th

D61 G

May 12th

G

May 19th

G

May 20th May 26th

D6 G

Description Monthly meeting of all Regional Project Coordinators. The general manager Professionalism and innovation puts forward the request of the head of District 6. Six RPCs volunteer to join a work group. 1ST MEETING: Exploration of the problem in an open discussion. Two members will visit the head of D6 to further clarify the assignment. RPC Youth will provide literature on Moroccan Youth. Two work group members visit the Manager of D6. 2ND MEETING: Enumeration of known problems with Youth in D6. In addition each work group member will ask his/her contact persons in D6 how they perceive the current situation. Incident August Allebe square. Simple booking escalates in massive fight between Youth and policemen. “The fat is in the fire”. Tense situation in the following weekend. Debriefing with full management of D6 and lots of policemen from the street. Discussion results in appointment of a Youth-council: they need to formulate a plan to make it through the summer. 3RD MEETING: Review of group process thus far. Team building costs energy and delays advice for D6 but is experienced as useful. Doubt how to proceed after the events in D6. In D6 a local Youth-coordinator is appointed for 3 years. Monthly meeting of all RPCs: The general manager Professionalism and innovation indicates that the group should focus on youth problems in the whole region in addition to those in D6. The team building process is judged as valuable. 4TH MEETING: Reformulation of the group assignment. Team building discussion on pigeon holing versus an integral view and multiple track policy. 5TH MEETING: RPC youth starts acting as group leader. Integral discussion of problems and solutions results in some great ideas. Also belief that the youth-council deals with short-term solutions and this work group should focus on long-term solutions. Intention to invite the RPC (civil) Services and the Youth-coordinator of D6 (who also heads the Youth council).

First meeting of youth-council in D6. 6TH MEETING: Dinner with one of the neighborhood team managers of D6. Conclusion that supporting conversations are more needed than recommendations concerning content.

These events are not attended by the researcher.

68

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE Date and type May 28th D61 June 3rd June 5th June 8th D61 th June 9 G

June 16th

G

June 25th

G

July 6th July 7th

G

Description Second, third and fourth meeting of youth-council in D6. Presentation of Youth-council to management of D6. 7TH MEETING: Introduction by the Youth-coordinator of D6. Distinction between content and process/culture in discussion. Choice to focus on process aspects in the advice. Appointments who will write the advice on July 6th and 7th. 8TH MEETING: Discussion on “need to be closer to actual practice” and on “culture and process” problems. 9TH MEETING: Evaluative group decision room session. Positive and negative experiences are shared and reflected upon. 10TH MEETING: Three RPC’s (selected from the work group) write an advice for manager D6 which is a proposal for coaching and empowering managers in D6.

The regional coordinators Youth and Social Integration have their offices in District 6 and in addition their discipline is most related to the problem the work group is dealing with. Not surprisingly they were the ones that dedicated most of their time on contacts with people from District 6. The coordinator Youth perceived a lot of commotion and dissatisfaction with managerial decision making on the work floor just after the incident on April 23rd (see Table 19). Repeatedly she tries to initiate consultation between management and employees on the work floor. This results ultimately in the debriefing on April 28th. The appointment of the local Youth coordinator and the Youth council is thus an indirect result of her individual efforts. The approach how to deal with Youth annoyance during the summer in District 6 (formulated by the Youth council) is communicated regionally by her. Furthermore the regional Youth coordinator achieves that top management adopts a less aggressive tone in press. This “natural militancy” was one of the reasons for reinforcement of escalation of Youth annoyance. Also she claims capacity for criminal investigation to detect the main offenders and she identifies the potential areas in the region where Youth annoyance could escalate in the near future. Finally two meetings are organized where top management and “Youth experts from the street” discuss current problems and police strategies. The regional coordinator Social Integration has numerous discussions with policemen and members of welfare organizations in District 6. These conversations aim at gathering the multiple perspectives of the various actors. Twice he arranges a course “communicating with Moroccans” for some lower level policemen who feel the need to improve their communication skills. Furthermore he influences and supports the strategies of the urban district council and some welfare organizations surrounding the Youth problem and the police force.

69

COORDINATING DISTRIBUTED WORK On the repressive side the regional coordinator Firearms arranges that questions on Firearm possession are standard included in the hearing of HALT2 settlements. Considering prevention the regional coordinators Robbery and Firearms participate in a school adoption project. Policemen that adopt a school are now instructed how to inform and educate pupils on these topics. Furthermore the regional coordinator Robbery plans to focus on young offenders during his yearly autumn project (when robberies are traditionally rising). Robbery and Youth professionals also brainstorm on possible strategies together. 4.2.2 Coordination mechanisms This section lists the coordination mechanisms observed in the work group and between the regional coordinators and District 6. Within the work group the following coordination mechanisms are observed: • the assignment from the head of District 6; • the (partly voluntarily) composition of the work group; • the 10 group meetings; • the appointment with the head of District 6; • the extra assignment from the Head Professionalism and Innovation; • the appointment of the regional coordinator Youth as chairwoman for the meetings; • several plans for future action (summarize literature, contact people from the work floor, invite people from District 6, writing of the final proposal); • actual invitation of the neighborhood team manager and the local Youth coordinator; • the meeting minutes. Considering the individual efforts of the regional coordinators and the coordination between their expertise field and the 6th District the following coordination mechanisms are observed: • individual appointments and discussions; • attendance of a management meeting; • regular meetings with professionals from their expertise field; • work procedures or directions (HALT; school adoption). • participation in temporary project groups (school adoption, autumn project, this work group) • courses communicating with Moroccans. A first tentative conclusion is that the police force makes use of a wide variety of coordination mechanisms on both the organizational and the group level, with a strong emphasis on communication and mutual adjustment (e.g. meetings).

2

Special procedure for young criminals.

70

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE 4.2.3 Coordination problems From the observations of the researcher, the individual interviews and the evaluative session a set of problems is gathered, which are discussed below. Coincidental selection of assignment and work group participants The establishment of the RPC work group is more a coincidence than an outcome of an existing and recurrent decision making process. Factors that contribute are the fact that the regional coordinators feel a need to increase their collaboration, the introduction of the researcher and the request from District 6. In future it is preferable to collect possible assignments, select the most important or promising ones, appoint work group members and discuss their progress more formally. Also the selection of work group members is anything but structured. Because the work group is seen as an experiment, participation is on a voluntary basis. In the end this proves to be important, as keeping up in hard times is a significant success criterion. However, in future work group members should both be selected on experience with the topic as well as on motivation for the project. Diverging views on assignment and lack of contact with principal During the meetings a lot of time is spent on discussing the exact aim of the assignment. Although the principal gave some very clear expectations, the group feels that they need a more open assignment. Of course the incident on April 23rd and the subsequent reactions in District 6 hamper and reinitiate this discussion. Furthermore during the 4th meeting the group concludes that they have 3 different assignments: an advice for District 6, an advice for the whole region and their own quest for collaboration possibilities between regional coordinators. Remarkably there is no further consultation with the principal from District 6 besides the appointment on April 17th. Although broad assignments may be cherished, it is recommendable to make the agreed upon formulation of the assignment explicit. Now the various regional coordinators keep a different interpretation of the assignment for quite some time. Also sharing of this assignment and progress with the principal is needed. The regional coordinators mark the lack of contact with the principal as a negative point during the evaluation. Noncommittal attitude in meeting attendance Only 3 of the 6 regional coordinators attend all meetings. The others miss respectively 2, 3 and 7 out of 10 meetings. Reasons are illness (once), date unknown (3 times), and conflicting appointments (8 times). The fact that dates are not known can be blamed upon the limited use of meeting minutes. The fact that other obligations are given priority is partly inevitable and partly caused by being absent when new appointments are made. The noncommittal attitude can be a result of a lack of belief in the usefulness of the collaboration. In future, assignments should be examined on their need for an integrative approach and group members need to be genuinely committed.

71

COORDINATING DISTRIBUTED WORK Absence complicates keeping track of the work group process and progress. That becomes especially apparent in this work group where communication between meetings is minimal and meeting minutes are hardly used. Group development and laborious communication The differences in backgrounds, perspectives, views, ways of working et cetera between individual coordinators are both an advantage as a disadvantage. In the beginning it heavily delays the work process. On the other side it eventually results in original solutions that none of the individuals would have reached alone. The meetings can be split in several phases. The first two meetings can be characterized as start up phase. There is much discussion on what should be done and how it should be done. During the second meeting people argue out of their own ‘pigeon hole’. They all place remarks like “I see it this way” and “I have done it so and so”. Relations between topics are sometimes mentioned but get lost in the discussion. In reaction to this dissension the third and fourth meeting are focused on team building rather than on the problems of District 6. The fifth meeting is experienced as a turning point in the whole process. Truly integrative discussion results in a set of creative ideas detached from the different backgrounds of group members. The next three meetings are a more detailed elaboration on these ideas. The final two meetings are reserved for evaluation and the writing of the advice. Poorly structured meetings During the first meetings there is no chair(wo)man, no agenda and no meetings minutes are taken down. Later (starting the fifth meeting) the meetings are headed, first by the Youth coordinators, later by rotating chairmanship. The agenda is from that point in time formed by putting forward the relevant issues at the beginning of each meeting. This agenda is followed quite well. The meeting minutes are not used during subsequent meetings and even not during the writing of the final advice. The ‘chaos’ during the first meetings is both increased by the differences in background and the different views on the assignment. During the evaluation this chaos is both interpreted positively as well as negatively, as illustrated by one of the comments of the coordinators: “Considering our short term advice for District 6 our work was rubbish. For our own collaboration this chaos eventually seemed to be functional, although I did not believe that in the beginning”. A negative effect of the chaotic work process is that valuable ideas, plans for activities and intentions to invite several people are forgotten during subsequent meetings. The activities are not executed or the results are not discussed, some of the people are never invited and the ideas are not incorporated in the final advice.

72

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE Individual actions are unrelated to group process During the analysis it becomes clear that besides the group process quite some individual initiatives are carried out with respect to youth in District 6. Especially the coordinators Youth and Social Integration are more involved because of their professional relation as well as physical proximity. This meant that they are more informed, and do know more of the ins and outs of the events in District 6. For the other group members there is no structural reporting of what happens in District 6. Sometimes they only discover in a late stadium that they do not know of something important. Although the group feels that it needs to increase the sharing of important individual actions, some coordinators express that they experience difficulties in judging what is relevant and what is not. Severe complaints arise about meetings in which time is wasted when people elaborate on issues that are of minor importance for the group. Reliance on group meetings as one and only communication mode Contacts between the group members outside the 10 meetings are very scarce, except between members who are working in the same building (Youth and Social Integration share an office in District 6; Robbery and Car Criminality share an office on Police Head Quarters). Problems related to this are late reactions on incidents (postponing until the next meeting), losing track of the process if you miss one or more meetings, and forgetting ideas that arise between two meetings. Also time is lost on information sharing: documents had to be read at the beginning of a meeting, while the authors had to wait and ‘do nothing’. Instead the precious time together could have been used for discussion. Long throughput time Everybody agrees that the throughput time of the whole process is much too long. The main cause is the double assignment (problem solving and team building). An alternative discussed during the evaluative session is to produce a quick advice by working together for one or two days full time. Comments on that alternative are the problem of inviting all relevant participants (with possibly conflicting agendas) and the danger that the advice becomes too quick and dirty (no time for development of creative ideas). Overall agreement is that a clear deadline from the principal would help. Lacking involvement of policemen from District 6 Although the importance of involving people near to the problem has been recognized repeatedly, only twice a policeman from District 6 is invited. This is evaluated as one of the major negative points. The final advice compensates this a little bit by encompassing a proposal for coaching neighborhood team managers, which would intensify the relation between regional coordinators and neighborhood teams. The final conclusion is that the main contributions from regional coordinators to the problem solving process in District 6 are some individual efforts. The work group is more a learning process than a result-producing project. When more work groups of this kind are formed in the future, collaboration and coordination can be improved on the points discussed in this section.

73

COORDINATING DISTRIBUTED WORK 4.2.4 ICT opportunities During the whole research period the researcher looks for possibilities to improve the work process by introducing ICT support. Also the work group participants and policemen from District 6 comment on possible ICT use. Possible ICT use is identified by wondering to what extent it would contribute to observed coordination mechanisms and how it could help in solving the described coordination problems. This results in the following ideas: • a bulletin board where police managers (or possibly also policemen from the street) could send in possible assignments for the regional coordinators to tackle collaboratively; • a possibility for regional coordinators to attach their name to one or more of these assignments indicating that they would like to join that work group; • if desired there could be a possibility to vote on these assignments to assist the decision making process which assignments to select for execution; • a shared folder including files with the agreed on project assignment, meeting minutes, literature reviews, interviews with people from District 6 and so forth; • a shared mail folder, that can be used to inform each other of important events in District 6 and about individual actions undertaken; • use of a calendar system (like Microsoft Outlook) to be sure that new meetings are scheduled on a time everyone is available; • if needed, a video recording device to be able to give people that were not present an impression of an important meeting or discussion (they can watch it later on); • a bulletin board where policemen from District 6 could contribute problem anecdotes and ideas for solutions that could be incorporated in the discussion of the regional coordinators; • the work procedures could be placed on an intranet for all policemen; • the key lessons from the course on communication with Moroccans could be put on intranet to reach a broader public and to motivate others to take the course; • the monthly meeting of professionals from an expertise area could be preceded by a discussion on a bulletin board identifying what issues they would like to address. It is clear that this list of ideas is not finite. ICT can assist coordination of distributed work processes and solving of experienced coordination problems in many ways. On the other hand ICT themselves will not solve the problems. You can provide shared folders, bulletin boards and intranets but when they are not used (properly) coordination will be worse. Available ICT (fax, mobile phone, p-mail3) are hardly used by the group members to coordinate their work. From this study it can be concluded that those work group members need both guidance in improving their coordination (see previous section) as well as in recognizing and utilizing ICT opportunities (this section). 3

At the time of this study the police force uses a version of pegasus mail, which they often refer to as p-mail. 74

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE 4.2.5 Factors hampering utilization of ICT opportunities The compelling question after the discussion in the two preceding sections is then why these group members do not improve their coordination and make use of the ICT opportunities if the solutions on both issues are so obvious. Below some possible hampering factors are discussed that are mentioned in interviews and discussions. An important comment is that these factors are not necessarily ‘true’ but that they can also be perceptions of the people involved (for instance the comment “technology is not user friendly” can be an objective measure, or a perception of one or more users, or both). The factors that may have hampered coordination improvement and ICT use are: • The position of the regional projects and the regional coordinators in the organization is unclear. Both in the organization surrounding them and within the group of regional coordinators there is doubt whether and how they need to respond to the request from the head of District 6. This doubt makes the focus unclear and the coordination more problematic. • Consequently the outcome of the work group process has a low priority for part of the participants and for the organization surrounding it. This absence of unanimous internal pressure and external pressure to produce something (within a certain time limit) allows the group to coordinate unproductively. • The group members have different backgrounds and views on the problems. Especially in the beginning this makes communication laborious and coordination counterproductive. Group members share the opinion that ICT would rather complicate than support this phase of a process. One of them illustrates this with the remark: “ICT stimulates information production and information spreading, however key to our process was better listening to each other rather than overwhelming each other with even more information”. • The problem at hand is new and unknown. The group is unfamiliar with addressing this kind of problems. The group needs to think creatively. Reaching this attitude costs some time and thus affects the efficiency of coordination negatively. • The group of regional coordinators has developed a ‘meeting-culture’ that can be characterized as ‘chaotic conversation’. It is for instance quite normal to interrupt people with jokes and chairmen are sometimes not accepted and appreciated (their interventions are then publicly interpreted as disturbing the cosy atmosphere). At times meetings can be very smooth and productive but those situations stay unstable. Reasons may be the fact that all coordinators are equal in power and are autonomous senior people, who get annoyed when someone dictates what they should do. ICT is not particularly suited to support these kinds of unstable processes and it is obvious that this ‘meeting-culture’ affects coordination negatively. • The regional coordinators express that in some of their individual conversations with people from District 6 lots of emotions are shared (also by reacting on non-verbal cues). They doubt whether these kinds of clues can be noticed by them in electronic discussions (for instance if the policeman would have put his story on a bulletin board). Also they believe that policemen are more motivated to express their view in a face-to-face conversation (which implicates your interest for their opinion), than in an electronic discussion.

75

COORDINATING DISTRIBUTED WORK • • •

• •

People are not used to using ICT frequently for supporting coordination: ‘unknown makes unloved’. There are some technologies available but these can be characterized as user unfriendly and immature applications. They suit for individual use (word processing, personal archives) but are very laborious for group use. Technology is available on computers in the office. However, the coordinators spend a lot of time traveling through the organization, meeting people on the street and in rooms where no technology is available for their support. Furthermore they spend some time meeting people from outside the police organization. In these circumstances there is no possibility for technology support. Within the police organization technology has a bad reputation. Current applications are out of date. Lots of stories exist where technology is hampering the execution of normal work. People are either enthusiastic or very skeptical. Privacy regulation is mentioned as a problem in daily operational work when someone wants to compile a register of offenders or an overview of youth in the neighborhood.

In addition, some factors that may foster coordination improvement and ICT use have been mentioned: • High importance of a task or project. An often-heard comment is that policemen excel in unique and critical tasks and perform very badly in repetitive standard tasks. • A technology champion can motivate a group to use certain ICT opportunities, as policemen are in general willing to experiment. If this champion has some positive prior experiences chances are quite good that the group will join. • Clear and easy to exploit benefits of a technology. The most striking results of the analysis of this work process in comparison with the other analyzed work process are discussed in the reflection in section 4.4.

4.3 The Regional Project (Fire)Arms The structure of this section resembles that of the previous one. After presenting the work process(es) of the regional project Firearms as a focal unit of analysis, four perspectives on this work process will be discussed. These are respectively the coordination mechanisms, coordination problems, ICT opportunities and the promoting and hampering factors. The exploration of the work process is started with studying some annual work plans and annual reports and subsequently a broad interview with the regional project coordinator Firearms to clarify the total structure of the regional project. After that the researcher joins the coordinator on some of his appointments and attends some meetings. Besides the fact that this illustrates the work practices of the coordinator it also helps in meeting other members of the regional project. In this way the analysis is continued. Key persons are interviewed and subsequently joined when undertaking some of their daily practices that illustrate their role. Views of particular parts of the whole process are discussed with 76

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE different participants to reach a common understanding of a process. Finally the main conclusions considering the effectiveness and efficiency of the whole process are presented on the monthly meeting of the regional project. Most of the qualitative conclusions are confirmed by the outcome of a questionnaire filled in by 88 members from the Regional project Fire-Arms (Bindt, 1998). 4.3.1 Work process Within the Amsterdam Police Force 4 general processes within the field of Firearms can roughly be discerned: tracing arms en firearm-suspects, handling arms and suspects, information analysis and regional knowledge sharing (see Figure 3).

tips reports

tracing arms and fire-arm suspects

arms suspects

information

regional knowledge sharing

handling arms and suspects

arms-form hearing report

report suspect

justice information analysis

Figure 3: The work process firearms A first important comment to prevent misunderstandings is that the regional projects are not ‘a goal directed undertaking with a beginning and an end’ as a project is normally understood to mean. They are more like communities of practice (Wenger, 1998): connecting people on a certain topic for as long as this topic deserves attention within the organization. On the other hand the activities belonging to the regional project are not

77

COORDINATING DISTRIBUTED WORK limited to the fourth process (knowledge sharing). That would not account for all the work the regional coordinator is responsible for. Besides knowledge exchange he also aims at making sure that the different organizational units that are involved in the three other processes work together smoothly and that their efforts are attuned to outside parties like nation wide departments and the administration of justice. We shortly discuss each sub-process below. For a more elaborated discussion and more detailed figures we refer to the case study report (Laere, 1999a). Tracing firearms and firearm-suspects (neighborhood teams, district units, regional criminal investigation department) There are four ways how a Firearm or a suspect who owns arms illegally can be traced. The first is by means of a tip processed by the criminal investigations department. If their information is valuable a neighborhood team or criminal investigation unit can get orders to confiscate the arms or arrest the suspect. Another way is that analysts write a proposal for criminal investigation based on information in police registration systems. After approval of this proposal by management a criminal investigation project is started which may result in confiscation of arms or arrest of a suspect. The third way is that civilians report something considering Firearm possession or use and the police (often the neighborhood team) can react to this. Finally there are the so called “chance hits”. During criminal investigations or observations on the street not primarily aimed at Firearms (for instance research on drug criminality or traffic controls) there could be a detection on Firearms. Handling arms and suspects (neighborhood teams, district units, regional criminal investigation department) After detection the firearms and suspects need to be handled correctly and if possible also quickly. There is a twofold aim. On the one hand information needs to be gathered to make juridical prosecution possible. On the other hand information needs to be gathered to fill police registration systems and feed the analysis process that can improve and initiate future detection of other criminal offenses. Information has to be submitted in a work flow system as well as on some forms. The firearm needs to be investigated from a technical perspective to detect its type (which influences the height of the penalty) and to see if the weapon has been used for earlier crimes. The suspect needs to be heard and his or her living place needs to be searched for more weapons. Based on the facts discovered an imprisonment follows. Information analysis (regional analysts) The aim of this part of the process is the increase of insight in trends as well as actual (future) crimes. This process is conducted from a regional perspective. By comparing records of all confiscations, arrests, and hearings the analyst tries to discover potential suspect groups, places where firearm trade occurs or places where firearms enter the illegal circuit. These discoveries can be the basis for a project proposal for criminal investigation as described earlier.

78

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE Regional knowledge sharing (regional coordinator, regional analysts) This sub-process encompasses the smoothing of coordination between the parties involved, the increasing of professionalism and knowledge sharing in the field of firearms, and the generation of innovative approaches to firearm criminality. Activities undertaken are for instance stimulating police departments to take up “floating information”. These are issued tips or valuable proposals for investigation that are not acted upon because of lacking capacity or priority. This kind of information becomes worthless when it lies too long. One should either act directly or pass it back so somebody else can use it within a few days. Furthermore the regional coordinator and the analysts organize a monthly meeting of professionals from the whole region, a newsletter, a possibility to call the central experts and consult them, work groups that write down procedures or develop innovative ones and meetings with justice and other regional and national external parties. 4.3.2 Coordination mechanisms The four sub-processes are now discussed again with a particular focus on coordination mechanisms. Almost all processes share resources with ‘non-firearm’ processes. Actors that work on the detection, handling and analysis of firearm criminality within their own team, unit or department can also be ordered to work on different topics. In general priority rules are used that state for instance that a firearm professional should dedicate 50% or 30% of his time on this topic. In practice the (lacking) priority of the topic in local policies or daily events may overrule this general rule. Also some professionals prove to be cleverer in informally arranging with their supervisor that they can spend time on their topic. The sequential dependencies between detection, handling and analysis are for the most dealt with by standardization of work in procedures. Some specific tasks (for instance technical investigation of a firearm) need to be done by a firearm- professional who has a course certificate. This can be characterized as a form of coordination of skills. Those procedures guarantee that the right actions are undertaken to transfer firearm and suspect to the next policemen in the workflow accompanied by the needed information. However, supplementary standardization of output (obligatory forms, obligatory deposition of confiscated firearms at central department) and informal adjustment (stimulation by central analysts) is needed to keep the process running. Within the detection process the decision making whether or not to carry out a criminal investigation or preventive projects is done in weekly management meetings. The decision whether or not to take smaller actions and how much time and capacity to spend on it is in the hands of the daily supervisor. Interrelation between different tips, investigations, crime reports or chance-hit-arrests can result from typing the name of the suspect or his or her address in the various police registration systems. Even for the most secret investigations there appears a warning to contact a team supervisor when the involved head suspect is by chance arrested for a minor offense.

79

COORDINATING DISTRIBUTED WORK The mutual dependencies between the regional support and the other processes are handled by all kinds of mutual adjustment (meetings, expert consultation). 4.3.3 Coordination problems Subsequently the coordination problems that occur in the four sub-processes are listed and commented upon. Little horizontal coordination A general conclusion is that there is too little horizontal coordination between the sequential processes which makes interventions from the regional coordinators and the central experts or analysts necessary (as they have the overview and see where it goes wrong). In the opinion of the central coordinator and analysts this is partly caused by the low priority of the topic on the decentralized teams and units, and partly caused by the lack of involvement of some of the local professionals. The members of the regional project blame their few mutual contacts on the limited amount of time they can dedicate to their profession, non-transparency of activities on other places and because the other professionals are difficult to contact. Not surprisingly, the participation in regional activities is also disappointing. Autonomy and non-transparency of decision making of teams and units The fact that almost all members of the regional project firearms can work on their topic only part of their time can eventually result in a situation where nobody is doing anything on the topic or firearms. The topic has a low priority, which leads to a lot of chance hits and few directed actions. Teams and units act autonomously. They try to circumvent the responsibility for potential firearm activities. That behavior may result in “floating information”. During handling of the firearms or suspects they may invest minimally by only addressing the ad hoc juridical process and not the long-term information position. Decision-making is non-transparent. Local members of the regional project seem not to be able to force a breakthrough in these situations. Everybody’s responsibility ends in nobody’s responsibility. Decrease of quality of work delivered The quality of the work delivered in the handling process decreases. This may have different reasons. On one hand the level of professionalism is decreasing because decentralized professionals at the neighborhood teams seldom use their expertise and the centralized experts are decreasing in number. Other reasons may be that the firearm professional is not always consulted in case of chance hits and the earlier mentioned lack of time or priority to handle a firearm process correctly. The central analysts perceive difficulties in contacting professionals when they want to give feedback on poorly handled cases. In addition policemen at the neighborhood teams often question the amount of data they have to fill in on forms or submit into workflow systems. They wonder if those data (which cost a lot of precious time to be produced) are ever used to catch a criminal.

80

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE 4.3.4 ICT opportunities Based upon the work processes and the experienced coordination problems the following ICT opportunities are identified: • A central electronic overview of firearm projects and activities (for instance on intranet). This creates openness of priorities and decision making at the different neighborhood teams and criminal investigation departments. It makes clear what is done and what is not. That may provide insight how assistance from neighboring teams can solve mutual problems. It also can help the central coordinator to show where continuation problems arise and who can solve them. Compulsory openness may increase attention and performance on the topic of firearms (everybody wants to measure up to the others) but attention has to be paid that no window dressing occurs. • A support tool (resembling the well-known internet chat tool ICQ) that shows the availability and accessibility of colleague experts for the regional project firearms. Information on expertise of a certain person combined with information on current or expected presence and contact information (phone, e-mail address) can dramatically ease the process of contacting someone who can assist in problem solving activities. It also eases personal feedback. • Procedures and lessons learned that are currently circulated in handbooks and paper newsletters could be stored electronically on intranet. This eases searching and maintenance. • Like in the youth work group there is a strong preference for synchronous face-toface meetings to coordinate work. Examples are the monthly regional meeting, meetings within various work groups within the regional project, and numerous one to one appointments. Asynchronous discussion on bulletin boards may make coordination in these processes much more efficient. Face-to-face meetings are currently used to ensure attention and participation, which is probably necessary given the low-involved professionals. When professionals are motivated and involved the ‘expensive’ face-to-face meetings can better be reserved for interactive discussions on complex or sensitive matters. Information gathering, communicating announcements and brainstorming can be done electronically in between these meetings. Which ICT opportunities can improve coordination in the Firearm process depends heavily on the interpretation of the problem situation. Opinions differ between the involved actors whether the problems are caused by low priority (not allowed to participate), low involvement (not wanting to participate) or low transparency (not being able to participate). ICT may bring more transparency and may ease contacting of dispersed colleagues but only if priorities and intrinsic motivation leave room for it. ICT is clearly not the solution for the major part of the coordination problem.

81

COORDINATING DISTRIBUTED WORK

4.3.5 Factors hampering utilization of ICT opportunities This section discusses some factors that may explain why coordination problems occur or why ICT opportunities have not been utilized. Factors that affect coordination and ICT use negatively are: • Time pressure and other duties is an often heard reason why activities that benefit others in the firearm chain are not executed. The official reading is that these kinds of activities do not contribute to local objectives. However, on the organizational scale the effect may be counterproductive as this self-interested behavior may cost a lot of rework for others. • In 1998 and 1999 there are very few computers with an intranet connection. The email system is outdated and user-unfriendly. Lack of connectivity and critical mass are therefore heavily impeding the realization of ideas like asynchronous electronic discussion, easily contacting colleagues and continuous insight in firearm activities of other teams. • The police force is caught in a decision-making paradox. Decentralization has given the neighborhood team a lot of autonomy in goal setting and capacity allocation. Moreover the policeman in the street can be characterized as a “street level bureaucrat” who in the end decides what is actually done and what is not. On the other hand external parties (local government, administration of justice) have a strong influence and the old hierarchy sometimes overrules local priorities. The regional coordinator has to maneuver between these central and decentralized decisionmaking processes. When ICT is used to formalize decision making and leaves no room for personal interests either management or street level policemen will reject it. • There is a tendency to support only the static information exchanges and coordination by standardization with ICT (for instance procedures on intranet), while dynamic information exchange and informal coordination can be seen as the pounding heart of the police organization. • There is too little attention for training policemen how to work with new ICT. Technology is introduced without properly integrating it in daily work practices. People have to learn themselves how to cope with the changed situation. • The traditional police culture can be characterized as “reactive, focus on short-term success, lack of involvement, mutual interest and information exchange, uncritically following and difficulty in sharing weaknesses and discussing shortcomings” (Coppes, de Groot and Sheerazi, 1997; Cozijnsen, 1989). This is exactly the opposite attitude compared to the mature attitude needed in a matrix organization. However, in the police organization one can come across policemen with a traditional as well as with a mature attitude. Some state that this old culture may not be an excuse for bad performance. These factors will be discussed in a broader perspective in the next section.

82

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE

4.4 Reflection In the final section of this chapter we reflect on the two case studies and relate the results of their analysis to our research questions. Our first three research questions respectively focus on conceptions of coordination, on factors hampering and fostering ICT adoption and on improvement of coordination (see section 1.7). We start addressing the various elements of coordination present in the case studies (work arrangements, dependencies and coordination mechanisms). Then we discuss the nature of the coordination problems observed. The description of different aspects of coordination starts with illuminating the impact of ICT, as this is one of our major objects of research. The behavioral and context aspects are included in the discussion on factors influencing ICT use. Besides identifying these factors we elaborate on how they interact. Finally we discuss how policemen picture and manage improvement of coordination. We conclude our reflection by translating the main observations in this reflection in a set of requirements for our design approach. Elements of coordination The work arrangements on the organizational and group level (see Table 11, page 45) are highly recognizable in practice. Although we focused on two work processes, we touched upon several groups, organizational units and the larger organizational structure in our analysis. The same holds for dependencies (see Table 12, page 46). The most striking ones are the work flow interdependencies, where several policemen contribute to a service to be delivered to civilians, and the process interdependencies present in the regional projects, where experts in the same expertise field need to share their experiences. These two dependencies are balanced in the matrix organization. Most coordination mechanisms listed in Table 15 (page 50) are clearly present in the two cases studied. On the one hand there are examples of rigid rules and extensive training of policemen, on the other hand there are targets to be achieved, norms to be followed and a large variety of meetings to enable mutual adjustment. During these meetings and in the interviews is observed how policemen construct and reconstruct interrelations. Furthermore, attention for division of work is visible. Formally the organizational structure and the duties of the organizational units determine who needs to carry out what task. Besides, lots of informal mechanisms are present to cope with ad hoc task allocation. For example, in the Youth annoyance group the question to invite additional group members is repeatedly raised. In the Firearm work process the regional coordinator proposes reallocation of tasks that tend to be postponed or canceled due to resource constraints. The diversity of dependencies to be managed and coordination mechanisms present implicates that the police organization is a hybrid organization. It does not rely heavily on one kind of coordination, but combines several centralized and decentralized coordination mechanisms. This complexity of their internal organization is inevitable given the complexity and diversity of their duties and the intertwinement of criminality and society. However, the complexity causes that coordination mechanisms that manage conflicting 83

COORDINATING DISTRIBUTED WORK dependencies may interfere. For instance, rules or work procedures that serve the interest of juridical prosecution or analysis by staff units may harm the interest of civilians that ask for a quick response to their crime reports. Similarly, professionals have to balance time spent on their topic of expertise against their duties for their local neighborhood team. We observe that in practice policemen have (or take) quite some freedom to balance these conflicting demands. As such they choose during execution of their work what coordination mechanism (and underlying dependency) gets most priority. In conclusion, the design of coordination may differ from the actual application of coordination mechanisms in practice. There are clear differences between the organization and its coordination described in documents, explained in interviews, and observed in daily work. Coordination problems At first sight people in the first case mainly struggled with ambiguity problems (diverging views on problem, approach and solutions). To the contrary the people in the second case particularly suffered from uncertainty problems (transparency of who does what and who could assist in tasks waiting for execution). Having a closer look, uncertainty and ambiguity problems are present in both cases. For instance, the Youth Annoyance work group suffers from uncertainty about what policemen in District 6 are undertaking with respect to youth criminality and youth annoyance and some of them are not aware of important incidents in District 6 and of individual activities of colleague group members. Also, in the Firearm work process examples of ambiguity problems are the different interpretations of the function of firearm professional and the different interpretations of work procedures. Although task descriptions and work procedures are standardized for all neighborhood teams, the attention for firearms and the quality of work differ highly from neighborhood team to neighborhood team. Our findings from our literature study that uncertainty problems require more information to be processed, whereas ambiguity problems require time for interaction and interpretation are confirmed. However, in both cases there is a heavy reliance on one communication strategy. In the Youth Annoyance case interactive communication is emphasized, whereas in the Firearm case information processing is stressed. There is not a clear distinction between different improvement strategies to be used for different kinds of problems. Consequently the less prominent coordination problems are not addressed adequately. ICT as a means of developing new coordination mechanisms Malone and Crowston (1994) state “One obvious way of generating new coordination processes is by considering alternative ways of communication”. This implicates that introduction of new ICT may create opportunities to develop new coordination mechanisms.

84

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE A first observation is that the difference between traditional coordination mechanisms and coordination mechanisms for distributed work processes seems rather arbitrary. Coordination of distributed work is a problem of all times, the Greek and Roman Empire already applied coordination mechanisms to coordinate distributed actors and processes. Similarly, work processes in the police organization are already more or less distributed. Furthermore, there is often already some kind of ICT support (telephone, fax, workflow system, databases, e-mail). That means that existing coordination mechanisms are already applied to coordinate distributed work processes in organizations. A second observation is that new ICT changes the artifact rather than the protocol (see Definition 7, page 50). For instance, the forming of a work group of 5 people from a group of 50 can be done by submission of names to a posted message on a shared bulletin board. This coordination mechanism (invitation to all members, reaction visible for everyone) was formerly also possible by sending and returning numerous letters or by asking the question in a meeting where everyone is present. The shared bulletin board does not really change the coordination protocol. It rather makes the execution of the existing protocol potentially (!) more efficient (faster sending of information, continuous interaction instead of once in a period meeting) and more effective (more group influence and control on work group composition, time to consider participation, shared awareness of difficult to start up projects). Similarly presenting work procedures on an intranet instead of in a paper handbook seems more efficient and effective as the procedures are then easier to update and distribute and are also easier to file and access by users. Also here, the coordination protocol of updating and consulting procedures stays the same. Another interesting effect is that the new ICT supported coordination mechanism does not replace the old one, but coexists with it. The intranet is developing, but handbooks are still present. Also coordination mechanisms that were formerly impossible from a cost/benefit perspective (discussion with colleague at the other side of the town) are now a serious alternative (calling or e-mailing this expert). This creates a choice problem whether one should coordinate it the old or the new way. The tendency to stick to old habits may worsen the confusion about the preferred coordination practice. One can thus wonder if the benefits of the new way of coordination are larger than the costs of the created coordination-choice process. Especially in a hybrid organization like the Amsterdam Police Force, the introduction of ICT for coordination purposes can lead to a multiplication of possible ways of coordination. One study (Okamura et al., 1995) reports of an organization addressing this problem by eliminating the old ways of coordinating, but it is questionable whether that is always possible and productive. ICT as a means of solving coordination problems We observe that uncertainty and ambiguity problems coexist in the Amsterdam Police organization. ICT can assist in solving both, but the solutions will differ. For instance, structuring discussions with brainstorm software or on a bulletin board and making explicit common interpretations can solve ambiguity problems. On the other hand providing overviews of tasks or showing available time and expertise of people can help lowering uncertainties.

85

COORDINATING DISTRIBUTED WORK ICT support contributes most obviously to coordination problems that are mainly related to efficiency. Examples are larger transparency, speed up information gathering, and ease information storage and retrieval. ICT support to address ambiguity issues needs more caution, care, patience and thus guidance and instruction for the participants. Providing ICT support for these unstable processes without other managerial interventions would most certainly worsen coordination problems because the main coordination issues are then not addressed. On the other hand, only addressing the managerial side (what actually happened in the first case) ends up in feelings of frustration about the inefficiency of the process. Combining both (managerial as well as ICT) seems most promising. ICT is not restricted to coordination between distributed actors or distributed work processes. Co-located groups can benefit from increased speed, storage and structure aspects of ICT. On the other hand, ICT cannot solve all coordination problems in distributed work. In both cases studied, the coordination problems are for some part not related to the dispersion of the participants. Poorly structured meetings, poor leadership, low priorities, lack of involvement and diverging views are not a result of the distribution of work processes but are caused by organizational circumstances or personal choices. Such coordination problems ask for managerial interventions addressing focus and effectiveness that are not directly ICT related. A final observation is that ICT can increase problems rather than solve them. When the ease of information exchange and storage results in gathering and storing large quantities of information that are not directly relevant, but may be ‘nice to know’, this superfluous information gathering may end up in information overload problems. Similarly speeding up a work process where coordination problems are ambiguity oriented rather than uncertainty oriented may harm rather than support the work process. Furthermore, the successful appliance of ICT supported coordination mechanisms depends on their interaction with numerous hampering factors, which we discuss hereafter. Factors hampering and fostering ICT adoption Besides illustrating the effects of ICT on coordination, the cases reveal a wide variety of hampering factors on a social, organizational and technological level. For the presentation of these factors we use the framework of Adaptive Structuration Theory (DeSanctis and Poole, 1994). Comparison with other field studies will be easier this way. AST argues that advanced information technologies trigger adaptive structuration processes that can lead to changes in the rules and resources that organizations use in social interaction. The impacts of technology on the group depend on: • the structural potential of the technology (its spirit and structural features); • how technology and other structures (such as work tasks, the group’s internal system, and the larger organisational environment) are appropriated by group members; • and what new social structures are formed. Appropriations, which initially occur in micro-level interaction, eventually may be reproduced to bring about adoption of technology-based structures across multiple settings, groups and organizations.

86

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE We distinguish between task related (Table 20), technology related (Table 21), group internal system related (Table 22) and organization related (Table 23) factors. The factors themselves are underlined. Then a short description follows. Finally the arrows point at the main effects each factor has on the organization. Table 20: Task related factors Task related factors Growing complexity of tasks: Besides hunting down criminals and handling reports by standard procedure there is a growing investment in preventive projects and community policing. Work that primarily focuses on local neighborhoods, suspect groups or criminal offenses may eventually interfere. The general policeman has to cope with this growing complexity and interrelation. However, he cannot know all procedures and best practices on any topic, nor solve all problems by himself. ⇒ Use of dispersed expertise is necessary to improve quality and innovation ⇒ The growing interrelation of problems and tasks makes one’s own added value unclear Fast response of general policeman to problems: A general policeman can be confronted with almost any kind of problem while on duty. Often (especially on the street) there is need for direct action and no opportunity to wait until the local expert is on duty or at the site. ⇒ Expert or procedures need to be consulted fast and at distance

Table 21: Technology related factors Technology related factors

Connectivity: Intranet is currently not yet available on every workstation4. Nor do policemen always have access to a computer (e.g. in the street). ⇒ Without connectivity computer supported consultation of dispersed expertise is impossible Complexity of interface: A lot of the police applications have a complex interface. The main workflow system and the e-mail system are regarded as difficult to use. ⇒ A complex interface hampers computer supported consultation of dispersed expertise Lack of critical mass: Groupware applications like intranet, e-mail, newsgroups and bulletin boards are only successful when most of the users are actively participating. Heavy promotion and instruction is often lacking when new information systems are introduced in the police organization. That may delay the adoption of these systems. A low percentage of early users lowers the experienced added value and puts off new users. ⇒ A low percentage of early users delays the adoption of groupware Focus on supporting formal coordination: In development there is currently a strong focus on systems that support formal coordination (procedures on intranet). In practice coordination proves to rely heavily on informal coordination (conversations, meetings). Developers should pay more attention to functionality of e-mail and discussion supporting systems. These are needed to interactively discuss complex matters that cannot be clarified on an intranet. ⇒ Neglecting informal coordination options makes computer supported consultation of dispersed expertise difficult and an incomplete alternative

4

By the end of 1998 there were about 130 PC’s with an intranet connection on 5500 users. During 1999 and 2000 all PC’s were connected to the intranet server. 87

COORDINATING DISTRIBUTED WORK Table 22: Group internal system related factors Group internal system related factors Reachability: There is little or no insight in where which expertise can be found. Also people work on different times due to the 24-hour service. That makes experts that could assist in solving problems difficult to locate, and thus difficult to contact quickly. ⇒ Without an overview of available expertise locating and contacting takes too much effort Consultation is not embedded in daily work: Neighborhood teams do marginally promote consultation and do a limited evaluation of quality and innovation. Only during major calamities experts are invited on the scene. Consequently there is little demand for expertise of the regional projects and little pressure for them to increase their contribution. ⇒ There is no experienced need nor pressure to make use of dispersed expertise Lacking commitment in availability of experts: The local neighborhood team manager determines how much time a professional can spend on his topic. That makes it hard for professionals to commit themselves to tasks to be carried out in the regional project. This is reflected in a noncommittal attitude: “I would like to join, but if priorities at my neighborhood team change I withdraw”. ⇒ The noncommittal attitude of local experts undermines the continuity of work for the regional project

Table 23: Organization related factors Organization related factors Decision making paradox: On the one hand the police force is highly decentralized. A neighborhood team has a lot of freedom to define their own policy and priorities (Terpstra, 2002). On the other hand the ‘old hierarchy’ still exists and pressure from (local) government, justice or the general police management can ask for a reverse in policy (Oosterhout, 2002). For professionals and project leaders such unexpected shifts can be frustrating if they are not justified. ⇒ Constant struggle about priorities decreases effectiveness and efficiency Results are difficult to perceive: The direct effect of police work on societal problems is often very vague. Did criminality decrease due to our surveillance and arrests or because of other reasons? Therefore evaluation of the effectiveness of work and its improvement is a hard job. ⇒ Hard to perceive results complicate evaluation and improvement of work Strong confidence in own knowledge and expertise: A policeman is by definition self-conceited (Coppes et al., 1997; Cozijnsen, 1989). He is selected to make quick decisions and act singleminded (which is important in crisis situations). Doubt or a change of strategy can have disastrous effects. Furthermore he is judged on individual performance. Mistakes are the responsibility of the policeman who made them; the expert giving a wrong advice is not blamed. Within the new (preventive) tasks, consulting of colleagues or handbooks is recommendable. A policeman in an ‘individual assessment culture’ often perceives this as unnatural behavior. ⇒ Consulting is unnatural behavior for a policeman

Limited ICT skills: Most policemen are lacking experience in handling ICT applications like e-mail and intranet. This implicates that they cannot exploit the possibilities of these technologies to the fullest. ⇒ Limited ICT skills inhibit the productive use of new technologies

88

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE Interaction of influencing factors The overview of influencing factors clearly shows the wide variety of technological, social, organizational and personal factors that interfere and make the outcome of distributed work processes or their improvement unpredictable. It confirms the view from adaptive structuration theory that technology and other structures mutually influence each other and are appropriated by organizational members. Here, we dwell upon some outcomes of the interactions between these factors that we observe in the Amsterdam Police Force. An often-heard complaint is that the impact of ICT is disappointing because new technologies are not used as expected (no contribution to bulletin board, not consulting latest version of intranet page). However, the problem of ‘improper application’ also holds for the traditional coordination mechanisms (silence during meeting, no response to letters, not consulting the paper handbook). In practice the ICT application is often blamed for the failing coordination when actually the people who do not follow the protocol should be blamed. When distributed work processes are not functioning as expected, one should not instantly blame the technology, but first look at the people working in that process or the process designers (possibly the same people). In a well functioning process the people involved would change the technology or search for alternative ways of coordination. Usual organizational practice, historical habits and personal beliefs about ICT value have a major impact on current coordination decisions and behaviors. Although ICT opportunities are sometimes quite obvious, people have a tendency to stick to their usual way of working. Face-to-face adjustment is preferred above computer mediated coordination just because one is used to it or because this is the way it is usually done in the organization. Furthermore expectations about how others will use the technology may limit its value. Computer mediated communication is associated with issues of minor importance. Face-to-face communication is associated with respect for one another and importance of the subject under discussion. It may be that not the technology itself (email) but the way it is used or believed to be used (“e-mails are not important so I read them rarely or just throw them away when the sender is unknown”) is determining the value of an ICT application. The comparison of the first case (work group Youth annoyance) with two other problem solving work groups stresses the conclusion that outcomes are unpredictable and depend on a wide variety of factors. Take for instance a work group that has to develop and document a new procedure for handling youth criminals. They operate in the same organizational context and with the same technological support systems. One of the members explains that they make heavy use of e-mail and phone between their once in a while face-to-face meetings. Also they are very satisfied about their collaboration process, the quality of their end product and their throughput time. Possible explanations for the differences with our case are clarity of the assignment, ICT skills of the members and their historical habits considering ICT usage.

89

COORDINATING DISTRIBUTED WORK A second example is a workshop organized in the same period of time with the regional coordinators aimed at developing all kinds of approaches to annoyance in the Amsterdam city center. During this workshop the regional coordinators make use of a Group Decision Room, a computer support system that enables electronic brainstorming, electronic structuring of ideas, and electronic voting. A first impression suggests that the process is much more productive than our case. During one day the group of 20 coordinators produce about 68 pages of rough information that is later condensed in 25 pages of valuable problem descriptions and ideas. Including a one-day preparation and a few days work in a small group to finalize the report afterwards, the whole process is carried out within two weeks. However, a first comment is that the key idea how to deal with the deteriorating annoyance (a flexible surveillance team on the street conducting a low tolerance policy) is developed after the Group Decision Room session in the small group of people that processes the meeting outcome. Secondly, one can question the amount of integration between the ideas of the 20 coordinators during the session. The majority of the contributions can be considered as a summing up of solutions from their own professional perspective instead of an integral discussion. The final comparison also makes clear how difficult it is to determine the effectiveness and efficiency of these two processes. Our youth case is productive from a team building perspective and a waste of energy from the client’s perspective. In the city center process there is a quick and clear advice, but its quality as well as its contribution to the team building process is questionable. Improvement of coordination When policemen ponder on examples of ‘good coordination’ often heard quotes are: “Reaching the objectives with minimal efforts”; “being on the same wavelength”; “working like a well-running machine”; and “not working at cross-purposes”. The underlying thoughts that can be derived are those of effectiveness and efficiency. Effectiveness is reflected in specifying common goals or aligning conflicting goals. Efficiency is articulated in reducing the amount of time and effort for communication and adjustment. However, we observe a tendency in the police organization to focus on uncertainty and efficiency issues when trying to improve coordination. Often the discussion is only about improving the currently used coordination mechanism (doing current things better). The relation with the underlying dependency that needs to be managed is seldom discussed. So the question whether choosing another coordination mechanism or another division of labor (doing different things) could produce much better results, is seldom asked. Earlier in this reflection we observe that the design of labor division and coordination in the police force is not strict and fixed, but quite open. Policemen have considerable room to alter labor division and coordination to adapt to situational circumstances. This decentralization design of labor division and coordination is understandable given the complexity of their organizational environment. This decentralization implies that improvement of coordination is not anymore a matter of centrally analyzing duties and demands and selecting a more appropriate fixed set of work arrangements and 90

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE coordination mechanisms to be implemented and executed. Instead improving coordination means supporting organizational members in making good coordination choices: choosing an ad hoc work arrangement and ad hoc coordination mechanisms from a set of available ones, guided by the objectives that need to be achieved and dependencies that need to be managed. From the above quotes of policemen we conclude that these coordination choices are based on enhancing efficiency and effectiveness: on comparing potential costs and benefits of coordination choices. The comparison of cases above confirms the proposition of Malone and Crowston (1994) that the extent to which an organization has improved its coordination cannot be answered directly, but depends on the process goals. The question whether for instance the Youth process is productive depends on the choice whether it is evaluated from the client’s perspective or from the perspective of the team building process. Moreover our cases illustrate the complexity of measuring the effectiveness or ambiguity side of coordination. Measuring the amount of time and money invested in two processes (efficiency) can be done. Measuring or comparing the quality of both advises is more difficult, and measuring the amount of team building objectively seems even worse. Value judgments considering these issues may be highly subjective. Also some goals may not be apparent at the beginning of the process, which implicates that evaluation criteria are determined or changed during the work process which could eventually end up in a kind of self fulfilling prophecy. The latter two points suggest that comparing the potential costs and benefits to enable good coordination choices is highly problematic. Consequently, we do not expect that accurate calculations of costs and benefits are feasible and desired to support coordination choices. What we observe in current practice is that policemen quickly estimate potential costs and benefits. However, we expect that coordination choices can improve when organizational members have more insight in the total picture of duties, environmental demands, dependencies, available coordination mechanisms and influencing factors. We argue that they currently have a limited view on for instance the complexity of the organization, the value of ICT supported coordination mechanisms, and the way how to apply these ICT supported coordination mechanisms successfully. Conclusions and requirements for our design approach First, we shortly summarize the main observations of the reflection. Then we conclude with some requirements for our design approach. We observe that the distinction between work arrangements, dependencies and coordination mechanisms is recognizable. However, due to the multitude of dependencies and coordination mechanisms present in the police organization, there is not a clear and fixed coordination mechanism for each dependency to be managed. Instead of a fixed design that is executed, we observe that policemen have considerable freedom to emphasize particular coordination mechanisms more than other ones. These so-called coordination choices are needed to deal with the complexity and dynamism of the organizational environment. By making different coordination choices policemen can 91

COORDINATING DISTRIBUTED WORK adapt the coordination of their work to situational circumstances. In addition, we observe that uncertainty and ambiguity problems coexist in practice and require joint solutions. The impact of ICT is diverse. ICT mainly changes the coordination artifact, rather than the coordination protocol. ICT supported coordination mechanisms coexist with their traditional counterparts (work procedure in handbook and on intranet), or make formerly unrealistic coordination mechanisms feasible (visiting or mailing expert at the other side of town). As such, ICT results in an even larger set of coordination mechanisms to choose from and consequently it complicates the coordination choices to be made. ICT is particularly useful to solve uncertainty problems, but can also contribute to solving ambiguity problems. Furthermore ICT can be useful for distributed as well as co-located groups. In all cases unbalanced use of ICT can also lead to counterproductive effects. Both for traditional and ICT supported coordination mechanisms holds that a multitude of interfering technological, social, organizational and personal factors can hamper the successful appliance of the coordination mechanism, which worsens rather than improves coordination. Finally the observed coordination choices prove to be the main starting point for improvement of coordination. Matching currently emphasized coordination mechanisms to situational circumstances and to the currently prevalent dependencies is more useful than trying to define one fixed coordination mechanism to be executed. In practice coordination choices are based on a comparison of the potential costs and benefits of different coordination mechanisms. These costs and benefits cannot accurately be determined, but are roughly estimated. They refer to the match with the dependency to be managed and situational circumstances and to the expected difficulties of applying the coordination mechanism. Given these observations, we argue that our design approach should pay attention to the following requirements: 1. The conception of coordination should fit with the well-known distinction between work arrangements, dependencies and coordination mechanisms that are observable on the group and organizational level. 2. In addition, ‘coordination choices’ should be a central issue in the conception of coordination. This resembles coordination practice in large service organizations and it offers possibilities for improvement of coordination: better coordination choices result in improved coordination. 3. The conception of coordination should pay attention to uncertainty problems (stressed in rational views on coordination) and ambiguity problems (emphasized in social views on coordination). 4. Our design approach should not only enable identification of ICT opportunities, but it should also explain how these opportunities can be seized successfully, thereby addressing the interaction of a multitude of technological, social, organizational and personal factors that influence ICT adoption. 5. Although uncertainty problems are easier to measure and to solve, it is important that our design approach gives clues how to jointly solve uncertainty and ambiguity related coordination problems. 92

CHAPTER 4: COORDINATION AT THE AMSTERDAM POLICE FORCE 6. 7.

Also, our design approach should point out that coordination can be improved by changing or improving coordination mechanisms as well as by changing or improving the underlying work arrangements and dependencies. To improve coordination choices the design approach should give insight in the potential cost and benefits involved in the selection of a certain coordination mechanism from a set of coordination mechanisms. These costs and benefits can relate to the suitability of a certain coordination mechanism to manage a certain dependency and they can relate to the costs and benefits involved is successfully applying the coordination mechanism (e.g. hampering technological, social, organizational and personal factors).

These requirements are addressed in chapter 5 and 6 where we respectively develop our conception of coordination and our design approach.

93

5 An integrated view on coordination of distributed work In this chapter we start the prescriptive conceptual phase of our research by constructing our own view on coordination. We fuse the observations and reflections from our conducted case studies (chapter 4) with the useful elements of the theories on coordination (chapter 3). The resulting framework for coordination of distributed work is the foundation for our design approach that is presented in chapter 6. To construct our view on coordination we elaborate on the concept of ‘coordination choices’. We answer questions like what are the main choices to be made, which costs and benefits are most prominent in making these choices and what is the impact of ICT on these costs, benefits and coordination choices? In section 5.1 we discern between four kinds of coordination choices. Each of them is discussed in detail in sections 5.2 until 5.5. There we describe the wide variety of factors that have an impact on the different coordination choices. We particularly focus on the benefits, costs, opportunities and challenges of ICT supported coordination mechanisms in comparison to traditional coordination mechanisms. Finally, in section 5.6 we discuss the use of the coordination framework as a whole.

5.1 Coordination choices From chapter 3 we recall that work has to be divided amongst actors which results in work arrangements. Work arrangements create dependencies between tasks or actors. Coordination mechanisms are ways to handle these dependencies in order to attain the objectives of the organization and its members. In our case studies we observe that: • work arrangements and accompanying coordination mechanisms to manage the dependencies created exist on the organizational and process level; • work arrangements and coordination mechanisms change continuously; • numerous work arrangements and coordination mechanisms coexist and certain groups of actors may use different ones than others for comparable tasks in the same organizational setting; 95

COORDINATING DISTRIBUTED WORK • •

mediating factors, related to work, technology, social setting, personal character or contextual circumstances influence the successful application of coordination mechanisms; ICT enables that more alternative coordination mechanisms become feasible for a certain situation, but may also have negative side effects.

The police organization has to cope with a wide variety of demands from the environment. Organizational design choices on the organizational level set some bounds for coordination choices on the process level, but multiple options are left open to make sure that the organization can react flexibly on the environmental demands. In a service organization like the police force more options can be left open than in a production organization, where constraints of physical machinery are much more prominent. To meet continuous change and dynamism the selection of work arrangements and coordination mechanisms needs to be conducted faster and more frequently. Increased flexibility results in more complicated coordination. As Kastelein (1990) already observed lowering the performance standard (less flexibility) is not really an option. Therefore we need to focus on improving the situated selection of a work arrangement and related coordination mechanisms. We argue that the choice to emphasize certain coordination mechanisms in a particular situation is not solely based on the dependency to be managed. First, there may be more than one dependency to be managed concurrently, which may lead to multiple, possibly conflicting coordination mechanisms. Secondly, the situated coordination choice is heavily affected by earlier choices on the organizational and group level. For instance, the invitation of a certain group member (process level, group composition) may depend on his earlier experience from other group processes (process level, group composition) or on training (organizational level, coordination of skills). Thirdly, the coordination choice may be changed to anticipate hampering technological, social, organizational and personal factors. We conclude that it is a too limited approach to identify possible coordination choices by solely looking at dependencies to be managed. Instead we turn to coordination problems. We assume that coordination problems can be related to coordination choices that were not made or that were made inefficiently or ineffectively (Hettinga, 2002). In accordance with our inductive research strategy we use our case observations to discover some central issues. By analyzing our case study results and by additional discussions with members from the police organization the observed coordination problems are grouped in four distinctive categories (for examples of problems, see Table 25): 1. problems related to which tasks have to be performed by whom; 2. problems related to motivating and directing organizational members that have interrelated tasks to work together towards a common objective; 3. problems related to the amount (how often) and character (how intense) of the adjustment between the involved organizational members; 4. problems related to the topic and kind of information they need to exchange. These four categories can be related to each other as is shown in Table 24. 96

CHAPTER 5: AN INTEGRATED VIEW ON COORDINATION OF DISTRIBUTED WORK Table 24: Basic coordination choices Quantity

Quality

Relation

Which actors and other tasks need to contribute to the task at hand?

Accommodation

How often and intense does information need to be exchanged between tasks and actors?

How important is this relation, how much commitment is needed, or how much effort to maintain this relation is justified? What information needs to be exchanged, and should it be to the point or extensive?

We argue that our case study material confirms that the basic elements of coordination are creating relations (concretized in work arrangements) and accommodating these relations. Problems can occur considering: • the quantity of relations and accommodations; • the quality of relations and accommodations. Just like coordination problems, coordination mechanisms can be related to these four categories of coordination choices. The overview in Table 25 reveals that in order to prevent or solve these different coordination problems appropriate coordination choices need to be made for each category. Coordination mechanisms that are grouped together can be seen as different ways of addressing a similar coordination choice. Table 25: Grouping of coordination problems and coordination mechanisms Coordination problems CONSIDERING QUANTITY OF RELATIONS Coincidental selection of assignment, lacking involvement of policemen from District 6, firearm professional is not always consulted

Coordination mechanisms Division of work, composition, task allocation, standardization of skills

CONSIDERING QUALITY OF RELATIONS Diverging views on assignment, lack of contact with principal, noncommittal attitude, team-building needed, long throughput time, autonomy of decision making of teams and units, lack of involvement, lack of commitment, lack of priority

Direct supervision, leadership, standardization of norms, interpersonal attraction, attention, agreement on rules, standardization of output, decision making,

CONSIDERING QUANTITY OF ACCOMMODATIONS Individual actions are unrelated to group process, reliance on group meetings as one and only communication mode, little horizontal coordination, non-transparency of decision making of teams and units, floating information

Mutual adjustment, communication, planning, action, interaction, feedback, anticipatory-, adaptive- and review coordination

CONSIDERING QUALITY OF ACCOMMODATIONS Differences in backgrounds, perspectives, views complicates communication, poorly structured meetings, decrease of quality of work, questioning the amount of data to be filled in on forms or to be submitted into workflow systems

Standardization of work, rules, labels, interrelating

97

COORDINATING DISTRIBUTED WORK We already observed in chapter 3 and 4 and in the beginning of this section that coordination choices are taken on the organizational and group level, and that choices on the organizational level set the bounds for choices on the group level. Therefore we discern between coordination choices for each of the four categories (see Table 26 and Figure 4). Managers or organizational designers make choices on the organizational level and employees make choices on the group level during execution of daily work. Table 26: Coordination choices at group and organizational level Category / dimension QUANTITY OF RELATIONS QUALITY OF RELATIONS QUANTITY OF ACCOMMODATIONS QUALITY OF ACCOMMODATIONS

Group level Composition Commitment Contact Content

Organizational level Competence Cohesiveness Connection Context

Group Work

98

C on te xt

ct ta on C

n io ct ne on C

Figure 4: Framework for coordination

result

C on te nt

assignment

t en itm m om C

C om po si tio n

n io es oh C

C om pe te nc e

Work execution produces results for a certain assignment and requires coordination choices on the group level on 4 dimensions, which are embedded in 4 related coordination choices on the organizational level on each dimension.

CHAPTER 5: AN INTEGRATED VIEW ON COORDINATION OF DISTRIBUTED WORK Each task to be performed on the group level involves four different kinds of coordination choices. Depending on the topic and nature of the job relevant people are assigned to perform that job (composition). To assure that they will contribute their participation has to be motivated or enforced (commitment). Furthermore the group of actors involved need to determine how often and how intense they need to exchange information (contact) and they need to deal with the question what information needs to be exchanged (content). Coordination choices on the organizational level ease the choice processes on the group level as well as limit them. For instance, recruitment, employment and training of organizational members assure that certain skills are available for composition of work groups and execution of work processes (competence). But it makes involvement of group members from outside the organization less attractive as relatively much effort is needed to locate and invite them. Secondly, mutual goals and a strong sense of shared mission and values within an organization may ease to arrive at a group of colleagues that are committed to a process or task (cohesiveness). Similarly the availability of multiple potential media (like regularly scheduled meetings, phones, intranet, e-mail) make it easier to contact a person when one feels the urge to do so (connection). Finally a clear picture of the organization's environment and the duty of the organization makes it easier to distinguish relevant information from redundant noise (context). To gain better understanding how choices on these various aspects are made more elaborated discussions on these coordination choices can be found in sections 5.2 until 5.5. Each section addresses one of the four coordination choice dimensions on the group level and the related dimension on the organizational level. The costs and benefits to be taken into account in each choice as well as the impact of ICT on these costs and benefits are discussed. Both theoretical contributions (chapter 3) as well as insights from the case studies (chapter 4) are used to clarify and illustrate these descriptions.

5.2 Composition and competence Whether it is on the drawing table articulated in standardization or planning, or in ad hoc adjustment during execution of the work, each work process involves choices about who needs to contribute. Although we stated several times that these choices are actually belonging to division of work, we also observed that they have a major impact on overall coordination effectiveness and efficiency. Therefore they are included in our view on coordination. The coordination choices that touch composition and competence are shown in Table 27. Table 27: Coordination choices regarding composition and competence Composition Determining skills needed for task at hand Selecting appropriate employees for that task

Competence Determining key competencies Training employees Recording experience

99

COORDINATING DISTRIBUTED WORK Composition basically means matching job characteristics and employee competencies. It aims at selecting the best employee(s) to do the job while spending minimal effort on this matching process. When the matching process is standardized (specific employees always carry out specific kinds of jobs) the costs of this coordination choice process decrease. However, if the characteristics of the work to be handled are highly variable or unpredictable more effort is needed to guarantee a flexible deployment of the workforce in order to obtain an appropriate fit between work characteristics and employee competencies. Otherwise standardized deployment rules may harm quality or speed of work, as the employees involved are not capable of executing the job properly. On the organizational level competence issues involve guaranteeing that certain often needed competencies are available. This implicates that key competencies need to be determined and that employees need to be trained in these competencies. Also providing for records which employee has what knowledge eases selection of employees for work processes. When a job on the group level requires skills that are within the scope of these competencies, the efficiency of the matching process is increased due to the efforts on the organizational level. However, the matching process should not turn into pigeon holing: framing the problem in such a way that it only requires available expertise (Mintzberg, 1983). Expertise outside the scope of the known competencies has to be organized whenever needed. The following costs and benefits have to be balanced in the competence and composition choices (see Table 28 below). Table 28: Cost and benefits regarding composition and competence Costs Time/effort for training and recording Time/effort for analyzing and matching

Benefits Faster matching of often needed competence Better performance (quality/throughput time)

In Table 28 it is reflected that spending more time and effort on training and recording on the organizational level leads to faster matching of often needed competence on the group level. Also spending more time and effort on analyzing jobs and available employees and matching them results in better performance regarding quality and/or throughput time. However, if the extra time and effort exceed the benefits, overall productivity is not rising at all. Now we have insight in what costs and benefits are involved in coordination choices regarding composition and competence, a next step is to ascertain what distinguishes a distributed setting from a contemporary setting on these coordination aspects. For this purpose we summarize the opportunities that ICT offers to improve coordination considering composition and competence. Again we draw upon our observations during our case studies. In contrast we discuss how hampering factors may undo these advantages. The main opportunities and threats are first presented in Table 29 and Table 30 and subsequently elucidated. 100

CHAPTER 5: AN INTEGRATED VIEW ON COORDINATION OF DISTRIBUTED WORK Table 29: Opportunities and threats regarding composition Opportunities Potentially more employees available to contribute to a certain process Making maximum use of experts More equal distribution of information intensive work

Threats Selection and matching costs too much effort Dependence on technology Experts may be disturbed too often Too much fragmentation of jobs makes one’s own added value unclear

The key advantage with regard to competence and composition is flexibility. Work and the employment of organizational members become potentially independent of time and place. Involvement of various dispersed actors becomes possible without losing time for traveling. However, a first snake in the grass is the word ‘potentially’. When organizational members do not consciously seek for chances to involve distributed actors or to contribute to distributed work processes none of those potential advantages may be experienced. Also, accommodating the actors involved to the task at hand costs time and effort and sub optimal choices may be made. Furthermore the independence of time and place turns into a dependence on technology. Technology breakdowns can be disastrous for productivity and employees may not like the idea that they cannot perform their tasks without technology. The general advantage of flexible employment has two sides. Concerning highly specialist work scarce experts can more easily be involved. To be approachable they do not have to be grouped at a central functional department, nor does their expertise need to be duplicated at every geographical site. Instead they can work decentralized and function as a (virtual) regional expertise center at the same time. When their expertise is not directly needed they can contribute to daily work in the primary process, instead of working on their own expertise while customers and front office employees have different needs. Whenever and wherever their expertise is needed they can be consulted or asked to participate. Depending on priorities they can interrupt their work and help instantly, or first finish their task and then contribute. In this way the specific expertise of the professionals is used when needed, but when priorities change and there is no work within their discipline they can automatically take up one of the existent assignments in their geographical business unit. Also costs of duplicating expertise and traveling are minimized. For more standardized information intensive work goes that it can easily be shifted to other employees. This can be worthwhile when peaks occur in one geographic site, while there is a surplus of capacity somewhere else. An issue of concern may be the disturbance of experts. Continuous consultation may distract them too much from their own work. Also too much fragmentation of the work may lead to delineation (Kastelein, 1990) as one does not recognize one's own contribution to the larger whole. Table 30: Opportunities and threats regarding competence Opportunities One central record of employee skills Training on the job

Threats Compiling and updating costs too much effort Training not taken because of low priority 101

COORDINATING DISTRIBUTED WORK To enable the unlimited involvement of employees in whatever work process they are needed there ought to be a clear and easy to consult overview of employee’s skills. Also here ICT offers an obvious opportunity. Centrally filled and locally consulted database systems can support the matching process in the desired way. However, finding the appropriate key words to link job characteristics and employee skills as well as updating the records after each new experience (training or practical) of an employee may cost lots of effort. Secondly, computer based education offers possibilities to learn during work time at the work site. In this way training does not directly involve a period of absence at the work, but makes the training itself independent of time and place. In the ideal situation both the trainee and the work can take advantage, as the training is more directed at individual learning needs and work related issues, and can be done on moments when work pressure is low. However, when work is always urgent and busy, and when the employee does not get room to learn, the flexibility may end in less training.

5.3 Commitment and cohesiveness In the previous aspect the focus has been on selecting who should contribute to what process. Commitment and cohesiveness deal with ensuring that the participation of the selected employees in the process is enduring and with giving direction to their efforts and contributions. The following coordination choices can be related to these aspects of coordination (see Table 31). Table 31: Coordination choices regarding commitment and cohesiveness Commitment Determine process goals Specify employees contributions Build social affection Visualize progress Evaluate process and results

Cohesiveness Disseminate organizational mission Determine organizational priorities Build cooperative culture Control performance

Commitment during process execution is an important element of coordination as it prevents withdrawing of crucial group members and for long throughput times. Here the benefits of a smoothly running process should exceed the efforts needed to gain commitment. The most obvious ways to arrive at a committed group of employees to carry out a work process or a group assignment is either to specify mutual process goals or to build social affection among them. Also sharing expectations about contributions of each other helps to build trust about each other’s future behavior in the group. Visualizing progress during execution of the work process may reinforce this trust as expectations are seen to come true. Furthermore evaluating the group process and its outcomes helps to build trust for future collaboration.

102

CHAPTER 5: AN INTEGRATED VIEW ON COORDINATION OF DISTRIBUTED WORK Gaining commitment on the group level can be smoothened by a strong sense of cohesiveness on the organizational level. This involves initiatives like developing a strongly supported mission or building a cooperative culture. In that case experts or employees are willing to assist others, although they also need to address the local duties of their business unit. Preferably priority setting and performance evaluation reinforce the mission and culture. Guiding the workforce to what is desired behavior and what is not eases struggling about directions and desired outcomes of processes on the group level. Table 32 shows costs and benefits in commitment and cohesiveness choices. Table 32: Cost and benefits regarding commitment and cohesiveness Costs Time and effort for determining mission, priorities and goals Limits freedom and creativity Time and effort to socialize Time and effort to show and interpret progress Time and effort to evaluate

Benefits Smooth process (high % productive hours) Faster throughput time Focus Willingness to help Trust Shared responsibility Trust and willingness to help in future Improved performance in future

Just like the previous aspects receiving the benefits has to be weighed up against the effort of getting them. This goes for instance for achieving a smooth process, a fast throughput time, willingness to help, trust and shared responsibility. Furthermore strong commitment and cohesiveness leads to focus, but on the other hand it limits creativity and freedom. It depends on the circumstances whether creativity and freedom or focus and fast performance are more valuable. When comparing building commitment and cohesiveness in contemporary work and in ICT supported distributed work, the opportunities and threats shown in Table 33 and Table 34 can be discerned. Table 33: Opportunities and threats regarding commitment Opportunities More participants in decision making Quick adjustment of changes Open decision making Decisions and goals recorded univocally Self management Monitoring and coaching Specialized relations

Threats Less speed due to a large number of participants Difficult to converge Members of multiple processes have conflicting goals Less freedom for interpretation of goals Growing complexity Big brother Easiness of dropping out 103

COORDINATING DISTRIBUTED WORK Computer supported decision making gives opportunities to involve more employees in decision making, which increases employee motivation and support for the decisions. Also they can more quickly react on changes, as they do not need to wait for the next meeting to bring up the issue. But these opportunities can also turn into threats. A large group of participants could slow down decision making and computer supported interaction proves to be less suited to converge in discussions. Besides involving more employees in decision making, ICT gives opportunities to share the progress and results of this process with everyone. Just like involvement, openness of the decision-making processes can stimulate motivation and commitment. Moreover univocal recording of for instance the process goals increases clarity about the priorities and reduces the chance of conflicts. However, when employees from different departments become involved in work processes, or when employees participate in multiple processes at the same time, the chance of conflicting priorities rises. Furthermore, explicit recording of goals and priorities leaves less room for free interpretation and improvisation. This can thus stimulate the arising of conflicts. Considering control ICT can also have different effects. On the one hand it gives the possibility of autonomy and self-managed work (teams) as information needed for decision-making can be provided for locally. On the other hand there are more than ever opportunities for centrally monitoring work processes and coaching when computer mediated coordination and communication processes make work execution more open and visible for the manager. Both opportunities could work out negative when selfmanaging employees lose the overview due to the increased complexity of ‘spaghettilike’ organizations, or when employees interpret the well-intended interference and support of managers as big-brother supervision. Finally the opportunity to establish brief and specialized relations (which is prominent in the discussion on competence) means that motivation relies heavily on the professional responsibility of the involved employee to deliver his contribution. However, when this relation is only a matter of professional performance in this process, and there are no other (historical, future, social, or work related) dependencies, it may be easy to drop out of the process if there are too many conflicting priorities. Especially in distributed settings it is less embarrassing to just stop contributing and neglect calls. Similar opportunities and threats can be recognized on the organizational level for cohesiveness. Table 34: Opportunities and threats regarding cohesiveness Opportunities Central record of mission and identity Anonymous interaction Explicit rewarding of results

104

Threats Absence of strong ties Inability to socialize Formalization Escalation of emotional conflicts

CHAPTER 5: AN INTEGRATED VIEW ON COORDINATION OF DISTRIBUTED WORK Although it may be less persuading than a penetrating speech, the advantage of diffusing a vision, mission and identity via for example intranet is that it can be consulted continuously and that employees are exposed to it repeatedly. Think of the top priorities blinking on top of each intranet page. In addition, ICT supported interaction can be organized as anonymous interaction, which could give employees room to share concerns or ask question they might not dare to discuss in face-to-face settings. There is, however, a lot of skepticism about the ability to build strong ties and to socialize in computer supported interaction. Partly these are legitimate concerns and partly these are selflimiting beliefs. There are numerous examples of strong relations developed in computer supported collaboration. There are also examples of failing socialization in face-to-face groups. Factors like the experience of the people involved in computer-supported interaction and their willingness to cooperate can compensate for the additional effort needed to build a strong relation. Finally ICT supported work stimulates explicit rewarding and a focus on performance output. The explicit reward can be more valuable than just a pat on the back or a simple thank you. On the other hand there is a danger of losing the personal touch or forgetting to reward the dispersed colleague(s) completely. Moreover, in computer supported discussions emotional conflicts can very easily escalate, as there is much room for misinterpretation.

5.4 Contact and connection When relations have been established and their intensity has been determined, the next point of interest is the amount of accommodation needed. For contact and connection the key issue is information sharing. Connection deals with activities on the organizational level to enable information sharing. Contact focuses on determining the amount of desired actual interaction on the group level. The coordination activities regarding these aspects of coordination are shown in Table 35 and discussed below. Table 35: Coordination choices regarding contact and connection Contact Timing of interaction Intensity of interaction Media choice

Connection Organizational meeting schedules Infrastructure development Maintenance and support

The amount of accommodation can be varied with respect to how often employees in a particular work process or work group contact each other. In addition one can choose how long each contact lasts and whether it is more or less interactive. Closely related to these considerations is the issue of media choice. Depending on the demands for urgency, level of interaction and duration one medium may be more or less suited than another. Connection involves offering various alternative ways for employees to contact each other, as well as ensuring the reliability of these media. Examples are the arrangement of 105

COORDINATING DISTRIBUTED WORK all kinds of meetings where certain groups of employees can meet each other, as well as the development of infrastructure for all kinds of ICT, like computers, a computer network, telephones and a telephone network. Furthermore these media (and the scheduling process) need maintenance and support to ensure that they function as expected and are used as anticipated. The costs and benefits related to contact and connection are shown in Table 36. Table 36: Cost and benefits regarding contact and connection Costs Loss of work time Time and effort for planning contacts Time and effort for scheduling Costs, time and effort for development, maintenance and support of infrastructure

Benefits Concerted action Early discovery of delays and misfits Just in time information sharing Regular contacts Guaranteeing untroubled connection possibilities

All time spent on information sharing cannot be spent directly on work, as illustrated by the following remark of a police employee in one of our group sessions: “If I had to attend all the meetings I am invited for, I would never be able to finish my work”. On the other hand information sharing enables employees to adjust their efforts and produce a meaningful whole in the end. Time and effort spent on deliberation about the timing and nature of contacts, on scheduling and infrastructure development and maintenance leads to the various benefits described in Table 36 above. They prevent too early, too long or too rich contacts that decrease efficiency of information sharing, or they prevent too late, too short, too lean or even completely failing contacts that hamper effectiveness. The introduction of ICT support for contact and connection related activities has a direct effect on these costs and benefits (which is not surprising as these technologies are denoted as ‘information and communication’ technologies). The opportunities and threats with regard to contact and connection are presented in Table 37 and Table 38. Table 37: Opportunities and threats regarding contact Opportunities Synchronous distance communication Asynchronous communication Best medium for specific purpose Record of communication logs

Threats Lack of awareness and critical mass Delay or lack of feedback Meta communication problems Information overload

The main opportunities for contact can be explained by looking at the time and place matrix (see Table 3, page 8). ICT make synchronous communication between employees at different sites possible. Distributed colleagues can be contacted easier and communication can become more efficient because more contacts can be dealt with in a 106

CHAPTER 5: AN INTEGRATED VIEW ON COORDINATION OF DISTRIBUTED WORK shorter amount of time. Furthermore traveling costs may decrease. However, more distance may result in a lack of awareness and the likelihood of contacts taking place may drop extremely. Asynchronous communication can also be executed more efficiently with the support of ICT. This holds especially for employees at different sites, but also for asynchronous contacts at the same site. In the first case the increased speed of information sharing is the most prominent advantage, in the second case the ease of structuring and updating is the biggest benefit. But when asynchronous communication is becoming the main mode for interaction, the delay or complete absence of feedback may impede the effectiveness of communication. All together the introduction of ICT leads to a rise in the available media and thus to the possibility to select the best medium for a specific purpose. However, these choices take time, may result in sub optimal choices or may even create confusion amongst employees which medium is used by others and should be checked regularly. These kinds of problems are often named meta-communication problems. Finally there is the opportunity for automatic registration and easy archiving of earlier occurred communication. However, when registration is too detailed the danger of information overload can arise. The enormous amount of (for a big part irrelevant) data makes consultation of the historical logs too time consuming. Table 38: Opportunities and threats regarding connection Opportunities More productive communication

Threats Dependence on technology Technology failures (unavailability, incompatibility, complex interfaces) High costs of purchase and maintenance

The purpose of providing for a larger variety of media is to enhance productive communication in the organization. This opportunity is utilized when employees make use of the flexibility offered. However, when employees get accustomed to this flexibility they become dependent on the technologies often used. Or those who do not like to become dependent on technology refuse to take advantage of the flexibility. Whether reliance on technology is seen as a negative or positive characteristic of work is closely related to the amount of technology failures. Clearly, failures like unavailability of networks, incompatibility of programs and complexity of interfaces have a negative effect on productivity. Finally when costs of technologies are high this can result in postponing purchases or disappointing cost benefit ratios.

5.5 Content and context The fourth couple of coordination aspects focus on what information is shared during the various contacts. Key issue for content is to distinguish between relevant information, distracting noise and too specific details. On the organization level this discrimination can be facilitated by providing context for all information in the organization, which eases interpretation and sense making on the group level. Table 39 lists the coordination activities involved. 107

COORDINATING DISTRIBUTED WORK Table 39: Coordination choices regarding content and context Content Selecting relevant information Determining desired profundity Building shared understanding

Context Picturing environment Justifying mission and goals Making sense of organizational processes

For each contact to occur the relevant information to be shared has to be determined. There needs to be a balance between a short overview of the key issues and a profound and extensive explanation of certain details. Besides just ‘transmitting’ the information (point of attention in contact), there needs to be a concern for building shared understanding. Feedback about whether the receiver has read, understood and acted upon the information received can provide clues about the level of shared understanding. Efforts on the organizational level like picturing of the environment, justification of organizational mission and goals and sense making of organizational work provide employees with a frame of reference to interpret information and ease the determination whether certain information is important for their work at this stage or not. The cost and benefits involved in content and context are shown in Table 40. Table 40: Cost and benefits regarding content and context Costs Time/effort for sender to determine relevancy Danger of omitting relevant information Time/effort in determining profundity Time/effort for building shared understanding Time/effort for picturing environment and sense making of organizational processes Time/effort for justifying mission and goals Blocking organizational renewal

Benefits Less time lost for receiver to interpret Prevention that important information gets lost between irrelevant details Better chance of reaching shared understanding Less misunderstandings and rework Less time involved in building shared understanding on the group level More effective goal directed behavior Ambiguity reduction

When the sender of information determines relevancy of information beforehand, the receiver can save a lot of time in interpreting the limited amount of information during or after information sharing. Another consideration is the danger of omitting too much versus the prevention that one cannot see the wood for the trees. Also the better the level of profundity is adjusted to the situation and to the employees involved, the larger the chance that shared understanding is reached with minimal effort during information sharing. Furthermore checking for clues regarding achievement of shared understanding prevents misunderstandings and rework. The effort dedicated to ambiguity reduction on the organizational level is recovered on the group level in easier, faster and more effective work. However, too much ambiguity reduction is dangerous and can impede necessary organizational renewal. 108

CHAPTER 5: AN INTEGRATED VIEW ON COORDINATION OF DISTRIBUTED WORK The impact of ICT on the cost benefit ratio of content and context follows from the opportunities and threats presented in Table 41 and Table 42 below. Table 41: Opportunities and threats regarding content Opportunities Fast and easy sharing of large amounts of information with multiple employees Focus on required context for task at hand Less distraction of irrelevant context

Threats Information overload Less time and attention for interpretation Loss of physical, social and situational context

The main ICT impact here is that it becomes so much easier to send large amounts of data to where ever and who ever you like. The immediate consequence is the often-heard complaint of information overload. Indirectly the increasing speed and increasing amounts of data result in less time and attention for interpretation (Weick, 1985b). Both information overload and lack of interpretation can undo the advantages of faster information sharing. Another threat of information sharing with electronic media is the loss of physical, social and situational context compared to face-to-face settings. Although this can hamper coordination as a lot of potentially important information gets lost, it might turn into an opportunity when each employee is forced to address the relevant context explicitly in his communicative acts. Table 42: Opportunities and threats regarding context Opportunities Corporate electronic memory Central record of sense making stories

Threats Too many irrelevant entries and bad indexing Stories difficult to relate to the situation at hand

ICT can support the compilation of a central and easy to update and consult electronic corporate memory for a work process, work group, or organization that could reduce ambiguity by capturing shared understanding. However, usefulness of such an electronic memory depends on whether one succeeds in continuous updating and useful indexing. Otherwise searching is either inefficient or ineffective. The opportunities and threats of a record of sense making stories that incorporate organizational culture and context are also comparable to these discussions. An overview of those stories can be very valuable to increase productivity in comparable situations, but the clue is whether the reader can translate the story to practical instructions for the assignment he is working on.

5.6 The coordination framework as a whole In this section we summarize how our coordination framework pictures coordination compared to existing coordination theories discussed in chapter 3 and we illustrate how the coordination framework can be used to improve coordination by solving coordination problems and seizing ICT opportunities.

109

COORDINATING DISTRIBUTED WORK Coordination framework and coordination choices The main deliverable of the descriptive conceptual phase of our research is the coordination framework presented in Figure 4 (page 98). The framework illustrates that coordination is the combination of: • composing a group of employees with appropriate organizational competencies, • ensuring commitment among this group supported by an overall organizational cohesiveness, • producing contacts between the group members enabled by available organizational connections, • and sharing content with each other interpreted in an organizational context, in order to produce the desired results for a given assignment. We argue that the coordination framework fits with the common view on coordination that distinguishes between work arrangements, dependencies and coordination mechanisms. The framework concretizes division of work and management of dependencies in 4 dimensions on the group and organizational level. Although dependencies are not directly recognizable in our framework, they have a prominent role in our view on coordination. We argue that coordination choices can be seen as a continuous monitoring of how the organization currently manages their dependencies on the group and organizational level, and how this management of dependencies can be improved by changing the coordination mechanisms or changing labor division. However, a ‘good’ coordination choice does not solely follow from an analysis of dependencies, but depends also on technological, organizational, social and personal factors involved. Furthermore we claim that the framework unifies most of the contributions to coordination theory discussed in chapter 3. Notions of composition, contact and content can be related to respectively actors, processes and information in business engineering. Commitment is more prominent in studies on work groups and in theories on group dynamics. The distinction between the organizational level and group level is apparent in organization theory as well as business engineering. Competence is related to the issues of core competencies and knowledge management in organization science. Cohesiveness embraces issues like leadership, decision making, goal setting, shared values and mission that are recognizable in group dynamics as well as organization science. The notion of connection is observable in most IT-research. Finally context refers to the outside environment of an organization (most apparent in organization science) as well as to the importance of linguistic and interpretative aspects of the information sharing stressed in (social)- psychological contributions to coordination theory. In the first and second chapter of this thesis we stress that human and contextual aspects of coordination need more attention in the conception of coordination. We argue that the concept of coordination choices incorporates these issues. Coordination choices are taken by humans and are context dependent. The choices are not just a rational balance of weighing pros and cons against each other, but are also influenced by social, psychological, technical and work related factors of the situational context. In addition

110

CHAPTER 5: AN INTEGRATED VIEW ON COORDINATION OF DISTRIBUTED WORK these choices are not taken sequentially, but rather more or less simultaneously or even chaotically (Briggs, 1994). Also the choices may inform one another (poor quality of the accommodation may reveal that more or different people need to be involved). Finally in social settings historical coordination choices may have profound effects on current coordination choices. As such the framework gives room for addressing human and contextual issues. Their impact is more apparent in the design approach in chapter 6 when guidelines for improving coordination of distributed work are developed. The coordination choices in this framework can be recognized in standardized as well as spontaneous coordination of work processes, and thus in traditional and modern, simple and complex organizations. The only difference is that in standardized coordination the choices are taken more automatically and subconsciously as their pattern is more or less defined beforehand. Often coordination choices become then more or less fixed during the design phase. In spontaneous coordination options are left open and become only fixed during the execution of work or even during sense making afterwards. From situated action approaches we learn that coordination choices are articulated in plans during the phases of work design, work execution and sense making and that these plans may be applied more or less rigidly (as routes or maps). Solving coordination problems and seizing ICT opportunities The overview in Table 43 shows how coordination problems and ICT impact are related to the dimensions of our coordination framework. Table 43: Coordination problems and IT impact in relation to framework Coordination aspect

Dimensions

Coordination problem

ICT impact

Quantity of relations

Competence Composition Cohesiveness Commitment Contact Connection Content Context

Complexity

Positive

Ambivalence

Negative

Unpredictability

Positive

Ambiguity

Negative

Quality of relations Quantity of accommodations Quality of accommodations

Complexity and unpredictability are specifications of uncertainty (Kuipers and Amelsvoort, 1992). When the amount of different types of work assignments and the differences between these types increase the selection of suitable employees and composition of work groups become more complex. More possible compositions need to be taken into account and as a result the choice process requires more effort. Consequently there is more uncertainty whether the best or a satisfying composition for this assignment has been reached. Similarly, in a situation where work group compositions are more variable and work group members are less acquainted to each other and less familiar with the problems to be solved, the timing and intensity of interactions become less predictable. This unpredictability of contacts is another aspect of uncertainty. 111

COORDINATING DISTRIBUTED WORK We observe that uncertainty issues are connected with the amount of relations and the amount of accommodations. Ambiguity issues appear to be related to the quality of relations and the quality of accommodations. This distinction is not completely blackand-white. For example employee characteristics that are used for selection can be equivocal or vague, which results in ambiguity problems in the framework element of composition. Also considering content there can be uncertainty about which information is relevant and which is not. But in general uncertainty issues are more prominent in the left-hand side of the coordination framework (for composition and contact), and ambiguity issues are more profound at the right-hand side (for commitment and content). For example, ambivalence of goals increases when work group or work process members are employees from different departments and different sites in the organization. They have to divide their time over several work processes, which can result in conflicting goals. Because employees are less acquainted and because of their in-between distance they are likely to be insecure of each other’s intentions. As a result ambiguity about whether there is enough mutual commitment rises in distributed work settings. Similarly, work group members who come from different departments and sites and who are little acquainted interpret information in different backgrounds and contexts. This may lead to misunderstandings. Also the increase of explicit and written information in e-mails, intranets and database systems can increase misunderstandings as this information is separated from its context. In addition employees tend to neglect the comparison of different sources of information as they focus too much on the shared virtual workspace. This leads to ambiguity of content shared. From our overview in sections 5.2 until 5.5 we conclude that ICT impact is generally positive for solving uncertainty issues (composition, contact), and generally negative for ambiguity issues (commitment, content). For example, in composition the opportunities of providing overviews of all employees, their characteristics and their involvement in work groups and processes generally exceed the threats of outdated records and higher required efforts. Also in contact the opportunities of easy and frequent consultation most of the time surpass the threats of unavailable connections. To the contrary the threats of insecurity about each other’s intentions and lacking trust seem to be bigger than the opportunities of recording and sharing clear goals (commitment). Also the threats of misunderstanding due to context lacking information sharing appear to exceed the opportunities of recording and indexing past experience and communication (content). Therefore improving coordination in distributed work settings should focus on the opportunities of composition and contact and on the threats regarding commitment and content. However, whether ICT reveal as an opportunity or as a threat depends heavily on how the employees in the organization apply the technologies in their daily work and coordination, and on the situational circumstances. When employees are aware of the ICT opportunities the chance rises that they will be turned into real advantages. Similarly knowing the threats helps circumventing them. How this balancing of opportunities and threats works is discussed in the next chapter.

112

CHAPTER 5: AN INTEGRATED VIEW ON COORDINATION OF DISTRIBUTED WORK

113

6 A design approach for coordination of distributed work Now we have developed a thorough understanding of coordination in distributed work settings, the next challenge is to guide improvement of coordination in these work settings. In accordance with our third and fourth research questions (section 1.7, page 18) we are particularly interested in guidelines for coordination of distributed work and in organizational prototyping techniques supporting the change process. Both are addressed in the development of our design approach in this chapter. The design approach is structured according to the analytical framework as described by Sol (1990). This framework has been used to construct methodologies for new problem areas (Meel, 1994; Vreede, 1995; Eijck, 1996; Uijlenbroek, 1997; Lohman, 1999). With this framework, it is possible to characterize design methodologies by their mode of thought (Weltanschauung), their modeling constructs, their working method and their management approach. We refer to these characteristics as the way of thinking, the way of modeling, the way of working and the way of controlling. In section 6.1 until 6.4 these four ways are described. Given the focus of our research questions most attention is paid to the way of thinking (related to research question 3: guidelines) and the way of modeling (related to research question 4: organizational prototyping).

6.1 Way of thinking: Framework and guidelines The way of thinking describes the view on the problem domain and provides design guidelines the approach is based on (Lohman, 1999). View on the problem domain The object of design is coordination of distributed work in large service organizations. For analyzing distributed work we adopt a systems perspective (illustrated in figure 1, page 4). In addition a coordination framework (figure 5, page 98) has been developed that integrates various views on coordination and that addresses human and contextual aspects of coordination. Finally the impact of IT in distributed work is articulated in opportunities and threats related to the 8 coordination elements (sections 5.2 until 5.5, page 99 and further). Here we elaborate on the question how coordination of distributed work can be improved. 115

COORDINATING DISTRIBUTED WORK In chapter 3 we distinguish between uncertainty and ambiguity problems. In our case studies in chapter 4 we observe that uncertainty and ambiguity problems are jointly present in the work processes studied. The coordination framework presented in chapter 5 indicates that coordination choices need to be made in each of the four dimensions during design and execution of a work process. As uncertainty and ambiguity problems are related to respectively the left-hand side (composition and contact) and the right-hand side (commitment and content) of the framework, this implies that both uncertainty and ambiguity issues need to be dealt with to arrive at better coordination choices on all four dimensions. Both theory and practice show that it is problematic to formulate straightforward guidelines how to improve coordination for distributed work. In chapter 3 we show that uncertainty problems and ambiguity problems require different, sometimes conflicting, interventions. For instance, an increase of information processed may decrease uncertainty, but may increase ambiguity. The process is more efficient (more information about customer needs is gathered), but may become less effective (the customer needs are incompatible or cannot be translated into services at all). On the other hand, more interaction and time for interpretation may decrease ambiguity, but increase uncertainty. The process is more effective (a clear distinguishing service is developed), but may become less efficient (it only serves a fraction of all the customers’ needs). The overall impact on productivity is hard to determine for several reasons. First efficiency and effectiveness are both changing. Secondly, these changes are counterproductive. Thirdly, effectiveness is hard to measure accurately in the case of services. The discussion becomes even more complicated when we take into account that for certain objectives of an organization a rise of variety (which implicates more uncertainty) and a rise of ambiguity are perceived as beneficial. Examples are companies offering a wide variety of services rather than specializing on one service and organizations that are known for innovative solutions and approaches. A similar line of reasoning can be followed for the impact of ICT support on coordination of distributed work. In section 5.2 until 5.5 (page 99 and further) we show for each dimension of the coordination framework that opportunities and threats often go hand in hand. An employee or work group has to pay attention to both seizing the opportunity and circumventing the threat simultaneously. Coordination choices have both costs and benefits and the overall impact is unclear and differs from situation to situation A choice between two alternatives that both incur severe problems or that both are equally attractive is called a dilemma (Twist et al., 1998). The essence of a dilemma is that it is characterized by value-complexity (George, 1980): “Value complexity refers to the presence of multiple, competing values and interests that are embedded in a single issue. When this is the case, it is difficult, if not impossible, for the decision-maker to formulate a single yardstick that encompasses and aggregates all of the competing values and interests.”

116

CHAPTER 6: A DESIGN APPROACH FOR COORDINATION OF DISTRIBUTED WORK A dilemma is by definition unsolvable in theory. In our research we focus on the question how to deal with a dilemma in practice. Boudreau and Robey (1996), following Jones (1995), argue that there are three strategies for coping with contradictions present in dilemmas. The strategy of denial refuses to acknowledge the contradictions and dismisses them as misconceptions or as the product of flawed research. The strategy of resolution seeks to demonstrate that the contradictions are actually compatible (which would resolve the dilemma and turn it into a complex, but solvable problem). For instance, an intervention may produce different effects in different contexts, and different situations may call for different requirements and premisses that may very well contradict those of other situations. Resolution occurs when these different situational factors are included in a more complete analysis. This strategy is evident in the contingency approach in organization theory. Finally, the accommodation strategy implies that contradictions should be accepted, studied and understood as inherent phenomena in social systems rather than ignored or resolved (Handy, 1994). In our research we adopt the third strategy: accommodation. For Handy (1994) accommodating with the dilemma means living with paradox or living with simultaneous opposites. He argues that understanding the paradox is the key. Balancing the opposites, or switching between them, must not be a random or haphazard act, but requires a clear rationale for what is happening. Similarly Twist et al. (1998) stress the importance of validation of the dilemma and the consolidation of learning experiences in the quest for ‘dilemma-resistant-arrangements’. Examples of balancing simultaneous opposites can be found in contributions from the competing values school (Bots and Bruijn, 2002; Bruijn and De Neree tot Babberich, 2000; Collins and Porras, 1997; Maidique and Hayes, 1984; Quinn, 1988; Quinn et al., 1996; Quinn and Rohrbaugh, 1983). Quinn (1988) describes how big organizations can choose between opposing management styles (director, producer, coordinator, controller, mentor, stimulator, innovator and intermediary). Effective managers do not choose for one of these competing management-styles, but try to apply all management-styles simultaneously. Maideque and Hayes (1984) provide an example of two opposing management-styles, one that provides room, and one that is more hierarchical. Hierarchy is used to boost innovation processes, subsequently room is offered to professionals to develop innovations and finally hierarchical intervention follows when commercialization of the innovation is needed. Correspondingly Bots and Bruijn (2002) distinguish between an actor approach (hands off, create opportunities for interaction) and an analytical approach (hands on, guide action by creating focus) in knowledge management. Whereas others (Hansen et al., 1999; Swan et al., 2000) advocate to choose for one of these approaches, Bots and Bruijn argue that it is dangerous to restrict oneself to a single approach. The risk of a unilateral choice is that the advantages of the other approach are not utilized. This may end in a continuous flight process where one approach is criticized and driven out by the other approach, after which the new approach is driven out by the earlier one and so forth. In that situation knowledge management as a whole is under constant criticism which is undesirable. Instead, both the conflicting approaches need to be combined: not by blending them into one gray indistinguishable whole, but by using them distinctly at the same time, all the time. 117

COORDINATING DISTRIBUTED WORK Design guidelines We continue our discussion by demonstrating how opposing issues in coordination of distributed work, like uncertainty and ambiguity problems and opportunities and threats of ICT support, can be balanced and addressed simultaneously. In Table 44 we present some dilemmas and guidelines for coordinating distributed work that we have identified. Each coordination dilemma and accompanying paradoxical guideline is related to one of the dimensions of our coordination framework. They are discussed in detail below. Their explanation is illustrated and underpinned with references to coordination theories (chapter 3), observations in our case studies (chapter 4), opportunities and challenges of ICT (chapter 5) and contributions from research in the fields of virtual groups, virtual organizations, knowledge management, and communities of practice. Table 44: Coordination dilemmas and paradoxical guidelines Dilemma COMPOSITION / COMPETENCE Few - many relations 1. Flexibility - stability Standardization 2. customization Specialization 3. generalization COMMITMENT / COHESION Formalization 4. socialization Freedom - control 5. Support - critique 6. CONTACT / CONNECTION Autonomy - interaction 7. Information sharing interpretation Standardization customization CONTENT / CONTEXT Little - much content Little - much context Codification - sensemaking

118

8. 9.

Paradoxical guideline Consult distributed expertise extensively, but keep groups small, focused and productive Select employees by standard competencies present and by unique competencies needed Deploy experts to specialist and generalist tasks Stimulate both social communication and formulation of clear goals early in process Rotate leadership among group members True support requires expressing a critical attitude Contact regularly to show progress and evaluate, but do not interrupt productive action unnecessarily Let fast information sharing give room to interactive interpretation Customized (unpredictable) media choice requires (predictable) communication standards for all media

10. Be to the point but do not omit relevant details 11. Add missing context and hide distracting context 12. Let experts assist consultation of knowledge systems

CHAPTER 6: A DESIGN APPROACH FOR COORDINATION OF DISTRIBUTED WORK Guideline 1: Expertise consultation paradox (composition) Consult distributed expertise extensively, but keep groups small, focused and productive This guideline focuses on finding a balance between extensive deliberation of available expertise (to improve quality of service) versus the need to quickly arrive at an answer to the assignment the group is working on. ICT clearly give the opportunity to quickly consult distributed expertise, but really committing unknown (and distributed) experts to the group process may involve more time. Also, in a larger group of people it requires more coordination effort to keep the group process integrated and group members can more easily lose overview on what others are working on. Finally diverging views on how to solve the assignment may slow down the group process. On the other hand small groups, with co-located, known colleagues who have a unified view on how the assignment needs to be dealt with can perform more quickly, but their quality of service may be unsatisfactory because they lack certain expertise. In conclusion consulting distributed expertise (experts + knowledge systems) is beneficial for quality of service, but it has an adverse effect on the throughput time and the coordination effort needed. The dilemma is how to consult ‘just enough’. For example, in our first case study the Youth problem-solving group regrets that they did not involve employees from district 6. On the other hand from the diverging views and the difficulty to schedule meetings may be concluded that the number of 7 group members resulted in coordination problems. For throughput time a smaller group might have been better. Also in the Firearm case the fact that the Firearm-professionals are not always consulted results in a decrease of the quality of service. To handle this dilemma consulting distributed expertise has to be stimulated, but employees have to be aware of the costs of consulting expertise. When the delay in throughput time and the extra coordination costs grow too large they have to limit consultation. Guideline 2: Employee selection paradox (composition/competence) Select employees by standard competencies present and by unique competencies needed To be able to select unknown employees with relevant expertise for a particular assignment, there needs to be an overview of available employees and their expertise. The most efficient way to organize this matching is by linking employees to a set of standard knowledge fields. When a work group needs expertise of a given knowledge field the employees belonging to that field can quickly be selected and consulted. However, when the assignment does not fit in the existent knowledge fields, matching is problematic. Moreover, always choosing the most obvious expert to be consulted may result in pigeon holing (Mintzberg, 1983), whereas experts from other knowledge fields may provide innovative solution alternatives. To overcome these deficiencies it seems more valuable 119

COORDINATING DISTRIBUTED WORK to present the problem centrally and let experts participate in the work group when they feel they have unique expertise to help executing the assignment. However, that way of matching requires much effort (experts have to monitor new assignments and determine their suitability continuously). For both options ICT provide valuable support. Standard competencies can easily be listed, organized, searched and updated with the help of ICT. Central presentation of assignments can be facilitated by for instance e-mail or intranet. The dilemma is to balance quick matching by standard competencies and thorough matching by unique competencies. To handle this dilemma an employee can combine both strategies (select some experts by standard competencies, and some experts by unique competencies for a single assignment). In the Youth problem-solving group some participants were selected by their expertise field (Youth, social integration), others joined because they felt they could contribute from another point of view (traffic, firearms). Both the experiences with the topic under study as well as the innovating viewpoints proved to be valuable during their work process. Another possibility is to alternate between both strategies from assignment to assignment, depending on the time pressure (standard competencies when time pressure is high, unique competencies when time pressure is low). Guideline 3: Expert deployment paradox (competence) Deploy experts to specialist and generalist tasks Most experts have a strong commitment with their profession and like to spend working time in their expertise field. Consequently they do not like to work on other duties and would prefer to be released of them and become a full time professional. For the organization it also seems wise to assign expensive specialists as much as possible to the complex assignments they are trained for. However, the danger of a central functional staff unit far away from the customer, where experts wait until front office employees consult them, is that the experts keep spending time on their expertise field even if customers and front office employees have different needs. They cannot flexibly be assigned to topics where capacity runs short. Also, the most experienced organizational members are far away from the daily interaction with customers. Instead experts can be members of an arbitrary business unit and serve customer request from all kinds in that business unit. Assignments within their expertise field are of course reserved for them. Although the expert now has an eye for customer needs and for links between his expertise field and other disciplines, a danger is that his specialist experience is lost because he executes too few complex assignments in his discipline to learn from. These struggles are apparent in the Amsterdam Police Force, where firearm professionals are too much detached from their regional project. The challenge in this dilemma is to let the expert participate in general assignments when capacity outside his discipline runs short, and simultaneously guarantee sufficient participation in assignments related to his expertise field. ICT can help bridge the dilemma of working locally on general assignments and organization wide on specialist assignments. Although experts are located in general business units, they can be consulted 120

CHAPTER 6: A DESIGN APPROACH FOR COORDINATION OF DISTRIBUTED WORK or asked to participate in a complex assignment in their discipline, whenever and wherever their expertise is needed. Depending on priorities they can interrupt their own work and help or first finish their task and then contribute. In this way the specific expertise of the professionals is used when needed, but when priorities change and there is no work on their topic they can automatically take up one of the general assignments in their own business unit. The concession in such a dual structure is the complexity and pressure for the expert to serve two different bosses. Guideline 4: Process start up paradox (commitment) Stimulate both social communication and formulation of clear goals early in process Social interaction aims at enlarging social affiliation, whereas formal goals emphasize that something needs to be produced. It is commonly accepted that social interaction has a positive impact on performance. But social interaction limits the time available for delivering the product or service. If employees are (virtually) gathering to collaboratively work on a product or service and the exact purposes are unclear, social interaction can be experienced as long winded and a possible waste of time. On the other hand, too much emphasis on formal goals without any room for social interaction may harm employee motivation and lower group performance. The dilemma of balancing attention for formal goals (task focus) and for social support (people focus) is clearly present in the Youth problem solving group. They eventually address the dilemma by recognizing their conflicting goals and by distinguishing between moments and meetings that aim at the formal assignment or social group development. The dilemma can be handled by explicitly addressing formal and social needs at the start of a process. When goals and expectations are clearly expressed each participant can decide how much time he would like to dedicate in this process and how important social relation building is. Also the group has a rough feeling of deadlines in proportion to work that still needs to be done. So they can also weigh the usefulness and efforts involved in social interaction. Guideline 5: Leadership paradox (commitment) Rotate leadership among group members In a distributed work setting the employee has a lot of freedom to manage his own work load and performance over time and place. A backlog is that it is easy to hide under performance or even drop out of a process without anybody noticing it for some time. Therefore leadership and monitoring face some new challenges. Basically distributed workers need to have a high self-responsibility. However, shared responsibility for a process amongst a (large) group of employees is dangerous in the sense that if the process fails no one is really responsible and everyone blames everybody else. Both the youth problem-solving group and the regional project Firearms suffer from the lack of leadership and control as everybody is equal and clear leaders with formal authority are 121

COORDINATING DISTRIBUTED WORK absent. On the other hand appointing one leader lays the burden of controlling the others upon him, which is difficult to do accurately from a distance. Both the Youth coordinator in the first case and the regional coordinator Firearms are reluctant to ‘act as a policeman to control the work process’. They argue that some self-responsibility must be expected from local professionals. In conclusion distributed work gives room for self-management, which can enhance employee motivation, but there need to be some guarantees to control overall performance. Rotating leadership proves to be a promising alternative. Here leadership is passed along from time to time. The knowledge that you have to monitor the others at a certain point in time brings about that you please the current leader with honest performance in order to be sure that others will not betray you when you are responsible for progress and results. Rotating leadership thus balances the values of freedom and high responsibility with minimal control efforts. Guideline 6: Support paradox (commitment/cohesion) True support requires expressing a critical attitude Strong group norms, cohesiveness and commitment to the group goals make groups efficient and effective performers. However, too much harmony may have negative consequences, like groupthink (Moorhead and Griffin, 1989). For instance, the relevance of the solution proposed by the Youth problem-solving group can be questioned as they did only slightly attune their intermediate and final conclusions with employees from the work floor and the principal. The dilemma for a work group or a group of employees involved in a work process is to be both enthusiastic about the current way of working and current goals and simultaneously criticizing them and suggesting changes (Handy, 1994; Wenger, 1998). Both expressions of support for the current process and constructive criticism should be shared clearly and openly. In relation to Guideline 1 and Guideline 3 employees (especially experts) should be willing to participate in other work processes when they are requested to do so. But on the other hand they need to withhold their support when they are too busy or when they feel their contribution is not worth the effort. Otherwise they are wasting precious time. In sum, experts should simultaneously be enthusiastic and critical about their participation. Guideline 7: Contact frequency paradox (contact) Contact regularly to show progress and evaluate, but do not interrupt productive action unnecessarily A common belief is that groups are more productive if they spend more time together. When this belief is existent in a distributed group it may cost them a lot of energy as it is more difficult for a distributed group to organize synchronous or even face-to-face meetings. However, most part of group work is individual work (Andriessen, 2002). 122

CHAPTER 6: A DESIGN APPROACH FOR COORDINATION OF DISTRIBUTED WORK Individual work is more productive when done concentrated on your own, because in the group you may suffer from distraction and disturbance by others. A distributed group can more easily utilize the moments of autonomy and concentration as long as they keep in mind when they need to interact. On the other hand they may lose sight of each other’s work and common goals when they interact too little. The regional coordinator Burglary (who attended only 2 of the 10 meetings of the Youth problem-solving group) and the Firearm professionals suffer from the lack of contacts with the larger collective. On the other hand time neighborhood team managers and professionals complain about the amount of meetings the professionals need to attend (locally, on the district level and on the regional level). The dilemma faced is that too few interactions may harm effectiveness, whereas too many interactions may harm efficiency. Clearly distributed groups are closer to the danger of too few interactions, but co-located groups are closer to the danger of too many interactions. Both dangers should be paid attention to in co-located as well as distributed work groups. Purposeful contacts need to be increased and distracting contacts need to be limited. Guideline 8: Contact quality paradox (contact) Let fast information sharing give room to interactive interpretation In the theories of information richness (Daft and Lengel, 1984; 1986), communication richness (Ngwenyama and Lee, 1997) and media-synchronicity (Dennis et al., 1998) a distinction is made between tasks requiring more and less richness/synchronicity. Rich (high synchronicity) contacts emphasize interpretation by high interaction and direct feedback, whereas poor (low synchronicity) tasks contacts stress fast and mutliple information exchange. On the one hand ICT provide opportunities to exchange information faster and in larger amounts, on the other hand communication with distributed experts can become more interactive with the help of ICT. However, in our case studies in the Amsterdam Police Force we observe that both features are potential, and that they are hardly utilized by the police employees. In practice work processes or even tasks may have contradicting requirements. On the one hand more information is required to reduce uncertainty, on the other hand interaction and direct feedback is needed to collaboratively interpret information. However, gathering too much information complicates interactive sense making, and collaborative sense making around a few issues leaves less room to gather and share information on other related issues. The challenge of this dilemma is to keep an eye on both competing values in media choice and to link the different communication processes. Results of fast information gathering and sharing can widen the scope of the discussion during sense making episodes, and outcomes of sense making discussions can limit the information sharing around topics that are judged as less relevant.

123

COORDINATING DISTRIBUTED WORK Guideline 9: Media use paradox (connection) Customized (unpredictable) media choice requires (predictable) communication standards for all media As has been noted in chapter 5 introduction of ICT leads to a rise in the available media and thus to the possibility to select the best medium for a specific purpose. This variability is a benefit as urgent matters do not have to be saved for the next meeting in two weeks (like in the Youth problem-solving group), but can be shared faster via another medium. But the link between purpose and preferred medium can be different for different situations and people involved. For instance one distributed group can prefer voice mail for urgent matters, while another can prefer e-mails. Here problems can arise like how to reach colleagues quickly and like insecurity whether messages are received and acted upon. Complete freedom is harmful for performance when group members are insecure about how others use different media. In that case they cannot compare costs and benefits of each medium for a particular situation and consequently media choice is not selecting the best medium but guessing which medium will be best. Therefore customization of media choice has its limits and needs to be balanced with predictable media use. It does not matter whether a group uses voice mail or e-mail for urgent matters, the most important thing is that employees have the same mental model considering media use (Haywood, 1998). Especially for asynchronous communication issues like availability and acknowledgment of message receipt and taking action upon it need to be captured in standards. So when voice mail and e-mail are used for communicating urgent matters, they should be checked several times each day. Explicating these standards needs to be a common discussion topic in the forming or storming phase of a group. Preferably similar rules apply to the whole organization. Guideline 10: Content sufficiency paradox (content) Be to the point but do not omit relevant details When electronic written media are frequently used for intra group communication information overload may threaten. The danger of fighting information overload is however that too much relevant content is omitted. For instance, oral communication much more focuses on the core of the message, but is less suited to communicate lots of details. The dilemma is to balance content overload against content shortage. Both senders and receivers have their share. The sender can clearly distinguish between the core of the message and the details. As a result the receiver can better divide his available time between reading and interpreting both aspects of the message. Furthermore, by clearly phrasing his question the receiver can give the sender clues what information can be omitted and so he helps to fight information overload by influencing the sending process. The challenge is thus not to simply increase or reduce content, but to distinguish between different kinds of content and communicate them simultaneously. 124

CHAPTER 6: A DESIGN APPROACH FOR COORDINATION OF DISTRIBUTED WORK Guideline 11: Context sufficiency paradox (context) Add missing context and hide distracting context Corresponding to the contact frequency paradox (Guideline 7) and content sufficiency paradox (Guideline 10) similar opposite requirements can be discerned for context. Distributed groups tend to suffer from a lack of context whereas co-located groups may suffer from too much task-irrelevant context. Context has been shown to be an important vehicle to ease interpretation of content (Daft and Weick, 1984; Weick, 1985b). Distributed groups therefore need to add missing context. On the other hand co-located groups may take the availability of context for granted and may forget to distinguish between relevant and irrelevant information. As such they may suffer from context overload. Although their starting points are different, both co-located and distributed groups need to discuss what context is important to addres and what context is distracting in their work process. Consequently both kinds of groups strive for adding and reducing context simultaneously. Guideline 12: Knowledge consultation paradox (context) Let experts assist consultation of knowledge systems Although especially novice employees can less rely on their own knowledge and need colleagues and expert systems to check their intended behavior, they may face more difficulties in using expert systems than real experts do. Novices lack an overview of the knowledge area and they may use wrong key words for full text search and for consulting the index. Also the art of linking the relevant bits of knowledge to a particular problem may be hard to capture in a knowledge system. In the Amsterdam Police Force police employees experience problems in consulting the intranet for knowledge questions. For experts the knowledge system is easy to use, but they do not really need it because they are either familiar with the knowledge in the system or can easily access and consult the old paper based alternative knowledge systems. Often it is argued that novices should consult the knowledge system for simple problems, and consult experts for complex problems. Otherwise experts are flooded with consultation requests or get bored by giving the same advice over and over again. However, whether a particular problem is faced as simple or complex can vary from employee to employee. Also problems can contain simple and complex parts and it may be inefficient to consult both the knowledge system and an expert for the same problem. In conclusion, there needs to be a balance between consulting a knowledge system and consulting the expert. Too much relying on the knowledge system may lead to inefficiency for the problem owner and a decrease in quality of service, whereas too much reliance on the expert may lead to inefficiency for the expert. To deal with this dilemma experts need not only to be motivated to maintain the knowledge in the system but should also be available to pinpoint novices where they can find what information. Pinpointing to 125

COORDINATING DISTRIBUTED WORK a specific part of the knowledge system is more efficient than each time explaining the whole procedure, approach or solution from zero and is also more efficient than letting novices search themselves. Of course, it is advisable to explain the searching process and the reasoning behind linking problem to solution so the novice learns how to solve similar problems without the help of an expert. When experts design as well as actively use the knowledge system the chances rise that content omissions and structural inadequacies are solved faster.

6.2 Way of modeling: Task-actor models and simulation-games The way of modeling describes the techniques used in the design approach to abstract from reality, e.g. make models of the problem domain (Hengst-Bruggeling, 1999). Modeling can be defined informally as the “art and science of leaving out” (Meel, 1994). Models help to provide the essential details of a problem situation, to facilitate understanding and to prove guidance towards solutions (Lohman, 1999). A distinction can be made between conceptual models and empirical models (Vreede, 1995; Eijck, 1996). Conceptual models define the problem structure and are more abstract. Empirical models enable detailed analysis and diagnosis of the problem or possible solutions (Vreede, 1995). We argue that no new modeling techniques need to be developed to support our design approach. Task-actor models (Vreede, 1995; Eijck, 1996) fulfill the need of providing a quick overview of current work processes and coordination mechanisms as well as sketching possible future alternatives. They are used as conceptual models in our design approach. In addition gaming-simulation fulfills the need of experiencing new work processes, new coordination mechanisms, new ICT to support coordination and future organizational context as a meaningful whole. By enacting the future situation organizational employees can both evaluate the new coordination alternatives more thoroughly as well as gain experience in dealing with the coordination framework and the paradoxical guidelines. Simulation-games can be seen as empirical models in our approach. Below the most important features and the way to apply both modeling techniques are explained. Task-actor models Building on the work of Bots (1989) and Dur (1992) Vreede (1995) and Eijck (1996) adopt an object-attribute-action worldview to conceptualize organizational coordination. We see organizations as systems, consisting of networks of communicating objects. Organizational situations can be conceptualized by describing the attributes and actions of perceived objects. The attribute part contains data elements describing the state of an object. The action part contains the tasks and decisions that reflect the possible behavior of an object. As the behavior of individual objects is seldom relevant, objects with similar characteristics are joined in object classes. An object class represents the attributes and actions of a set of similar objects. Both the task structures of Bots and Dur as well as the task-actor models of Vreede and Eijck have proven to be easy to communicate.

126

CHAPTER 6: A DESIGN APPROACH FOR COORDINATION OF DISTRIBUTED WORK The main advantages of the object-attribute-action view are (Bots, 1989, Dur, 1992, Eijck, 1996, Meel, 1994; Vreede, 1995): • the notion of an object quite naturally describes both what it is (attributes) and what is does (actions); • the distinction between an object class and an object instance naturally represents the difference between the conceptual and empirical level of abstraction; • autonomous behavior by organizational actors can be adequately represented; • the close relationship between the phenomena of the problem situation and their object equivalents increases the comprehensibility of the model and facilitates its communication to others. In order to conceptualize organizational processes and organizational coordination Vreede and Eijck distinguish between eight ‘leaf” object classes that inherit certain properties from a number of superclasses. There are three basic superclasses: • Item: passive object that can be communicated or exchanged between item processors. Items cannot carry out any actions themselves, they only contain information in their attribute values that may trigger the receiving item processor to select a particular course. Items can be divided in the subclasses messages (direct information representation), products and persons (indirect information representation). • Item processor: active object that can perform actions like sending and receiving items, altering its own status attribute values and altering the attribute values of the items it processes. Item processors are further divided into nodes and links. A node stores, retrieves and processes items and can be an actor or a repository. A link connects and routes items. • Actor action: transition between preconditions and postconditions. A distinction is made between tasks (involving a delay, attribute assignments and possibly subtasks) and decisions (a choice between two or more courses of action taken with the rationale of a decision rule). The eight object classes of the object hierarchy form the basic building blocks for three aspect models that each highlight different aspects of the way in which an organization achieves the coordination of its processes and activities. Together these models provide a multi-perspective analysis of the modeled situation. These models are: • An interaction model, which aims to facilitate the understanding of coordination relationships within an organization by highlighting the exchange of items (message, product, person) between nodes (actors, repository). • A process model, which aims at clarifying how actors in the organization handle items, by depicting the interdependencies of actors’ and repositories’ actions, that together are aimed at achieving an organizational goal. A process is thus a sequence of actions carried out by interacting nodes in order to achieve a certain goal. • An actor model, which aims at facilitating the analysis of the behavior of an individual actor by highlighting the interdependencies of the actions this actor has to perform and coordinate in order to achieve his goals. An actor model consists of the same elements as a process model.

127

COORDINATING DISTRIBUTED WORK Considering the building of these models participant observations and interviews with key stakeholders can be held to gather information. Vreede (1995) describes how a first version of the models can be built by a researcher or analyst (who has experience in model building) and by a small group of employees (who can implicitly verify the first models). Next these versions can be evaluated and improved with key stakeholders individually or participatively (Vreede, 1995). Simulation-games The second modeling technique we present is gaming-simulation. We argue that simulation-games provide a richer and deeper insight in the interactions between work, humans, technology and organizational context of a possible future situation. The taskactor models give only a rough overview of a new work process and its coordination (by highlighting the actors involved, the information they exchange, and the sequence of processes they perform). Gaming-simulation is a particular form of simulation. Simulation can be defined as the process of designing a model of a concrete system and conducting experiments with this model to understand the behavior of a concrete system, to evaluate various strategies for operation of the system, and/or to study the impact of scenarios representing a particular path to a hypothetic future situation (Meinsma, 1997; Shannon, 1975; Sol, 1982). This definition makes a distinction between the simulation model and simulation as the process of building the model and experimenting with it (Bockstael-Blok, 2001). Different types of simulation can be distinguished based on the degree of human involvement and interaction in executing the simulation model (Bockstael-Blok, 2001): • computer simulation (no human involvement during execution); • interactive simulation (part of the model is executed by the computer, part of the model involves action and decisions of one or more human actors, but these actors do not interact directly with each other); • gaming-simulation (humans interact and make almost all decisions). A simulation-game is a specific kind of simulation model that represents a complex problem area and contains different roles to be played by participants and game administrators, intended objectives to be achieved, activities to be performed, constraints on what can be done, and payoffs (positive and negative) as a result of their actions and the actions of other elements in the system (Caluwe, 1997; Caluwe et al., 1996; Greenblat, 1988; Meinsma, 1997). In a simulation-game the game roles, goals, activities, constraints and consequences and the linkages between them are patterned for real life, or simulate the above mentioned elements of real-world systems (Greenblat, 1988). Consequently, some games (like for instance card and board games) are not simulationgames, as they are not designed to represent real-world systems. Similarly, many simulations that do not incorporate human decision making or focus on the training of one actor rather than the interaction between actors are not simulation-games. Greenblat (1988) states that for that reason also role-playing differs from gaming-simulation in the degree of structure and formalization (respectively freewheeling versus interacting system) and in the focus on playing individual roles rather than interaction processes.

128

CHAPTER 6: A DESIGN APPROACH FOR COORDINATION OF DISTRIBUTED WORK

The terminology for gaming-simulation is not applied consistently (Duke, 1995; Meinsma, 1997). Gaming-simulation is also called simulation, gaming, simulation/gaming, gaming/simulation, policy-exercise, free-form simulation, role playing and social simulation in literature. In this research we use the term “gamingsimulation” to refer to the modeling and experimenting process, and the terms “simulation-game” or “game” to denote the models itself. A simulation-game (Geurts, 1995; Meinsma, 1997): • can help to understand the behavioral complexities within a concrete system; • can facilitate the integration of multiple perspectives within the model; • may use the experience of individuals productively; • and enables problem owners to experiment with scenarios representing a particular path to a hypothetical future situation and to study the impact of these scenarios. Simulation-games can also be seen as educational instruments for teaching and training. Possible objectives are (Caluwe et al., 1996; Geenblat, 1988): 1. Increasing awareness, motivation and interest; 2. Teaching by conveying information or reinforcing information already given; 3. Skill development; 4. Practicing communication and collaboration; 5. Attitude change; 6. Self-evaluation or evaluation by others; 7. Integration of learning experiences. Gaming-simulation turns passive recipients of information in a classroom into active participants. Thereby it stresses the importance of experience based learning, reflected in the learning cycle of Kolb (1984): concrete experience, observation and reflection, formalization of abstract concepts and generalizations, testing implications of concepts in new situations. The belief in the importance of experience based learning is reflected in the following old-Chinese proverb: I hear and I forget I see and I remember I do and I understand In addition several authors pinpoint at the joy involved in learning with gamingsimulation. Elgood (1988) states: “interest, excitement and enjoyment are emotions naturally associated with the game format: people get excitement from variety, novelty, color, humor and the unexpected”. Corbeil (1999) even states that participants learn in a game as a result of the joy they experience. He underpins this by referring to two opposite forces involved in learning described by Morris (1993): Neophilia (the curiosity to learn something new) and neophobia (the fear for new situations and the tendency to fall back on older methods). Joy stimulates the natural tendency for neophilia and gamingsimulation provides a safe environment that reduces neophobia.

129

COORDINATING DISTRIBUTED WORK Several authors describe an approach for designing a simulation-game that resembles general design approaches (Caluwe, 1997; Caluwe et al., 1996; Greenblat, 1988): 1. draw up a functional set of requirements; 2. analysis of processes and mechanisms; 3. design of an integral framework; 4. draw up a game-technical set of requirements; 5. construction of a prototype; 6. testing of the prototype; 7. final adjustments to improve the game. These steps need to be executed iteratively rather than sequentially. Furthermore it is important to involve some employees from the organization under study in the design team and share intermediate results of the simulation-game with a larger group of potential participants. The challenges that need to be faced during design are (Duke, 1974): • create order in real-world complexity; • make the complexity manageable; • make the small world in the game understandable and consistent; • make the game content and rules fascinating and sticking into the participant’s memory for a long time; • arrive at the pre-formulated goals and objectives as easy as possible; • create and preserve motivation of participants during the simulation-game. Furthermore Vrolijk (2000) lists a lot of gaming-simulation technical decisions that need to be taken such as: symbolic or realistic scenario, degree of complexity, degree of competition, degree of dynamics, macro-cycle (preparation, follow-up), micro-cycle (amount of rounds), real-time or symbolic time, what location, what roles, attendance of supervisor, amount of players, strict rules or relative freedom and whether or not to use computers.

6.3 Way of working: Problem solving and appreciative inquiry The way of working describes the activities that need to be taken to address the problem of improving coordination of distributed work. Hereby the appliance of the guidelines presented in section 6.1 and the appliance of the modeling techniques presented in section 6.2 are integrated. We adopt a problem solving perspective on changing organizational coordination and complete this with some aspects of the method of appreciative inquiry. In the problem solving literature there are numerous strategies proposed to solve (ill-structured) problems (Ackoff, 1978; Bots, 1989; Brightman, 1980; Dunker, 1945; Mitroff et al.,1974; Sol, 1982; Simon, 1973). Basically these methods all make use of variants on the following steps: diagnosis, analysis of goal, development of alternatives, evaluation of alternatives, and choice and implementation. 130

CHAPTER 6: A DESIGN APPROACH FOR COORDINATION OF DISTRIBUTED WORK Visser (2001) and Cooperrider (2000) criticize traditional organizational change methods for focusing too much on things that are wrong and that do not function. He states that this is caused by an underlying assumption that organizations are machines and that defect parts need to be replaced. He argues that searching for problems not only results in finding problems but also in creating them. By focusing on problems, problems are emphasized and enlarged unconsciously. These defect-oriented methods create a negative atmosphere and may result in loss of energy, cynicism and growing resistance for more organizational change. Instead they propose the method of appreciative inquiry that pays attention to looking for and reinforcing already existing solutions. The method is grounded in principles like the strength of our imagination to create new futures and like the experience that positive (appreciating) questions and views lead to positive data, willingness to cooperate and a more sustaining change process. Typical steps in the appreciative inquiry approach are: • appreciation: value the best of what is currently present; • envisioning: imagine how the organization could develop; • dialogue: organization wide conversation how the organization should develop; • innovate: implementing the chosen change alternatives. Variants on the basic steps are Definition, Discovery, Dream, Destination and Initiate, Inquire, Imagine, Innovate. Although we acknowledge the critiques of appreciative inquiry on traditional problem solving, we do not think that abandoning the traditional problem solving methods (as is advocated) and completely neglecting the problems in the diagnosis phase is wise. The tension between traditional problems solving and appreciative inquiry is comparable to the tension between approaches and views of ethnographic researchers (focus on present processes) and IT design researchers (focus on new opportunities) in studies on computer supported cooperative work (Artman and Wearn, 1995). Addressing problems and opportunities is both valuable. Therefore we incorporate both views in our design approach. This results in working method of five steps, which are clarified below. 1. Analyze current work processes Drawing upon the earlier cited guideline that understanding needs to be developed before design (Sol, 1992), we strongly advocate to analyze current work processes and coordination as a start of the design process. Given the discussion above the analysis should focus on: • providing an overview of existing processes, their interrelations, and the actors involved in executing them; • identifying and describing best practices; • identifying and describing experienced problems (worst practices). As suggested above both problems as well as good working practices can be an inspiration for improvement.

131

COORDINATING DISTRIBUTED WORK To collect the data for these three purposes we propose not only to rely on organizational archives and reports and interviews with managers and employees, but also to counterbalance these sources with direct observation of work practices. During execution of the actual work processes the effects of macro level structures and meso level coordination mechanisms are combined with the merits of informal coordination. It is only in actual execution of coordination that we really see the impact of all work related, technology related, social, psychological and organizational factors as a whole. In this phase task-actor modeling is used to provide insight in processes and their interrelations. The coordination framework can be used as a general checklist to arrive at a complete overview of coordination mechanisms, and to identify problems and best practices concerning coordination. 2. Identify promising opportunities for distributed work In the second step of the approach proposals for improvement are developed. In the beginning the aim of this phase is to divert to as many creative ideas as possible. The change alternatives can develop from problems to be solved, best practices to be reinforced, ICT opportunities to be utilized and by out of the scope thinking (what if our organization would be designed from scratch, regardless of its current history). In general distributed work can be beneficial where multidimensional dependencies between customer, work, resources and location exist. The findings in the previous step and the opportunities for the various aspects of the coordination framework can serve as a checklist, but should not limit the scope of the discussion. At the end of this phase a rough cost benefit analysis needs to be made for each of the generated alternatives with the help of the coordination framework. The expected costs and benefits can be listed and compared. Finally the most promising opportunities need to be selected for further investigation. 3. Design coordination of distributed work In the third step each of the selected opportunities is developed into an enduring coordination mechanism. This involves defining the appropriate interventions regarding competence, cohesiveness, connection and context on the organizational level and developing options and strategies for situationally managing composition, commitment, contacts and content on the group level. The strategies should picture how benefits can be maximized and costs minimized, and how threats can be circumvented. As we observed when formulating our general guidelines in section 6.1, seizing opportunities in distributed work involves coping with dilemmas. Consequently the strategies should include clues how conflicting goals can be reached simultaneously. Here not only the coordination framework and related opportunities and threats, but also the guidelines described in section 6.1 are useful.

132

CHAPTER 6: A DESIGN APPROACH FOR COORDINATION OF DISTRIBUTED WORK 4. Develop and play simulation-game In the fourth step a simulation-game is developed that reflects the challenges of coordinating distributed work in this particular organizational context. Thus far the new coordination mechanisms have been treated as isolated elements. Also the four dimensions of coordination have largely been designed independently. By developing and playing the game employees can gain experience with coordinating their distributed work processes and find out which strategies are most promising for them under different circumstances. They learn to apply the relevant paradoxical guidelines in this particular case in their particular organizational context. As a result the simulation-games can lead to improving certain newly developed coordination mechanisms. Another outcome may be that the developed coordination mechanism should not be implemented because the costs prove to outweigh the benefits. Also unexpected benefits can be developed into mature change alternatives. As such the simulation-game serves as an organizational prototype to evaluate new coordination mechanisms in an early stage of the design and implementation process. 5. Implement the new work processes Finally the developed design (possibly altered after playing the game) is implemented according to the design guidelines presented before. In this stage technology-use mediation by CSCW mediators as described by Orlikowski (1995) can be very beneficial to quickly react on problems during implementation like negligence of certain functions or features, technological failure, unintended use or unexpected benefits. In conclusion we summarize the main distinctive features of the design approach: • 5 steps need to be undertaken to develop and implement distributed work processes and distributed coordination, namely analysis, opportunity finding, design, gamingsimulation and implementation; • opportunities can be valued with help of the coordination framework by defining their benefits as well as costs; • 8 coordination elements and 12 related paradoxical guidelines deserve attention during design, prototyping and implementation; • gaming-simulation is valuable to fine tune the design and explore design and implementation dilemmas in an early stage of the change process.

6.4 Way of controlling: Adaptive and incremental Now we have related the way of thinking and modeling in this description of the way of working we only need to address the way of controlling to complete our design approach. The way of controlling describes the control of the working procedures and the models used to design coordination of distributed work. This implies project design, checkpoints, documentation and time management (Eijck, 1996; Turner, 1980). Hence, the way of control represents the managerial aspects of the problem solving process (Lohman, 1999).

133

COORDINATING DISTRIBUTED WORK Meel (1994) discusses two extreme control strategies. Within an expert strategy (Taylor, 1947) a technical expert or analyst is responsible for the actual problem solving and reports to a decision-maker, preferable the most important stakeholder in the client organization. Such a control strategy is very efficient, but results may be difficult to implement because of the “not invented here syndrome”. In contrast within a participative strategy (Emery and Emery, 1974; Mumford, 1983) the actual problem solving is done by (representatives of) the different stakeholders in the client organization. Design decisions are made on a basis of consensus between (representatives of) the different stakeholders involved. Such a control strategy can ease implementation, but can be very inefficient as it asks a lot of effort from all parties (Eason, 1988). An adaptive control strategy tries to combine the best of both worlds by picturing the design process as an adaptive process of learning for both analysts and the stakeholders of the client organization (Keen, 1980; Lohman, 1999; Meel, 1994; Sol and Bots, 1993). The strategy should motivate stakeholders to bring forward their views and (tacit) knowledge of the organizational characteristics like business processes, bottlenecks in current functioning and possible improvements. On the other hand analysts should be motivated to bring forward their knowledge and experience with respect to techniques, methodologies and tools and can suggest possibly overlooked alternatives, both organizational and technical. As such stakeholders can learn from analysts and vice versa by closely cooperating during all phases of the design process (Meel, 1994). To facilitate learning the design approach and modeling process should be “middle out” (focusing on a small but critical part or the organization that is recognizable) and “incremental” (quick delivery of an initial model that can be revised and enhanced in a number of cycles) as argued by Sol (1992). Thereby quick feedback is gained, which is not only beneficial for learning and the quality of the design, but also helps to strengthen management and employee support for the change process. Furthermore it should be noted that although we presented the steps in section 6.3 in a linear fashion, they do not have to be undertaken fully chronologically in practice. In contrast their application will be iterative rather than sequential, in order to enable continuous interaction between the different steps to tune the results (Lohman, 1999). Before continuing with the action research studies, in which we applied our design approach, we would like to stress that our approach is not prescriptive in the sense that all activities must be carried out (Vreede, 1995). The design of organizations and information systems is such a complex, ill-structured problem, that a universal design approach is very unlikely to exist (Checkland, 1981; Sol, 1982). As a consequence Vreede (1995) concludes that a design approach like ours is conceptual. During application of our approach (guidelines, modeling techniques, activities, managerial interventions) the analysts and stakeholders have to specify which of the elements of the approach they want to apply in a particular situation. For instance, for a particular coordination problem it is not likely that all 12 guidelines are equally important. Combined with its iterative nature the design approach is thus not a cookbook, but should be used in a flexible way that is determined by the nature of the problem and the style of the analyst. 134

CHAPTER 6: A DESIGN APPROACH FOR COORDINATION OF DISTRIBUTED WORK

135

7 Future coordination at the Amsterdam Police Force This chapter contains the set up and results of two action research studies within the Amsterdam Police Force. The aim of these studies is to apply and evaluate the design approach presented in chapter 6. One action research study focuses on the involvement of distributed expertise in daily problem solving in neighborhood teams (Laere, 2000; Vrolijk, 2000) and has been carried out from September 1999 until June 2000. The second action research study involves the coordination of filling and maintaining the intranet site and answering and verifying answers on the message boards of the regional projects (Blom and Korf, 2001; Laere, 2002a) and has been carried out from October 2000 until August 2001. In section 7.1 we describe the set-up of both action research studies. Sections 7.2 and 7.3 report of the activities undertaken in both studies and present the resulting outcomes. The reflection upon these findings is presented in chapter 8.

7.1 Action research set-up Similar to the set-up of our case studies (section 4.1) we discuss respectively the units of analysis and data to be collected, the case characteristics, and the data collection sources. As we want to reflect on the applicability and usefulness of the design approach our data collection needs to focus on the appliance and use of the elements of the design approach. Therefore the main units of analysis are the steps of the way of working (see section 6.3) and the aspects of the way of thinking and the way of modeling related to these steps: 1. Current work processes: • task-actor models (section 6.2); • coordination framework (section 5.1). 2. Opportunity selection: • coordination framework (section 5.1); • costs and benefits of opportunities (sections 5.2 until 5.5). 3. Future work processes: • paradoxical design guidelines (section 6.1); • task-actor models (section 6.2). 137

COORDINATING DISTRIBUTED WORK 4.

Develop and play game: • game design (section 6.2); • paradoxical design guidelines (section 6.1); • game results. Because of time constraints it is not in our reach to carry out the fifth step of our approach (implementation). The reports of the action research studies are structured following these units of analysis. That structure enables a smooth linking of the findings of the action research studies to answering the research questions in the reflection in chapter 8. We select two action research cases that are different in nature to increase the learning opportunities. Similar to our exploratory case studies the action research studies can be characterized as shown in Table 45. Just like in the case studies the first action research can be seen as task oriented problem solving groups, whereas the second action research is dedicated to networks of experts or communities of practice. Table 45: Characteristics of the two action research studies Characteristic TASK VARIABILITY TASK COMPLEXITY GROUP SIZE GROUP MATURITY

Neighborhood team Ad hoc Multi-disciplinary Small Short term relation

Regional project Routine Mono-disciplinary Large Long term relation

The following data sources are used during our action research (see Table 46). Table 46: Data sources in the action research studies Quantitative data source Questionnaires Game logs

Qualitative data source Organizational documents Organizational archives Interviews Observation of meetings Observation of work practices Group discussions Game debriefings

Neighborhood team X X

Neighborhood team X X X X X X X

Regional project X X

Regional project X X X X X X X

During the analysis of and reflection on the action research findings we use our research questions, the coordination framework and the design approach steps to structure our conclusions.

138

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE

7.2 Collaborative problem solving at the neighborhood teams In the second exploratory case study one of the problems observed is that the quality of work decreased (see page 80). Besides factors like a low priority for the task, time pressure and lack of capacity, the difficulty of consulting expertise is indicated as a reason for poor quality of performance. From the analysis follows that regional projects invest heavily in anticipatory coordination like training experts, developing work procedures, writing newsletters, organizing meetings and workshops. However, little attention is paid to the appliance of all this generated knowledge and skills for daily problem solving. Then local experts may not be present, handbooks may be inaccessible or out of date and distributed experts are hard to contact. As such all the anticipatory efforts are lost when adaptive coordination during execution is failing. Therefore the first action research focuses on daily problem solving in the neighborhood teams. The major question to be answered in this study is whether and how consultation of distributed expertise can improve the quality of daily problems solving in neighborhood teams? To answer this question the design approach described in chapter 6 is applied. The five subsections below subsequently address the analysis of work processes and coordination problems, the selection of opportunities, the design of new coordination mechanisms and the development and results of a simulation-game. 7.2.1 Work processes and coordination problems In the second case study in chapter 4, where we analyze the regional project Firearms, we already touch upon most of the central work processes of the Amsterdam Police Force. Here, we focus on the tasks carried out by the neighborhood teams. The data are gathered by studying organizational documents and archives, by observing work practices, attending meetings and by interviewing employees from different levels in some neighborhood teams (January-March 1999). The neighborhood teams have the following basic tasks: (1) emergency service, (2) take down notifications, (3) criminal investigation, (4) project work, (5) surveillance, and (6) community policing. A neighborhood team consists of about 60 to 80 policemen that have to perform these tasks in shifts in order to deliver a 24 hour 7 days a week service. The first three tasks are more or less the traditional reactive tasks the police force has to carry out. Policemen in patrol cars are assigned to handle emergency calls that come in via the central incident room or via the local police office. Policemen at the neighborhood team office take care of citizens who want to report a crime or ask a question. Depending on priorities some incidents from the emergency service and some crimes reported can result in criminal investigations. Information gathered during these three processes is entered into the police workflow system. Project work differs from criminal investigation because it aims at solving series of crimes instead of handling single cases. When a neighborhood team suffers from continuing youth annoyance or burglaries, resulting in repeated crimes on that topic, management can decide to start a project “youth annoyance” or “burglaries”. These projects can have a reactive (hunting down the criminals) or proactive (preventing future offenses) nature. The latter two tasks are pro-active tasks that aim at preventing criminality. Surveillance aims at being close to civilians to answer questions 139

COORDINATING DISTRIBUTED WORK and at being present in risk areas to deter potential criminals. Finally community policing involves the adoption of a part of the neighborhood by one policeman. He is the intermediary between the police force and the people in that part of the city. As such he learns to know citizens as well as criminals in this area and can sometimes relate incidents to suspects or can anticipate on crimes. For more detailed discussions of these tasks we refer to Vrolijk (2000) and Laere (2000, 2002a). In all these processes specialized expertise can be needed at any given time. In re-active work the policeman at the scene may not always be the expert. You never know whether an incident or notification will involve a youth suspect, a firearm, a burglary, drugs or whatever else. The policemen are all generalists: anyone should be able to deal with any problem, at any time, at any place. If the generalist needs help, he should consult the local expert or other sources where expertise can be found (for instance a procedure handbook about that topic). The main way to do this when you are out on the street is to contact the central contact person at the neighborhood team office who is responsible for walkietalkie traffic. He can call the person you need. When you are in the office you can directly phone the expert yourself. In the pro-active projects one of the group members can be the local ‘youth’ or ‘burglary’ professional. And he or she just knows or can ask his or her network of colleague experts if similar problems have been dealt with earlier. Also when surveillance is aiming at patrolling in a street where car burglaries are expected the expert responsible for car criminality can brief or join the surveillance team. Several problems involving the coordination of distributed expertise are identified and summarized in Table 47 below. The elements of our coordination framework are used to categorize the coordination problems. Table 47: Coordination problems for involvement of distributed expertise Group level problems COMPOSITION Variability and complexity of tasks Professional not always involved COMMITMENT Lack of involvement and commitment CONTACT Different shifts due to 24 hour service No transparency who is on duty Low percentage of e-mail users CONTENT Too much irrelevant information

140

Organizational level problems COMPETENCE Decreasing knowledge of experts No overviews of available expertise Limited ICT skills COHESIVENESS Shifting priorities Limited time for working on expertise CONNECTION Limited access to computer, intranet, e-mail Few options for informal coordination CONTEXT Generalist may not recognize the problem Generalist may not understand the procedures

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE Most of these problems are already discussed in chapter 4. Although generalists are confronted with tasks they cannot perform according to the desired quality standard, the involvement of expertise is perceived to be difficult. Local experts are often not available, handbooks are not available outside the office and may be outdated, and there is no transparency considering the current expertise of distributed professionals and about how they can be contacted. Besides the local experts there is no overview which policemen are members of what regional projects. Therefore it is very difficult to consult experts outside the own neighborhood team. There is a regional help desk for professional knowledge (with a strong focus on juridical expertise), but they are only consulted for highly complex matters. A general complaint is that their response time is too slow for assisting in the solution of day to day problems. On the contrary lots of information about new working procedures and know-how are perceived as overwhelming and irrelevant when they are received on a moment when they are not directly needed. Furthermore shifting priorities make it difficult to heavily invest in local expertise building. Also the strong focus on local duties make cross-neighborhood team collaboration hard to do. Finally the observation that generalists may not be able to link the actual problems to the available procedures and know-how stresses the need for involvement of an expert. 7.2.2 Opportunities for coordination of distributed work The opportunities presented in Table 48 arise from individual analysis by the researcher (March-April 1999), a group session (April 27th 1999) and some additional one to one discussions with police employees (April-May 1999). Table 48: Opportunities for improving consultation of distributed expertise Opportunity

Benefit

Cost / threat

Any time, any place participation (composition)

Better quality of work

Central record of employee skills (competence) Recording of goals and contributions (commitment) Specialized relationships (commitment) Best medium for task at hand (contact)

Faster matching of task and expertise Univocal objectives Clarity about duties and performance More efficient collaboration More productive communication

Disturbed experts Matching time Increased complexity Time and effort for compiling and updating Less freedom Big brother

Easy insight in other processes (contact) Fast sharing of large amounts of information (content)

Learning opportunities Understand larger whole More efficient communication

More interaction and context sharing with experts (contact/context)

Better interpretation of information that has been shared

Lack of social interaction Costs of drop outs Dependence on technology Purchase and maintenance costs Complexity of media choice Time spent on reading Information overload Loss of context Less time for interpretation Time for interaction Easily forgotten

141

COORDINATING DISTRIBUTED WORK Most of the opportunities identified are discussed in the general overview in chapter 5. The most prominent opportunities are those of any time, any place involvement and communication and the ability to select the best medium for the task at hand. However, the rise in flexibility of composition and contact also increases the complexity of selecting experts and media. Other difficulties may be expected considering motivation of experts to contribute to work outside their own neighborhood team and balancing the amount of information to be shared and the time needed for interpretation. 7.2.3 Design of coordination for distributed work processes In the neighborhood team’s work process we are confronted with policemen who sometimes need to consult colleagues or other information sources to achieve good quality of the work process. By one to one discussions with policemen (May-June 1999), by the researcher’s analysis (July and August 1999) and by a group session with the policemen involved (August 26th 1999) we have determined the steps this consultation process should consist of. These are: 1. be aware of the problem and define the problem; 2. decide to consult a colleague or other source; 3. contact colleague or source; 4. consult colleague or source and receive a set of alternatives; 5. choose solution from the set of alternatives; 6. apply the solution; 7. evaluate the solution and the advice of the colleague or source; 8. refresh the overview of best practices and overview of personal expertise. In the current situation some of these steps are little motivated or rewarded (2, 7), some are hard or impossible to do with regard to distributed expertise (3, 4), and some are not executed at all (8). Therefore directions for coordination are derived from the problems and opportunities identified with the help of the coordination framework and the design guidelines. In the overview in Table 49 these directions are related to both the steps of this consultation process as well as to the coordination framework and the guidelines of our design approach. Each of the coordination directions is discussed below. Coordination during the execution of the primary process Currently coordination of expertise relies too much on anticipatory coordination and pays too little attention to link available expertise to daily tasks and problems to be solved. ICT provides opportunities related to composition and contact to link supply and demand of knowledge and expertise just in time, because instantaneous data or voice connections can substitute traveling times. However, Guideline 1 and Guideline 8 remind us that there needs to be enough time to commit experts and align their thoughts, and to interpret the information gathered. Consequently distributed experts and expertise may be easy to contact, but possibly hard to commit or to comprehend. In such cases organizational members either take more time (by partly using anticipatory coordination or by holding up execution), or limit the consultation of distributed expertise to manageable amounts.

142

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE Table 49: Directions for consultation of distributed expertise Directions for coordination

Step

Element

Guideline

Focus on the coordination of skills and expertise during the execution of the primary process Let situational needs or “knowledge gaps” of policemen in the primary process guide group composition and information sharing Stimulate consultation and distributed collaboration by management and by culture Make sure that policemen on duty (and especially experts) can continuously be reached via ICT Utilize all knowledge available in the organization by consulting expertise via various media Selection of the best course of action should be done by the local generalist (problem context) and the expert (know-how context) together Always show progress and evaluate Motivate and reward updating of the central record of competencies

1, 2

Composition Contact Composition Content

1, 8

1, 2 2, 4 3 4 5 7 8

1, 10

Commitment Cohesiveness Contact Connection Composition Contact Context

3, 4, 6

Commitment Commitment Competence

6, 7 2

7, 9 1, 2, 8, 12 11, 12

Guidance by situational needs or “knowledge gaps” It is counterproductive to overwhelm generalists with all kinds of information they may never need. Instead their specific needs at a certain moment (depending on the problem to be solved, the generalists’ prior experience, and other situational circumstances) should guide which experts to involve, what information to share, and what media to use. An important deliverable in the first step of the consultation process is therefore the formulation of an “expertise-need”: What expertise does the generalist who needs to deal with the problem miss to solve it successfully? When information sharing considering know-how and expertise becomes more demand or needs oriented, there is caution for information overload (Guideline 10) and for too large work groups (Guideline 1). Stimulation of consultation and distributed collaboration Besides the fact that experts are difficult to locate and contact, there prove to be more barriers (like social and psychological) that cause that consultation does not happen often. Consultation needs to be motivated and rewarded actively, otherwise will not make the effort. Motivating and rewarding consultation and assistance can be done more or less formally (by management assessment) or more informally by development of a cooperative culture where consulting and assisting colleagues is common. From Guideline 4 can be concluded that it is wise to combine formal and informal rewards to let them reinforce each other. In addition to motivating the generalists to consult experts, experts need to be motivated to assist the generalists. First of all the experts or their neighborhood team managers need to balance their own priorities and the larger organizational priorities so that experts get more room to spend time on work for other neighborhood teams (Guideline 3). According to Guideline 4 assisting generalists should be incorporated in organizational culture as 143

COORDINATING DISTRIBUTED WORK well as formally acknowledged. However, Guideline 6 warns that experts should balance the need of assisting the generalist against their other duties. Assistance is thus advocated, but it is also limited. Guarantee that policemen can continuously be reached When a variety of media is available policemen, and especially the experts, can always be consulted when they are on duty. However, some caution is needed to prevent that continuous availability does not become a burden rather than a blessing. When experts are disturbed too often, or for questions which answers could easily have been found elsewhere, their productivity and motivation to help decreases tremendously. Generalists thus need to balance between the needs for autonomy of the experts and for interaction with the experts (Guideline 7). Furthermore continuous availability is only true when policemen are aware of all incoming messages (voice mail, e-mail etc.). There should be standards what media are suited for urgent and less urgent matters, or for emotional or complex matters (Guideline 9). Utilize all knowledge available in the organization ICT provides the opportunity of consulting multiple experts and other sources like databases without time and place constraints. Here the emphasis is on diverging to gather as much information on your expertise-need as possible, whereas the aim of formulating expertise-need is to converge and to limit group size and amount of information. When the expertise-need is adequately formulated this combination of convergence and divergence results in a focused in depth search for know-how and expertise. It can result in consulting electronic databases or in distributed collaboration between generalists at the scene and experts (for that specific problem) somewhere in the organization or a combination of both. By combining various media the benefits of different media (fast information sharing versus intensified interaction) are both utilized (Guideline 8). A first issue of concern is that there is a limit to gathering knowledge on your expertise need (Guideline 1, Guideline 10). Secondly database systems that contain procedures or best practices are not always efficiently usable by generalists. Often their inexperience with the topic heavily delays their searching speed or decreases the success rate of their searches. Therefore combining the database search with some pinpointing of an expert where the answer can be found could be a useful compromise (Guideline 12). Thereby the expert should not fall in the trap of completely explaining a simple procedure or best practice, as this would harm his productivity. Finally Guideline 2 reminds that in gathering knowledge for a problem to be solved, a generalist should look for a balance between proven search-paths and the most obvious experts to be consulted on the one hand and for unknown and illogical search options on the other hand. In that way solutions can sometimes be fast and according to procedures, and sometimes be innovative. Selection of the best course of action During the problem analysis it has been concluded that generalists may experience problems with understanding and applying the procedures and know-how they gather. On the other hand it is impossible for experts to list all clues about when certain elements of a 144

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE procedure of best practice need or need not to be applied (Guideline 12). Therefore it is recommended that experts at least play a minor role (agree or disagree on the final intended approach) in the choice for a particular course of action. They are more experienced in matching particular elements of best practices to particular characteristics of the problem. However, they need the generalist at the site to provide details about the context of the problem situation. As such, for less straightforward problems there needs to be some interaction between know-how context and problem context (Guideline 11). Always show progress and evaluate In order to speed up the own problem solving process and stimulate repeated assistance by experts in the future it is important to give them feedback on the effects of their advice. When experts see that their contributions are used and have profound impact they will be eager to contribute more to the process. Whereas in co-located groups the effect of their contribution is directly visible, distributed groups need to address feedback explicitly (Guideline 7). Also constructive criticism can motivate the expert to assist better in the future. The fact that a generalist invests time in criticizing on the advice shows that he is willing to support the expert in improving his consultation (Guideline 6). Motivate and reward updating of the central record of competencies Just like evaluation is important for commitment in future processes, updating the central record of competencies is crucial for effective future composition (Guideline 2). So, when a problem solving process is finished and everyone is looking forward to the next assignment explicit attention for preserving lessons learned is needed. When generalists have gained experience with a certain topic they can show this experience in order to be available for advice to less-experienced colleagues. Similarly, procedures and best practices need to be updated. As this is a bureaucratic and possibly dull stage of the collaboration process monitoring and assessment may be needed to assure execution. Although the design thus far gives a lot of clues how coordination of distributed expertise could be improved, the combined effects of the improvements are still unclear. For instance, if we would provide a “mobile office” for every policeman in town (connection), and for a central database with all skills and expertise of each policeman (competence) what would happen? Would coordination of skills and expertise and the quality of problem solving improve, or would factors like self-conceit, power of geographical supervisors, the police culture or difficulties in sharing context in distributed communication still impede distributed coordination? And if so, is it then really necessary to completely change the culture of the police force (considering showing of weaknesses and settlement processes) or are a set of attractive incentives also sufficient to achieve the preferred behavior? Other examples of design questions that were unanswered thus far are: How should participation on bulletin boards in this particular organization be managed, and what should be the content of intranet sites aimed at the coordination of skills and expertise? For the analysis of these combined effects we proceed to the next phase of our research approach, the development and playing of a simulation-game.

145

COORDINATING DISTRIBUTED WORK 7.2.4 Game design We recall from the discussion of our design approach in chapter 6 that the development of a simulation-game consists of the following 7 steps: 1. draw up a functional set of requirements; 2. analysis of processes and mechanisms; 3. design of an integral framework; 4. draw up a game-technical set of requirements; 5. construction of a prototype; 6. testing of the prototype; 7. final adjustments to improve the game. The execution and results of these steps are discussed below. Functional set of requirements The game serves two interests: • A design interest: we would like to explore to what extent our new forms of coordination of distributed work are successful, and how they can be improved. • A training interest: we would like to make policemen aware of the issue of coordinating distributed expertise and related problems as well as to guide them how to behave in future. In order to enable evaluation of the new working method and to provide learning opportunities for the policemen the following functional requirements for the simulationgame are defined: • Key element of the game is the consideration of a policeman whether he will or will not make use of expertise from outside his own neighborhood team to solve a daily work problem (composition/commitment). If so, a second key element is to determine how he can make use of this expertise productively (contact/content). • In the game more information- and communication technologies are available than in the daily work setting. Policemen can be reached by their colleagues any time, any place (connection). • Policemen are motivated to improve the coordination of skills and expertise (referring to the necessity of management support addressed under cohesiveness). The design of the new way of consultation is presented as an option. The policemen are stimulated to verify and experience whether this new working method brings the expected results. They are also free to suggest and try out other new coordination mechanisms. • At the end of the game the group of participants needs to conclude whether, or under what conditions, the new working method leads to improvement of performance. • To support this forming of opinion the quality of performance is measured and the participants receive continuous feedback on their performance. • The game should come close to reality to ease judgment whether the new working method works in reality.

146

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE •

Key actors in the game are the neighborhood teams (consisting of a mix of generalists and experts), a jury (who assesses the solutions to the problems) and the simulation-game administrators. Furthermore there is an environment panel who play the parts of organizational members not directly represented in the game. Thereby participants in the simulation-game have the opportunity to contact anybody within the police organization they might need.

Processes and mechanisms The second step, the analysis of processes and mechanisms has for a large part already been carried out in the first (section 7.2.1) and third phase (section 7.2.3) of our design approach. For the game the 8 steps of the consultation process (section 7.2.3) are reduced to 5 steps. The main reason for this is that this reduces the complexity of the simulationgame. The resulting five steps are: • formulate your expertise-need (1); • search for know-how or contact experts (2, 3, 4); • develop an approach for the problem (5); • apply the approach (6); • evaluate and give feedback (7, 8). The formulation of the expertise need makes game-participants aware of the need for consultation. The search for know-how or experts can be coordinated in various ways. In the game five options are offered: 1. rely on own expertise (when the generalist is familiar with the problem or when it is not that complex); 2. search on the intranet (when the generalist knows enough of the problem and the problem context to be able to locate the related know-how on the intranet and interpret it correctly); 3. contact knowledge center (when the generalist is under time pressure or knows too little about the problem area to locate know-how or experts himself, he can ask the knowledge center to locate the know-how or experts for him); 4. contact one or more experts (when the problem is too complex to be solved by only consulting the intranet); 5. put a question on the bulletin board (when the generalist wants to involve a larger group of colleagues, for instance for a multi-dimensional or highly complex problem or when a creative solution is desired). Of course the participants in the simulation-game can also choose to combine several of these expertise sources. Subsequently an approach has to be developed from the gathered information. In the game the approaches for the problems are not really applied. The assessment of the approach by the jury is seen as appliance of the approach. When the feedback from the jury is received the participants in the game need to pay attention to evaluation and related actions (like thanking the experts that assisted or suggesting adjustment of the know-how on the intranet).

147

COORDINATING DISTRIBUTED WORK To support the successful execution of these consultation steps, some tools for coordination are offered to the simulation-game participants. These are: • An overview of skills, phone numbers and e-mail addresses of experts in the police force; • a search-tool that can list all policemen from a particular expertise area or with a specific skill who are currently on duty; • the existing police intranet, with a roughly categorized overview of existing procedures and best practices; • a phone and an e-mail connection for each neighborhood team and shared bulletin boards (there are separate bulletin boards for each expertise area). Integral framework Figure 5 below shows the relations between the actors in the simulation-game. neighborhood team

administrator send assignments

assignment solve assignment

questions

answers

solved assignments

jury judge solutions

judgements

other team environment panel answer questions

judgements

Figure 5: Structure of the neighborhood team game The game administrators send assignments to each of the neighborhood teams. These teams consist of 2 or 3 policemen, who are a mix of generalists and experts. When developing an approach for an assignment, a neighborhood team could consult intranet or one of the other neighborhood teams or the environment panel. The environment panel also represents the knowledge center. When an approach is finished it is sent to the jury. The jury judges both the solution to the problem (content) as well as the coordination the neighborhood team has undertaken to arrive at the solution (process). Their assessments are sent both to the neighborhood teams and to the game administrators.

148

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE Game-technical set of requirements Although by now a clear picture has arisen of the basic structure and processes of the simulation-game, still a lot of lower level design choices need to be made. These lower level choices are discussed below. For a more elaborate motivation of each of these choices we refer to Vrolijk (2000). Caluwe (1996) states that gaming-simulation for organizational change asks for a strong similarity with the organizational reality. Duke (1974) stresses that participants normally have enough fantasy to accept less realistic elements of the game as reality. The content of the assignments and the coordination mechanisms in our game are very close to reality. The main reduction from reality is that assignments are not incidents that are really occurring, but descriptions of situations on paper. The motivation for this is that the focus and challenge of the simulation-game should be on locating, consulting and applying distributed expertise, and not on solving the problems. The neigborhood team game is kept as simple as possible in structure and interfaces. The administrators can vary complexity by sending more complex assignments. By varying the complexity the administrators can prevent that the game is too difficult in the beginning, and too simple after a while when the “tricks” have been discovered. In the invitation a short description of the objective and the content of the game is given. The invitation stresses that the aim of the game is not to test some new technological tools, but focuses on the question how some new but simple tools can be applied heedfully in order to improve the coordination of skills and expertise. The day starts with a briefing in which the current problems with coordination of skills and expertise are discussed, the new working method is presented, and the structure and functioning of the simulation-game is explained. Next there are two playing cycles. In both cycles the participants are invited to try out the new working method by developing approaches for a number of assignments in their neighborhood team. In a first debriefing between the two cycles all teams come together and share their learning experiences. Based on these experiences new learning objectives may be formulated for the next cycle (for instance if a particular aspect of the method did not get enough attention). The second cycle is similar to the first. After the second cycle a large debriefing follows. Besides the learning experiences during the second cycle the question whether and how this new working method should be implemented in the police organization is answered. This results in recommendations for police management (considering organizational prerequisites) and personal intentions (considering more frequent or different consultation of expertise). If possible, organizational recommendations are elaborated in a project group. The game can be played in a day or in half a day. It seems neither desirable nor productive to claim a group of policemen for a longer period of time. Literature strongly recommends that a simulation-game is played in one room. This enables control, manageability and face-to-face briefing and debriefing. However, a particular feature of our problem situation is that expertise is distributed over geographically dispersed teams. Therefore we would prefer to place the neighborhood teams in separate rooms. The expertise center at the head office of the police department 149

COORDINATING DISTRIBUTED WORK provided for a nice compromise. Some small rooms separated by glass walls served as team offices. In this way communication between teams was limited to phone and e-mail. However, for discussing astonishing experiences there was also room to quickly visit one another. Furthermore there was a central large room for briefing and debriefing. To make sure that during the debriefings attention can be given to each participant’s opinion, a maximum of about 15 participants (5 teams with 3 members) is recommended. Furthermore a minimum of 6 to 8 (3 or 4 teams with 2 members) is needed to assure enough dynamism and complexity. Although the use of computer in simulation-games is strongly discouraged by experienced game designers (Elgood, 1988; Caluwe, 1996) it is inevitable in our game. Computers are less flexible and may therefore inhibit participant’s desire to change game processes. Also, intensive focus on computer use can limit social interaction between the participants. Finally there is the danger of computer failure which may disturb the game or may make continuation of the game impossible. However, in our case the elements email communication, intranet consultation and bulletin board discussions are crucial elements of the new working method that require computer use. In addition we choose to use the computer for distributing assignments, gathering answers, distributing assessments and calculation of scores. In that way receiving assignments and assessments and sending solutions costs less effort for the neighborhood teams. That in turn leaves more time and effort for the distributed problem solving in between those processes, which should be the core of the simulation-game. Moreover the computer support for assignment, solution and score logistics provided big benefits for score calculation, immediacy of feedback and data collection for research purposes. Besides solutions for each assignment the participants register their expertiseneed, their coordination strategies and their evaluations in the computer. These data, categorized per team and per assignment, are available for later analysis. Furthermore three questionnaires are filled in. One considering the view on the current problems considering coordination of expertise is attached to the invitation. Participants are asked to send this one to the administrator before the date of the simulation-game. The second one, considering their opinion of the new working method beforehand, is handed out after the first briefing but before the first play cycle. Finally an evaluative questionnaire, considering their opinion about the new working method and related problems, is handed out and filled in after the last play cycle, but before the final debriefing. In this way the last questionnaire registers the individual opinions of participants before the group’s shared opinion influences them. Construction of the prototype For the development of the computer support for the game we choose to use MS Access ’97. The main reason for this is that this software is available on the police computer network and that it supports building of a relational database (team, assignment, approach, coordination strategy, assessment, score et cetera). During the simulation-game the participants can make use of the existing intranet (internet explorer 4) and e-mail 150

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE facilities (pegasus mail). As there is no bulletin board software available yet the bulletin board feature as well as an overview of experts and their skills are also developed in the MS Access environment. By interviewing some experts in the police organization we develop about 60 assignments varying in topic and complexity. For each assignment a “standard-solution” is described to support the jury in the assessment process. Testing of the prototype and final adjustments to improve the game In October 1999 5 members of the police organization test a first prototype of the simulation-game. On December 13th 1999 a trial-game is played with 10 policemen from neighborhood team Nieuwmarkt. The main questions for the game developers is whether the assignments are sufficiently realistic and challenging, and whether the MS Access interface is clear and easy to use. The first test is a very positive experience. Although the developers focus attention on “trying out all the buttons” the policemen soon start to really play the game. They are eager to find the best solutions or to tease each other (“sorry, our youth expert is currently sick at home, please try another team”). Furthermore it is promising that some mistakes in the Access design can be changed while the policemen are continuing to use the database. The policemen stress that there is no need for video films or role-playing to illustrate assignments. Policemen could easily use their vivid imagination. Adjustments to the original design are the simplification of some of the interfaces and changes in some of the assignments because there are flaws in the problem formulation. The trial-game results in two major changes. In the original design the emphasis in the first playing round is on experiencing the current situation (so the list with experts’ skills and the discussion list were not yet available). The idea is that a contrast between the first and second round would create learning opportunities. However, the policemen stress that they prefer to play the new situation in both rounds to make maximal use of the simulation-game. They argue that they already experience the current situation during their daily work, so that the desired contrast is already there. Furthermore we first put time constraints for some of the assignments to see what coordination strategy will be chosen under time pressure. The policemen argue that this will harm the key learning opportunity of the game: discovering the value of different coordination strategies. They emphasize that the absence of time pressure is a very valuable aspect of the game in comparison to daily practice. This convinced us of not putting time pressure on the assignments at all. After the game is played several times some more adjustments are made: • Besides a list with experts and their skills a search tool is added that could make selections from this list. Possible selections are experts on a specific topic, or experts who are available on a certain time. • The registration of coordination strategies is standardized to ease automatic calculations of average use and average score of a certain strategy.

151

COORDINATING DISTRIBUTED WORK •

A separate possibility for registering the consultations of the knowledge center is introduced. Thereby we get more insight in the content of these consultations, which are primarily done by phone.

7.2.5 Game results The neighborhood team game has been played 9 times: 1. On December 29th 1999, with 4 policemen from the neighborhood teams Prinsengracht, ’s Gravezandeplein, Diemen and Flierbosdreef. 2. On April 12th 2000, with 6 Firearm professionals from District 3. 3. On April 25th 2000, with 12 participants from the course Firearms. 4. On May 24th 2000, with 8 participants from the course Firearms. 5. On May 29th 2000, with 12 policemen from the neighborhood team Admiraal de Ruyterweg. 6. On June 7th 2000, with 10 policemen from District 4. 7. On September 19th 2000, with 9 policemen from District 9. 8. On January 25th 2001, with 12 policemen from neighborhood team Lijnbaansgracht. 9. On March 28th, with 8 policemen from district 1. In the remainder of this subsection we discuss the analysis of our data sources: the computer logs from the Access database, the outcomes of the briefing and debriefing discussions, and the results of the questionnaires. After that we draw some final conclusions combining the results from the various data sources. Briefing and preliminary questionnaire The briefings are both aimed at discussing how policemen currently consult distributed expertise and at explaining how the game will be played. The briefings take about 1 hour. When confronted with the research findings that consultation of distributed expertise is often problematic, policemen start to give examples of disappointing experiences. During these discussions the game-administrator asks for positive examples of consultation of distributed expertise. Furthermore results from the preliminary questionnaire (quality of work at our team is bad; intranet is used less than once a day) are used to stimulate discussion how participating policemen think about their current work performance. The results from the preliminary questionnaires do not really differ from the outcomes of prior research. Locating experts is perceived to be difficult and new media, like intranet of pmail, are marginally used. Overall the policemen confirm that coordination of distributed expertise needs to be improved. The open discussions, with lots of practical examples, create a relaxed atmosphere. The purpose of the game and the operation of the Access database are easily understood by the participants. Sometimes some minor questions are answered at the beginning of the first play round.

152

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE Figures Table 50 shows the number of assignments per game. On average 30 assignments are solved in each game. The figures on media use and scores below are based on these 261 assignments. Table 51 presents the usage of each coordination strategy. Table 50: Number of assignments Game Assignments

1 38

2 16

3 26

4 35

5 30

6 37

7 29

8 29

9 21

Total 261

6 73 13 0 16 3

7 59 41 21 21 7

8 83 45 3 28 3

9 78 39 7 19 4

Mean 29

Stdev 7.3

Table 51: Coordination strategy usage (in %) Strategy Own expertise Intranet Help desk Expert Bulletin board

1 76 24 0 16 0

2 44 56 6 38 0

3 54 42 8 27 8

4 77 46 0 6 0

5 47 73 17 57 3

Mean 66 42 7 25 3

Stdev 14.7 17.2 7.6 14.9 2.9

In 66% of the assignments policemen rely on their own expertise or the own expertise plays a distinctive role in the development of the solution. The most used coordination strategies to consult distributed expertise are searching on the intranet (42%) or consulting an individual expert on the topic (25%). The help desk and the bulletin board are only marginally used, despite the fact that the simulation-game was an experimental situation where exploring new avenues for consultation was strongly promoted. The figures vary considerably from game to game, resulting in high standard deviations. Still we can conclude that policemen rely more on own expertise than that they consult intranet (p = 0.019; student t-test, one tail, matched pairs, α = 0.05). Also intranet is more often consulted than experts (p = 0.001), and experts are more often consulted than the help desk (p = 0.001) and the bulletin board (p = 0.001). The difference between help desk consultation and bulletin board use is not significant (p = 0.055). From these figures we conclude that: • The great differences in strategy use between games reflect that some groups of policemen have more explored or adopted the new coordination strategies than other groups. Learning experiences and level of adoption may therefore differ strongly from group to group. • The ranking order of the most used coordination strategies exactly reflects the ranking order of strategies with the least psychological costs. Relying on own expertise and intranet consultation are individual strategies that do not force you to show in public that you cannot solve your problems on your own. Consulting an individual expert or the help desk show your ‘inability’ to only a few people. The bulletin board has the highest psychological costs. This may imply that the fear of showing your weaknesses is a very big inhibitor (Ackerman, 1994; 1998). 153

COORDINATING DISTRIBUTED WORK •

Policemen still rely mainly on their own expertise (which is not wrong at all), but the figures for intranet and expert consultation are much higher than in current practice (where it is almost zero). This implies that policemen have at least explored and possibly adopted these new coordination strategies.

Next Table 52 shows the average scores on assignments for each coordination strategy. Table 52: Score per coordination strategy (on a scale from 1 to 5) Strategy Own ex. Intranet Help d. Expert Bul. Bo. TOTAL

1 3.48 3.67 3.67

2 4.14 3.44 5.00 4.17

3.55

4.00

3 4.07 4.55 3.50 4.00 3.50 4.19

4 3.78 4.63 4.50 3.94

5 4.14 4.23 4.60 4.12 4.00 4.23

6 3.30 4.18 4.33 4.00 3.54

7 3.47 3.75 3.33 3.33 3.00 3.41

8 3.38 4.00 5.00 4.5 5.00 3.52

9 3.27 4.17 3.00 3.74 1.00 3.23

M 3.67 4.07 4.07 4.04 3.42 3.73

SD 0.37 0.40 0.90 0.40 1.36 0.36

The best average score is reached with the coordination strategies of intranet and help desk consultation (4.07), closely followed by the average score of expert consultation (4.04). The average scores when relying on own expertise (3.67) and when using the bulletin board (3.42) are lower than the overall average score (3.73). For the help desk and the bulletin board this figure suffers from higher standard deviation, which may partly be due to fewer observations. The scores for consulting intranet or another expert are significantly higher than when relying on own expertise (respectively p = 0.024 and p = 0.026; student t-test, one tail, matched pairs, α = 0.05). The score for help desk consultation is not significantly higher than relying on own expertise (p = 0.19). From these figures can be concluded that: • It seems to pay to consult distributed expertise. Consultation of intranet and experts outperform reliance on own expertise. • Intranet and experts are more reliable sources than the help desk and bulletin board. For the last two the average scores fluctuate dramatically. To gain insight whether the participating policemen improve their consultation strategies during the game we compare the average coordination strategy use and score per coordination strategy from the first and second play round of each simulation-game. We also look for differences between games 1-4 where most participants are experienced policemen (who could possibly rely more on their own expertise) and simulation-games 5-9 where we focused more on the required mixture of inexperienced generalists and experienced professionals. However, for both comparisons no significant differences in coordination strategy use or score are found.

154

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE Debriefing and evaluative questionnaire In all games participating policemen are enthusiastic about the practical opportunity to explore and discuss new ways of working. This confirms our expectations that gamingsimulation is a useful technique for the police organization, as policemen are “doers” (Vrolijk, 2000). Often the simulation-game is indicated as “close to reality”, and the assignments are perceived to be recognizable. During all games there is a positive atmosphere. There is a lot of joy and laughter about successes and failures, but there are also fierce discussions about the quality of solutions or the benefits and drawbacks of certain coordination strategies. Overall the game is judged as fun and instructive. Only in a few cases there is some loss of concentration by some players at the end of the afternoon. When this is sensed by the administrator the debriefing is started. Apart from the trial-game we experience no technical problems with the Access database or the other ICT (intranet, e-mail, phone). Policemen who play the game are positive about the new opportunities for coordinating. Both in the debriefing and in the evaluative questionnaire they indicate that policemen more often need to consult distributed expertise. However, they also identify problems with regard to consulting distributed expertise: • Considering the lack of insight in what expert has what expertise a first problem is that not every Firearm professional has the same type of knowledge. Some know more about firearm identification, others know more about legislation, but on paper their function is the same. So the function name is not sufficient to select the most suited expert. Secondly, for a single problem several people have part of the solution. So often more people need to be located and contacted (each time with reasonable difficulty) to arrive at a complete solution. • Intranet is currently seldom used for consulting professional knowledge. Most policemen are not aware of the professional knowledge available on intranet. Lots of them only visit the page of their own neighborhood team, the page with “most wanted criminals” and the page with advertisements. • Although intranet is already implemented a lot of policemen do not master the basic skills to navigate on intranet. They state that they would make more use of the intranet if there had been a course to teach them those skills. A clear outcome of the simulation-game is that numerous policemen “discovered” the large amount of professional knowledge on intranet and state that they certainly will make more use of intranet in their daily work from now on. • Policemen are reluctant to bother colleagues with their problems. Their basic strategy is first trying to search for the needed knowledge on their own. Only if they do not succeed they contact a colleague. This confirms the earlier observation that there is a culture of “not showing your weaknesses” and it also explains why searching on intranet is the most used coordination strategy despite the drawbacks of intranet that will be discussed below.

155

COORDINATING DISTRIBUTED WORK When the policemen are asked for their main observations after exploring the new consultation process the following comments are given: • Few problems are experienced with regard to the formulation of the expertise need. Some teams took this step a little too easily, but directly experienced feedback by a lower score. • The most problematic step is not surprisingly the second step (search for know-how or contact experts). The benefits and drawbacks of the different strategies and media in this step are discussed hereafter. • Another problematic step is the evaluation and feedback. Although everybody agrees on the necessity and importance of this step it is often forgotten (after the job is finished, or in game-terms … after the points have been earned). During the evaluative discussions it is for instance stressed that one cannot expect regional projects to improve their intranet pages if they never receive any comments on them. • A last important consideration is the need for management support and cultural norms and values to stimulate consultation. Both in the questionnaires and in the concluding discussions these issues score high. Although the policemen value the demand-based way of coordinating distributed expertise, they stress the need for stimulating this new behavior. Without motivating factors the tendency to hide weaknesses and the hurry to finish the job will exceed the importance of consulting distributed expertise and improving quality of work. Policemen mention that current unavailability of most options on the street need not be an excuse. In a lot of cases asking the own office by walkie-talkie to search for certain information on intranet or to contact an expert could be valuable. The proposed new coordination strategies receive constructive criticisms during the debriefings. Each of them is discussed below. Many policemen experience frustration while consulting the intranet. They complain about the badly performing search engine, the inconvenient structure, and the longwinded or incomplete texts. A noticeable remark is the repeated request for a better search engine. There is a very strong belief that a better search engine will solve all problems of locating the needed expertise. This belief is grounded in the experiences with internet. However, most policemen are not aware of the consequences of the big differences between internet and intranet for the search behavior. For instance, their police intranet is much, much smaller which makes it easier to locate information by navigating through the basic structure. In contrast, most policemen start almost immediately by using the search engine. In addition, the intranet structure is not ideal. The regional projects have developed their intranet pages quite independently. Therefore common elements are placed under different headings or located in different areas. For the neighborhood team policemen it means that every time he visits another topic he has to rediscover the complete structure of this knowledge area which is very time consuming. Finally a lot of the regional project’s intranet pages are too long and too general. Policemen desire more examples of practical cases where general procedures or laws are applied. With regard to consultation of the central help desk and individual experts there is a strong desire that the central help desk is available 24 hours a day. Especially during the 156

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE nights it is hard to locate and contact an expert as fewer people are on duty. Furthermore policemen like the idea of a search-tool that helps to select the best expert for the job who is currently on duty, but they make a sharp distinction between expertise information and contact information. They stress that expertise information should be specific and detailed: not only topic Firearms, but a break down in knives, firearm legislation, firearms in catering, firearms at schools, firearm technical recognition et cetera. Furthermore each policeman himself needs to actively maintain these data. In contrast contact information can be rough (night shift / morning shift / afternoon shift), and should require minimal maintenance. Preferably the duty-lists are automatically integrated in the electronic phone book. Especially the experts in the game have experienced the double role a neighborhood team has. Sometimes they consult others and sometimes others consult them. Only twice an expert expresses that he is disturbed heavily in his own work, but they believe that in practice this will not be the case so often, as the amount of assignments per time interval is not that much as in the simulation-game. From the questionnaires it follows that policemen do not have problems to consult unknown experts, or to assist unknown colleagues. The bulletin board is received with great reservation as may be clear from the marginal use percentages (see Table 51, page 153). The main reasons for this are that the policemen experience too much uncertainty with regard to the speed and reliability of the answers. However, despite the rare use policemen stress that the bulletin board can become a valuable alternative for consulting distributed expertise under the following conditions: • Distinguish between questions needing a quick and reliable answer and more broad discussions focused on exploring opinions or gathering and developing innovative ideas. • Central control is needed to guarantee quick and reliable answers. Furthermore measures should be taken to have youth related questions answered by youth professionals. A broadly valued point of this coordination alternative is that policemen who are not directly participating in solving the problem can learn a lot by just reading the questions and answers. For other strategies the consultation process is not visible for policemen that are not directly participating. Final conclusions From these results we conclude that both the training- and design objectives are met. By playing the simulation-game policemen become aware of the need of consulting distributed expertise, and they experience and learn how new technologies can assist them in consulting distributed expertise during daily problem solving. Furthermore the game results and evaluative discussions revealed numerous points of attention and nuances that can ease the eventual implementation of the new working method. The outcomes of these 9 gaming-simulations and the preceding design project result in numerous detailed recommendations for the police organization (Laere, 2000; 2002a). Some of the identified shortcomings with regard to consultation of intranet and bulletin board are tackled the second action research described in section 7.3 hereafter.

157

COORDINATING DISTRIBUTED WORK

7.3 Electronic interaction in the regional projects One of the most important outcomes of the first action research is that the content and structure of the regional projects’ intranet sites are not linked to the questions that neighborhood team policemen have in daily practice. Furthermore it is not clear how quality and speed of answers on the bulletin board need to be guaranteed. It is obvious that the regional coordinator of a certain topic cannot perform both these tasks himself. Often he is not an in-depth expert on the topic, but rather a good coordinator and manager. Besides he cannot look after the questions and answers on the bulletin board 24 hours a day, 7 days a week. Consequently it seems that the distributed experts working at the neighborhood teams, who are members of the regional project, need to play a role in maintaining the intranet site and guaranteeing answering speed and quality on the bulletin board. As such the tasks of maintaining the intranet site and managing the bulletin board can be seen as a work process in which distributed actors need to coordinate their work. Therefore a second action research is initiated that focuses on the electronic offering of expertise by means of intranet and a bulletin board. Again the design approach articulated in chapter 6 is applied. The subsections below address respectively the current way of working, related problems and opportunities, the proposed new way of working and the design and results of a simulation-game. For more detailed discussions of these steps and intermediate results we refer to Blom and Korf (2001) and van Laere (2002a). 7.3.1 Work processes and coordination problems In the previous action research the process of consultation and related problems are discussed and analyzed from the perspective of the neighborhood team, the requesting party. Here, we focus on the role of the regional project, the party that provides expertise. All 20 regional projects generally have the same objective and tasks. However, differences in their history, the characteristics of their topic (preventive or repressive, little or much collaboration with external parties, and little or high priority on neighborhood teams) and the personality of the regional coordinator have caused a big variety in the way the tasks are executed. The research is limited to one of the objectives of the regional projects, namely “increasing the know-how and expertise of policemen on a particular topic”. For this duty the following tasks need to be performed: 1. Keep up with developments in their discipline; 2. Developing new know-how; 3. Deliver or gather new know-how for publication; 4. Validation and verification of know-how and classification of reliability; 5. Publishing and distributing know-how; 6. Answering questions; 7. Judge reactions to distributed know-how; 8. Adjustment and updating of earlier distributed know-how; 9. Monitoring whether all previous tasks are carried out. 158

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE In between tasks 6 and 7 neighborhood teams apply the know-how as has been studied in the previous action research. The regional project coordinator is involved in almost all tasks. Experts are only strongly involved in tasks 2 and 3 (development and collection of new know-how) and in task 4 and 5 (distribution of know-how and answering questions at their local level). However, in the second case study in chapter 4 is observed that these contributions to the regional project are sometimes overruled by local duties at the neighborhood team. Finally the regional project coordinator has often arranged some assistants to help with the publishing process. Know-how can be published in official procedures (enforced by top management), handbooks for a certain topic, newsletters, the internal police force newspaper, or announced in regional project meetings or during daily briefings at neighborhood teams. Normally several of these media are combined as some have a more official status (procedure, handbook) and some are more aimed at promotion (newspaper, meetings, and briefings). In addition the intranet is used by 10 of the 20 regional projects. For answering questions most regional coordinators can be contacted when the local expert is not available. Only a few of them have a 24-hour, 7 days a week pager service. Finally feedback on usability of distributed knowledge is preliminary received during the monthly regional meetings or during visits to neighborhood teams. Common problems are that: • execution of too many tasks in this process depends on the regional project coordinator, who cannot perform these tasks on his own, 24 hours a day, 7 days a week; • police management is reluctant to assign supporting staff to the regional projects, as they prefer investing in policemen that work civilian-needs oriented (in neighborhood teams); • experts in the neighborhood teams are not allowed or may not be motivated to perform tasks for the regional project, as they are rewarded for the work they do for their own neighborhood team; • experts who do not work regularly on problems of their own expertise are losing experience and thus losing value for the regional project; • there is little horizontal coordination worsened by a low attendance at regional meetings, a lack of insight in who are the current experts per neighborhood team, and because only a few policemen regularly use e-mail and intranet; • the content of the expertise supplied may not be specific or profound enough to meet the neighborhood teams’ needs, but improving this situation is hard as there are few opportunities to get feedback on usability; • the large differences between regional projects cause that some are easy to consult by phone or intranet, and some are not. In addition, the arrangement and classification of similar topics on the various intranet sites of the regional projects may differ, which complicates searching for the neighborhood team policemen. • finally, when reliability of know-how is unclear generalist policemen may not use it as they strongly rely on official and proven methods to ensure juridical prosecution.

159

COORDINATING DISTRIBUTED WORK These problems are related to our coordination framework in Table 53 below. Table 53: Coordination problems for electronic offering of expertise Group level problems COMPOSITION Too much dependence on few actors COMMITMENT Lack of commitment of experts CONTACT Little horizontal coordination Low attendance at regional meetings Few experts use p-mail CONTENT Not aimed at needs neighborhood teams Classifications difficult to comprehend Little feedback on comprehension and usability of know-how

Organizational level problems COMPETENCE Experts loose experience Overview of experts not up to date COHESIVENESS Autonomy of neighborhood teams Priority of civilian related duties CONNECTION Most media not available when experts are working on the street CONTEXT Reliability of certain know-how may be unclear Little insight in neighborhood teams’ needs

Improving the consultation process, as described in section 7.2, may solve part of these problems. This will lead to more consultation and better feedback on needs. However, as explained earlier, it is questionable whether the regional projects can quickly and reliably answer the growing amount of consultations when the regional coordinator needs to handle them all alone. Therefore the key challenge seems to be how to involve distributed experts in more tasks. 7.3.2 Opportunities for distributed coordination Table 54 shows the opportunities for improving electronic offering of expertise. These opportunities are generated during participatory observation of work and coordination practices and several one to one discussions with policemen (October and November 2000). Roughly summarized the challenge of quickly responding to consultations is related to opportunities within the composition and contact elements of our coordination framework. ICT provides opportunities to involve more people, and to contact them by more or faster means. Similarly, the challenge of supplying reliable and profound knowledge is related to the elements of commitment and content. To guarantee reliable answers experts need to be strongly involved and motivated to provide the best answer for a requesting colleague. Also, explicit knowledge (handbooks, intranet, bulletin board) is only reliable when the author is known as a reliable source. Finally one of the key-issues considering content is the balance between profundity and concision.

160

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE Table 54: Opportunities for improving electronic offering of expertise Opportunity

Benefit

Cost / threat

Any time, any place participation (composition)

Central record of employee skills (competence) Recording goals and contributions (commitment) Monitoring and coaching (commitment) Specialized relationships (commitment) Show progress of distributed tasks (commitment) Synchronous and asynchronous distance communication (contact) Best medium for task at hand (contact)

More actors available Experts exploited both centrally and locally Local experts keep expertise up to date Faster matching of task and expertise Clarity about duties and performance Performance visible for everyone More efficient collaboration Stimulates participation and speed of work Contacts consume less time Expert always in reach More productive communication

Experts can spend less time on local goals Matching time Increased complexity Dependence on technology Time and effort for compiling and updating Less freedom

Corporate electronic memory (content)

Knowledge always accessible

More interaction and concern for context sharing (contact/context)

Better interpretation of information that has been shared

Time and effort for monitoring Big brother Lack of social interaction Costs of drop outs Time and effort to show it Danger of forgetting to show it Lack of awareness Delay or absence of feedback Dependence on technology Costs of purchasing and maintenance Complexity of media choice Bad indexing Loss of context No direct feedback on usability Time for interaction Context sharing easily forgotten

Most of the benefits, costs and threats listed in Table 54 have already been discussed in the general overview in chapter 5. Related to the process of offering electronic expertise the regional coordinator can take advantage of the composition opportunities by assigning more tasks to experts supported by an up to date overview of their skills. Also there are numerous ways to increase the currently lacking commitment of the experts. Making expected contributions explicit and showing progress and performance combined with active coaching could result in efficient collaboration with sufficient mutual support. However these interventions should be balanced with the threats of too little social interaction and decreasing autonomy (big brother). Furthermore the intranet and the bulletin board increase the options to select the best medium for the task at hand, but the suppliers of expertise need to be aware of the strengths and weaknesses of synchronous and asynchronous media. Both in the intranet and in the bulletin board interaction between suppliers and users of knowledge and expertise is needed to arrive at a smooth matching process between request and supply and to guarantee the proper application of general knowledge in a situated context.

161

COORDINATING DISTRIBUTED WORK 7.3.3 Design of coordination for distributed work processes The opportunities above are developed into a design of a future work process for the electronic offering of expertise. This is achieved by organizing a computer supported group discussion session (November 27th 2000), by one to one discussions with keyactors in the work process (January and February 2001), and by discussing the intermediate and summarized results during two presentations (January 22nd 2001 and April 10th 2001). First tasks and actors are identified for the work processes of offering expertise via intranet and via a bulletin board (see Table 55). The tasks are a differentiation of the currently executed tasks described in section 7.3.1. For a detailed explanation of all tasks and actors we refer to Blom and Korf (2001). Here we shortly discuss the noticeable ones. Given the experienced importance of quality control some tasks address validation and verification of intranet content or judgment of the quality of questions and answers on the bulletin board. In both cases this control can be executed before and after publishing, thereby balancing the dilemma of speed versus quality. Furthermore the general task of adjustment and updating is in the case of intranet split in adjusting structure and adjusting content. These tasks need not to be performed by the same actor(s). Table 55: Different divisions of tasks amongst actors

14 15

Tasks intranet Keep up with developments Develop new know-how Collect new know-how Verify before publishing Publish know-how Verify after publishing Judge reactions Update structure/hyperlinks Update content of intranet Monitor previous tasks Tasks bulletin board Ask questions Categorize questions Judge received questions and link to earlier answers Answer questions Judge answers to questions

16 17 18

Match question to expert Pass on FAQ's to intranet Delete old questions

1 2 3 4 5 6 7 8 9 10 11 12 13

162

Division of tasks amongst actors Recommended Centralized Decentralized Core group RPC All experts Core group RPC All experts All experts Core group All users Core group RPC All experts Central expert Adm. Empl. All users Core group RPC All experts RPC / expert Adm. Empl. Core group Central expert Adm. Empl. All users Central expert Adm. Empl. All users RPC RPC RPC Recommended Centralized Decentralized All users All users All users All users All users All users Moderator Moderator Core group All users Core group After posting Moderator Moderator Moderator

All users Core group Before posting Moderator Moderator Moderator

All users Core group After posting Core group Core group Core group

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE

When the RPC wants to assign certain tasks to other policemen there are more choices than just asking all experts on his topic that are working at the neighborhood teams. First of all he could ask police management to assign an administrative employee or one of the experts of the neighborhood teams to him in order to guarantee full time (about 40 hours a week) central support. Secondly, he could try to establish a core group of experts who keep working at their neighborhood teams, but who are part of their time responsible for region wide offering of expertise. Of course their neighborhood team managers need to give them room for these additional tasks. Finally some of the bulletin board tasks could be assigned to a central moderator of the communications department. The choice which actor performs what task is guided by matching task characteristics (reliability, readability, comprehensibility, usefulness, timeliness of intranet information and questions and answers on bulletin board) and actor characteristics (like topic expertise, large network of relations, written and verbal communication skills, computer skills, manageability, time available). For all tasks three different actors are suggested in Table 55 to stress that regional projects have the freedom to choose for the recommended or for a more centralized or decentralized actor for each task. Directions for coordination are formulated to cope with the largest tensions between the different alternatives. The coordination directions motivate the choice for the recommended actors and highlight the benefits and drawbacks of the more centralized and decentralized alternatives. Furthermore they help to identify preconditions for effective task execution. The coordination directions are presented in Table 56 and discussed below. Benefits and drawbacks of all choices are more elaborately discussed in Blom and Korf (2001). Table 56: Directions for coordination of electronic offering of expertise Directions for coordination All relevant knowledge needs to be mapped quickly and orderly on the intranet site Everybody should be able to bring up any additional knowledge for the intranet site Knowledge on the intranet needs to be reliable

Task 1, 2

5, 8, 9

Element Composition Commitment Commitment Content Contact Content Content

1, 3, 8, 9, 11, 12 1, 11, 12

The structure of the intranet requires uniformity as well as room for diversity Execution of intranet tasks need to be controlled Balance answering speed and quality on the bulletin board Progress and participation on the bulletin board need to be monitored

7, 10

Commitment

1, 6, 11

14, 15

Composition Contact Commitment

1, 3, 8

3 4, 6

13, 16 17, 18

Guideline 1, 3, 6, 10 4, 6, 11

1, 5, 11

163

COORDINATING DISTRIBUTED WORK Map all relevant knowledge quickly and orderly The first set of tasks (1 and 2) includes keeping up with new developments and developing new expertise. Important requirements for these tasks are a large number of internal and external relations (network) and topic expertise (to determine quality and relevance). The core group is proposed as a compromise between involving all experts and involving no experts at all (in accordance with Guideline 1). The RPC has too many tasks to perform and may lack sufficient expertise. Assigning the task to all decentralized experts makes the group unmanageable. A prerequisite is that participation of these experts is granted by their neighborhood team supervisors. Therefore experts are not claimed full time, as their expertise would then be lost for daily problem solving in the neighborhood team. These experts do not only have experience in their own discipline, but are also experienced generalists, and thus a valuable resource for their neighborhood team. (see Guideline 3). The challenge for the experts who are member of the core group is to distribute time over their local and regional duties. For both processes they have to be eager to contribute, but also critical to distinguish between valuable contributions and time wasting (Guideline 6). Finally, when gathering information for the intranet site the core group needs to balance comprehensiveness and briefness in accordance with Guideline 10. Stimulate bringing up of additional knowledge Making the core group responsible for scanning developments and developing innovations guarantees a minimum input of knowledge for the intranet site. But of course local experts should be motivated to bring up their own experiences (task 3) that are difficult to track by the central RPC or core group. The main problem is in commitment rather than in contact. Currently RPCs need to beg for contributions. Neighborhood team supervisors can also stimulate this behavior. Following Guideline 4 it would be wise to incorporate these central duties in decentralized performance evaluations (formal motivation) and in war stories (social motivation). A final point of interest is the need for experts that send in their experience to add relevant contextual information (within what circumstances their experience is or is not valid) as is illustrated by Guideline 11. Guarantee reliable knowledge on the intranet site The strong attention for quality control (task 4 and 6) follows from the requirement that information on the intranet needs to be reliable. Besides topic expertise insight in the needs of the generalist policemen and handling speed are valuable actor characteristics. The core group is here the most obvious actor as these experts have both topic expertise and are still engaged in daily work at the neighborhood teams (Guideline 3). Furthermore they are still a manageable group that can guarantee a reasonable throughput time (Guideline 1). To facilitate judging reliability and usefulness for generalist policemen some other guidelines need to be addressed. First the core group needs to be able to adjust communication media to different needs (fast information sharing versus interaction, Guideline 8). Furthermore this flexible use of media requires shared communication standards to prevent misunderstandings and delayed work processes (Guideline 9). Reliability and quality of information also depends on whether context is clearly connected to content on the intranet (Guideline 11) and whether assistance is provided for 164

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE generalists that cannot locate the expertise they need themselves (Guideline 12). Otherwise the information on intranet is reliable, but its application to enhance quality of daily problem-solving is not guaranteed. Leave room for diversity in the uniform intranet structure Considering the publishing and updating of information on the intranet site (tasks 5, 8 and 9) one needs to balance uniformity versus diversity. The regional projects have different needs given the different nature of their topics. However, generalist policemen searching for information indicate that a uniform structure and presentation would ease their searching dramatically. Therefore the actor that performs this task should have topic expertise and master general communication skills. Appointing a central expert for each regional project is believed to serve these requirements best. To arrive at a coherent intranet site one publisher is more suited than a group of publishers such as the core group. They would need to coordinate too much about how to structure the site and about who will adjust what considering updating (Guideline 1). On the other hand a central administrative employee, who could for instance be responsible for maintaining several intranet sites of different regional projects, would lack the topic expertise to provide additional context where needed (Guideline 11) or to anticipate search difficulties (Guideline 12). Monitor performance and task execution on intranet For the evaluation of feedback on the intranet site (task 7) and the monitoring whether all tasks are carried out the main requirement for an actor was a person that had a central overview (manageability, Guideline 1). In this case the RPC seems the most logical choice. For problems with execution of tasks (task 10) Guideline 4 indicates that control can be organized formally and socially. Considering evaluation of feedback (task 7) Guideline 11 points at the importance of interpreting comments in their context and thus requesting for sufficient context with comments to judge them adequately. Here a central topic expert may be more valuable than the RPC to interpret the context. The advantage of a central administrative employee is that he has an overview over different sites of several regional projects and consequently can distribute valuable general comments to other regional projects. Balance answering speed and quality The interest in quality of bulletin board answers is acknowledged in the verification tasks (tasks 14 and 15). As a consequence the core group is here the preferred actor because they master more topic expertise than the moderator does (Guideline 1). In addition they can also easily anticipate on generalists’ needs as they themselves have a dual duty (Guideline 3). A major point of discussion is the consideration whether answers need to be controlled before or after publishing. Although policemen commenting on the design agree on the danger of publishing and consequently using an unverified answer, quite a few people point at the benefits of quick answers and the danger of big delays because of heavy controlling before publishing. To optimize both quality and speed verifying after posting is recommended, whereby verified and non-verified answers are clearly distinguished. This combines the different advantages of ICT (Guideline 8), namely fast 165

COORDINATING DISTRIBUTED WORK information sharing (non-verified answer directly visible) and time for interaction (anybody can comment on a wrong or incomplete answer). Monitor progress and participation on bulletin board continuously In contrast with the intranet monitoring the bulletin board involves much more guarding the answering speed rather than just the quality of the information and persuading actors to contribute. Therefore the dominant requirement here is continuous attention for the bulletin board, which can impossibly be expected of a RPC. A central moderator seems more suited to perform this task as he or she does not suffer from the coordination burden of for instance a core group of experts, who have to arrange by schedules who watches over the bulletin board (Guideline 1). Also, if the members of the core group decided to watch over the bulletin board all the time, this would either end up in a waste of resources or in complex schedules to coordinate who is on duty (Guideline 5). In conclusion the thread of the design discussion is the balance between speed and quality. Opportunities in composition and contact increase the reach and speed of offering of expertise. But simultaneously the increased handling speed results in a more critical dependency on strong commitment of actors and good understanding of the content shared. Only when these efficiency and effectiveness criteria can be optimized both, productivity will really grow. From this discussion it is hard to tell whether the developed paradoxical directions for coordination are applicable for policemen in daily practice. Also, it is unclear whether they will prefer the recommended actors to perform the tasks, or whether they will choose for some centralized or decentralized options for certain tasks. Consequently, we again feel the need to further explore and fine-tune the design in a simulation-game. 7.3.4 Game design In this subsection we describe the execution and results of the following 7 steps: 1. draw up a functional set of requirements; 2. analysis of processes and mechanisms; 3. design of an integral framework; 4. draw up a game-technical set of requirements; 5. construction of a prototype; 6. testing of the prototype; 7. final adjustments to improve the game. Functional set of requirements The game needs to serve a design interest (explore the feasibility of the new ways of coordinating), and a training interest (create awareness and guide future behavior). The following functional requirements are defined: • Key element of the game is the challenge how the members of the regional project need to organize their work in such a way that relevant knowledge is clearly and timely presented on the intranet site and that questions on the bulletin board are quickly and adequately answered.

166

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE • •

• • • •

In the game more information- and communication technologies are available than in the daily work setting. Policemen can be reached by their colleagues any time, any place. Policemen who are a member of the regional projects are motivated to improve the electronic offering of expertise, referring to the requirements that followed from the neighborhood team games. The new way of coordinating is presented as a possible alternative. The policemen are stimulated to verify and experience whether this new working method brings the expected results. They are also free to suggest and try out other new coordination mechanisms. At the end of the game the group of participants needs to conclude whether, or under what conditions, the new working method leads to improvement of performance. To support this forming of opinion the quality of performance is measured and the participants receive continuous feedback on their performance. The simulation-game should come close to reality to ease judgment whether the new working method would work in reality. Therefore there needs to be a possibility to really alter (a copy of) the intranet site of the regional project. Key actors in the game are the regional project coordinators and their central assistants, decentralized experts on the neighborhood teams and the simulation-game administrators. The central administrative assistants and the central moderator are included as pseudo roles. Furthermore each regional project can decide whether or not to appoint a core group of experts.

Processes and mechanisms To prevent the game from becoming too complex the 18 tasks (Table 55, page 162) are clustered into 5 key tasks. These key tasks are: 1. verify changes on the intranet site before publishing (task 4); 2. publishing on the intranet (task 5); 3. verify changes on the intranet site after publishing (task 6, 7, 8, 9); 4. moderating the bulletin board (tasks 13, 16, 17, 18); 5. verifying answers on the bulletin board (tasks 14 and 15). For each of these 5 tasks the game participants need to decide before they start playing who is going to carry out what task. In between play rounds each regional project can reflect on their experiences and choose for a different division of work and a different way of coordination in the next play round. For the other tasks (tasks 1, 2, 3, 10, 11 and 12) there are fewer uncertainties who can best perform them. Knowledge that could be relevant for the intranet site is inserted in the game by the game administration (task 2, task 3). The regional project coordinator needs to be aware of new developments that are published in the newspaper (task 1) and can be assisted by his experts or even other policemen in the game. Furthermore the regional project coordinator is responsible for overall monitoring (task 10) and all participants in the simulation-game can ask questions or add answers to the bulletin board (task 11 and 12).

167

COORDINATING DISTRIBUTED WORK We shortly elaborate how the tasks and the related deliberations concerning coordination are modeled in the simulation-game. Participants are introduced to the choices in the briefing and receive a sheet with options for each of the five tasks and the main benefits and drawbacks of each alternative division of labor. For detailed discussion of the deliberations involved we refer to the previous subsection. The RPC, the core group or all experts can do the verification of changes. In the first case the RPC can directly publish the ideas that come from newspaper or other sources and actually this means that there is no verification at all. In the other cases the RPC needs to consult a proposal for change with one member of the core group by e-mail, several members of the core group by e-mail, orally with the whole core group, or with several arbitrary experts by e-mail. The main dilemma is that of quality versus speed. Publishing on intranet can be done by a central expert belonging to the regional project, or by a central administrative employee (who takes care of the intranet sites of several regional projects). The latter is modeled as a pseudo role: the central experts from the involved regional projects become members of a central administrative unit (geographically separated from their regional coordinator). To assure their “lack of expertise” they are not allowed to publish knowledge or do alternations on their own intranet site. As a consequence all regional projects have to make the same choice for this task. The dilemma involves balancing the need for topic expertise versus the need for standardization of the structure of intranet sites. The verification after publishing can be done by the RPC, the core group or all experts in the Police Force. Here one needs to balance difficulties of commitment of a large group versus the limited availability of a single actor. Moderation of the bulletin board can be done by a central moderator (pseudo role of the game administrators), by the RPC, by the core group, or not at all. The core group can decide to monitor the bulletin board continuously with all members, or they can design a schedule stating who will monitor it within particular time frames. Considering the verification of answers on the bulletin board two choices need to be made. First answers can be verified before or after posting. In the first case answers are only visible after an expert has judged them to be adequate. In the second case answers are directly visible for everybody and experts can judge and alter them as soon as possible. In both cases the core group can decide to monitor with all members continuously, or to arrange a schedule for monitoring. Besides inserting knowledge into the game (via newspaper, personal e-mails, comments on the intranet site or questions and answers on the bulletin board) the game administrators guide the assessment process by sending assessment requests to neighborhood team members to judge how other regional projects maintain their intranet and bulletin board. These assessment requests refer to concrete examples of (not) published knowledge and (not) answered questions.

168

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE Integral framework Figure 6 shows the resulting information flows between the various actors in the simulation-game. For the decentralized publishing alternative the central expert can directly change the intranet. For centralized publishing changing the intranet requires interference by the central administrative unit. new knowledge comments on intranet questions answers checking of answers

bulletin board

change assignment

Regional project (RPC and central expert)

comments on intranet comments on change-suggestions comments on changes

central administrative unit questions wrong answers change questions change answers checking of answers

changesuggestions changes

intranet Game adminstration

Neighborhoodteam (core team, decentralized experts) judgements new knowledge request for judgement general judgement

score

Figure 6: Integral framework of Regional Projects Game 169

COORDINATING DISTRIBUTED WORK Game-technical set of requirements Considerable effort has been dedicated to formulate realistic knowledge developments, intranet comments and bulletin board questions and answers. Secondly, an important feature of the simulation-game is that alternations can be made to (a copy of) the current intranet and that there is a really working bulletin board available. Complexity has been kept as low as possible by limiting the amount of tasks where coordination can be varied distinctively to 5. However, from the beginning it has been clear that this second game is more complicated than the first. Another difference with the first simulation-game is that besides briefing, playing rounds and debriefing there are smaller planning and reflection rounds just before a play round where members of regional projects decide how to organize their coordination in the next play round. To make sure that debriefings are manageable the maximum amount of participants is 15. This results in 3 regional projects consisting of 5 actors (a regional project coordinator, a central expert and 3 experts that work decentralized on a neighborhood team). This is also the minimum, as 3 regional projects are needed to provide enough dynamism and complexity. The actors are distributed over a limited set of computers (where they can access e-mail, intranet and the bulletin board) to resemble their limited availability in practice (see Table 57 below). For instance, 2 experts sharing a computer on a neighborhood team have to negotiate when each of them can devote time to his regional project. As a result 4 neighborhood teams (with dispersed members of the core group) as well as 3 regional project offices (with regional coordinator and central expert) are present for the decentralized publishing option. For the centralized publishing option there are 4 neighborhood teams, a shared office for the three regional coordinators and two central offices for the central webmasters. Table 57: Distribution of actors over computers Comp. 1 2 3 4 5 6 7 8

Decentralized publishing RPC A + central expert A RPC B + central expert B RPC B + central expert C Expert A + Expert B Expert A + Expert C Expert B + Expert C Expert A + expert B + expert C Game administrators

Centralized publishing Admin. A (for B and C changes) Admin. B + C (for A changes) RPC A + RPC B + RPC C Expert A + Expert B Expert A + Expert C Expert B + Expert C Expert A + Expert B + Expert C Game administrators

Finally a set of indicators is developed for each actor to give them insight what tasks need to be addressed first. Examples are “amount of intranet comments not judged yet”, “amount of proposals for intranet change not yet verified”, “amount of questions unanswered”, “amount of answers not yet verified” and “amount of assessment-requests not yet distributed”.

170

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE Construction of the prototype Again we use MS Access ’97 for building the main structure of the simulation-game. The Access database supports: • The distribution of “new knowledge” by newspaper or by personal mail; • commenting on the intranet sites of a particular regional project; • discussion about proposals for adjustment and change between the actors that verify knowledge and the actors that publish knowledge; • posting of questions and answers on the bulletin board as well as verification of answers; • sending e-mail messages to other players; • and the routing of assessment-requests and assessments and the calculation of scores. In addition a copy of the police intranet is available. By means of MS Frontpage the players that have publishing tasks can change the intranet sites. These changes are immediately visible for the other players. The Access database consists of different menus for different players. Given the login name of the player behind the PC the forms are dedicated to that user. This can for instance imply that some optional buttons are not available, but also permissions on the bulletin board may differ (for instance an expert Firearms is allowed to approve answers on the bulletin board considering Fire-Arms, but not those considering Youth). Furthermore the game administrators can enter what division of labor has been chosen by each regional project and these modifications can also influence the appearance of players’ menus and their permissions. For instance the choice whether answers are verified before or after publishing regulates when these answers appear on the screen for different users. By interviewing some experts in the police organization and by studying police archives we develop numerous knowledge developments, intranet comments, bulletin board questions and (false) answers for each topic area of the participating regional projects. Although players in the simulation-game are stimulated to enter their own comments and questions, the game administrators can maintain a constant stream of inputs for each regional project. For each new knowledge development, intranet comment or bulletin board question an assessment request is automatically generated that can be forwarded by the game administrators to the neighborhood team that needs to assess the performance in reaction to this input. For prepared inputs “standard solutions” are added to support the assessors. Testing of the prototype and final adjustments to improve the game During development of the game continuous testing of the developed forms and functions results in incremental improvements of the overall structure and functioning of the prototype. This is done in numerous meetings between the database developer and the researcher.

171

COORDINATING DISTRIBUTED WORK In addition the simulation-game is tested twice on the police network. Both the interactions with the Access database and with the intranet-copy by means of MS Frontpage function adequately. Only the score calculation proves to take too much time. By extending the updating interval this problem is solved. 7.3.5 Game results The regional project game is played two times: • On August 24th 2001 with the regional projects, Amazone (AM, hard core criminal youth), Foreigners (FO), Firearms (FA); • And on August 27th 2001 with the regional projects Extreme harassment (EH), Burglary protection (BP) and Environment (EN). We discuss the analysis of our data sources: the computer logs from the Access database, the outcomes of the briefing and debriefing discussions, and the results of the questionnaires. After that we draw some final conclusions combining the results from the various data sources. Briefing The briefing sessions take about 1 hour, just like the briefings during the simulation-game for the neighborhood teams. The problems of monitoring quality and speed of knowledge sharing via intranet and a bulletin board are easily recognized by the participating policemen. Some of the policemen indicate that they are a little bit overwhelmed by the possible options for labor division and by the complexity of the simulation-game. Also, the Access-interface proves to be less user-friendly than during the first simulation-game. A major complaint is that it is sometimes not clear what other actors can see or cannot see. Despite these undesirable IT-troubles the participating policemen express that the simulation-games are instructive. Figures For each regional project the following data are available: • the labor divisions chosen during the simulation-game; • the amount of new knowledge developments, intranet comments and bulletin board questions (input); • the amount of proposed changes for the intranet site, comments received on change proposals and changes carried out (output intranet); • the amount of answers, verified answers, unanswered questions and unverified answers (output bulletin board); • the scores for intranet and bulletin board performance. The regional projects do not really experiment with the different divisions of labor available. Most of them stick to their first choice. Apparently there is more need to explore this single way of working than to experiment with different ones. Only in the second simulation-game the regional projects decide to shift from decentralized publishing to centralized publishing.

172

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE Table 58: Labor divisions chosen Task Verify before publishing Publish

AM RPC Expert

FO Core (1 agree) Expert

Verify after publishing Moderate

RPC

Core

Admin

Verify answers

Core after Cont.

Core Sched. Core Before Cont.

Regional project FA EH Core Core (1 agree) (FTF) Expert Expert/ Admin RPC RPC RPC

RPC

Core Before Cont.

Core Before Cont.

BP Core (1 agree) Expert/ Admin All exp.

EN Core (1 agree) Expert/ Admin Core

Core Cont. Core Before Cont.

Core Sched. Core Before Sched.

Furthermore we can conclude that the regional projects chose for quite different options to organize their work process of electronic knowledge supply. Considering the verification of intranet changes before and after publishing some rely on centralized control (RPC or face-to-face meeting of core group), whereas others rely on more decentralized control (one member of core group has to agree, monitoring afterwards by core group or even by all experts). The options where all members of the core group have to agree electronically beforehand (or even 6 experts) are judged as not practical and too dangerous for throughput time. For the moderation all options (central administrative moderator, RPC and core group continuously or following a schedule) are used. Only the complete abundance of this task is not selected. With regard to the verification of answers on the bulletin board most regional projects rely on control before posting. Most of them prefer continuous attention of the core group to scheduled attention. The figures shown in Table 59 are discussed for each regional project separately, but some general comments need to be made beforehand. First (s)low performance (for instance few changes to the intranet site) does not necessarily mean bad performance, but can also mean much learning. Some of the participating regional projects already have a quite sophisticated intranet site and are thus already used to (quickly) adjusting it. Others do not have an intranet site at all and really start from scratch in the game. As a consequence more deliberation is needed to make adjustments. Also, some changes on intranet require much work, while others involve just a few mouse clicks. Secondly, there are problems with the assessments. Experts in neighborhood teams are so busy commenting on and questioning other regional projects, and maintaining their own intranet site and answering and verifying questions on their bulletin board, that they lack time to assess performances of other projects. In addition it proves to be difficult to figure out what the reaction of a regional project is to a particular comment on their intranet site. For instance, if a comment is wrongful, the explanation of the regional project is not 173

COORDINATING DISTRIBUTED WORK visible and there is no change on the intranet site either, which can lead to a bad assessment. As a result the scores severely lag behind performance and as such there is no direct feedback on performance. Although the researcher perceives this as a major shortcoming, the participating policemen argue that the indicators (amount of intranet comments not acted upon, amount of questions unanswered and amount of answers not yet verified) do provide direct feedback on performance. Furthermore, undesirable changes on intranet and badly answered questions prove to result in new comments on intranet or on new questions. In conclusion the scores are not very valuable for our analysis. Instead we give a rough value judgment of the performance of each regional project based on analysis of the data after playing the game (intranet quality and speed and bulletin board quality and speed in Table 59). Thirdly, given the fact that participants are already overwhelmed by dealing with intranet comments and bulletin board questions, the game administrators insert only a few knowledge developments. At the beginning of each game intranet comments and questions on the bulletin board are mainly contributed by the game administrators, but very soon all participants start to post comments and questions. A low number of proposals does not necessarily mean that few proposals are done. As these figures reflect the situation at the end of the game, early proposals can have been turned into changes executed. Table 59: Value of indicators at the end of the simulation-game Indicator # knowledge developments # intranet comments # intranet comments not acted upon # proposals without comments # proposals with comments # change assignments # comments on proposals / assignments # changes realized Intranet quality Intranet speed # questions # answers # questions not answered # answers not verified Bulletin board quality Bulletin board speed

AM 0 12 0 0 0 0 0 8 +/++ 12 14 0 0 + +

FO 1 16 0 3 1 4 2 1 +/18 23 1 6 ++ --

Regional project FA EH BP 1 0 0 13 3 7 0 1 0 0 0 0 0 0 2 0 0 0 11 2 6 13 2 5 + + + ++ +/+ 15 6 9 19 9 22 0 0 0 3 0 0 + + + + ++

EN 0 5 0 2 0 1 0 2 + 17 24 3 18 + --

The Amazone project chooses for rather centralized options considering labor division. This is understandable from the fact that their approach to youth criminals relies on a centralized information position and centrally coordinated action. As such the RPC keeps 174

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE a strong influence on the content of the intranet site. To prevent that the RPC drowns in his duties the bulletin board tasks are handed over to a central moderator and a core group that continuously verifies the answers. In the indicators this centralized labor division is reflected in quick handling speed with regard to intranet changes. On the other hand some comments of users are ignored which can indicate that there is a lack of attention for needs of the policemen that use the information from the intranet site. Also on the bulletin board questions are quickly answered and verified, even without noticeable interventions from the central moderator. Answers on the bulletin board are to the point and clear. Only once there is a discussion on the bulletin boards considering vague directions for action in one of the answers. In contrast to the regional project Amazone the two other regional projects in the first game (FO and FA) choose for a more decentralized labor division. Their performance is quite different. The regional project FO has difficulties with the maintenance of both the intranet site and the bulletin board. With regard to the intranet site the RPC and publishing expert receive few comments on their proposals for change. Also 4 proposals and 4 assignments are still in progress and not yet executed. Within the regional project FA the RPC and publishing expert are more profoundly describing their proposed changes and explicitly stating what points they doubt. As a result they get more and more detailed comments from their distributed experts. This is a clear link with Guideline 10 (provide sufficient content). The relation is more productive because the central employees direct the decentralized experts in what information they would like to receive as comments. Furthermore it is noticeable that the regional project AM changes the intranet without any comments from their distributed experts, whereas the regional project FA receives 11 comments on 13 changes. Looking at the regional projects AM and FA shows that both ways of working can be productive. Looking at regional projects FA and FO shows that less attention for commitment or content can result in a less productive work process. Both regional projects FA and FO prove to have problems in keeping up with the incoming questions on the bulletin board. At the end of the simulation-game both have answers that are not yet verified. Given their choice that answers need to be verified before they become visible, this means that these answers are not visible for policemen who might be waiting for an answer. In contrast the regional project AM has verified all the incoming answers, while answers from everybody are immediately visible at their bulletin board. Possibly this immediate visibility urges them to verify more quickly. On the other hand they also had less incoming answers from unknowns sources than the others. Another reason that especially the regional project FO is a little behind in verification of the answers (and changing their intranet site) may be that their topic is quite complex which makes the verification more time consuming. During the second game the regional project EH organizes their intranet maintenance process with regular face-to-face meetings to discuss changes. The major motivation for this is that they do not have an intranet page in reality and that the first moves require 175

COORDINATING DISTRIBUTED WORK intensive interaction between experts close to the work floor and the central RPC. This behavior resembles the often-argued media use that groups in their early stages require rich media. Clearly their inexperience or their time consuming communication results in a slow rate of changes. For the remainder their performance is sufficient. The regional project BP has received 6 comments on 5 executed changes and 2 proposals. Again it is observed that comments from distributed experts are more valuable when the proposals are more clearly formulated and are asking for specific feedback. Also the discussion on the bulletin board of this regional project shows that one question can gather multiple answers. As such it requires much effort from the core group to verify all these answers. Just like in other regional projects the verifying experts need to correct inaccurate answers now and then. Finally the regional project EN encounters great difficulties in coordinating their distributed work process. With regard to intranet they do not receive a single comment on the few proposals for change they do, although their labor division does aim for it. Also they fail to verify the answers on their bulletin board on a regular basis, which results in a large amount of answers not visible for the remaining policemen. Partly this is caused by a lack of ICT skills in combination with an indistinct interface. But in addition one of the decentralized experts says that he does not feel engaged and that he perceived that his contributions are not valued. So apparently there are also commitment problems that are not addressed by the RPC EN. Debriefing and evaluative questionnaire First of all the regional projects focus on quality of information and quality of answers to questions rather than on speed (content, composition). For instance, answers are most of the time verified before they become visible, despite the longer answering time. The danger that a policeman uses the wrong answer immediately, and does not (or too late) notice the correction that is made afterwards, is perceived as more important than the decrease in handling speed. However, from the figures can be concluded that verification speed differs largely between regional projects. For the neighborhood teams, who ask for both quality and speed, this is unacceptable. So quick verification needs to be guaranteed. In order to make it easier for policemen to find information on their site, regional projects are prepared to standardize the structure of their sites (content). In addition they are willing to make the information more readable and applicable (content). They do not believe or are not willing to remove “for neighborhood teams irrelevant information” from their sites, although policemen commented in the other game that too much irrelevant information complicates searching and influences future consulting (content). Consequently it seems that active commenting of neighborhood team members on the intranet sites of the regional projects is required to force them to pay more attention to their needs (contact, content). In the end some regional projects prefer to have their own web-master, some prefer a central web-master who edits their intranet site for them (composition). Both options should be supported to answer these different needs. However, good agreements on the 176

CHAPTER 7: FUTURE COORDINATION AT THE AMSTERDAM POLICE FORCE standard structure between the different web-masters are necessary to guarantee a concerted representation for users. During the two games it proves that technology is a necessary tool in distributed settings, but that user-unfriendliness and unfamiliarity can be great inhibitors (connection). Uncertainty how RPC’s need to coordinate with their distributed experts (the e-mail program Microsoft Outlook was not yet available and some of its properties had to be imitated in Access) gave rise to misunderstandings in communication. The simulationgame can be improved by replacing Outlook features for Access features when possible, but these experiences stress the importance of developing functional, but not too complicated IT facilities. Furthermore the need for diversity of media is confirmed (connection). E-mail, telephone and face-to-face conversation are used by different regional projects for the same task with satisfactory results. This shows that employees need to have the opportunity to select a medium that fits their situation best (contact). Although assigning a single task to a group of distributed employees (the core group) complicates mutual adjustment, this option has been chosen frequently for several tasks (composition). The benefits (continuous occupation, larger contribution of expertise) are valued more than the ease of making one person responsible. This confirms the identified need for distributed work processes. In some regional projects distributed experts decrease their contributions when they do not feel acknowledged (commitment). In some other regional projects the RPC is actively checking by phone whether e-mails had had the intended effect and whether the group as a whole was still one-minded. It seems that the latter regional projects are more productive with regard to speed as well as quality. In contrast a remarkable outcome of both the questionnaires and the debriefings is strong agreement that experts do not need to be explicitly awarded for their contributions: “It is just their job”. Finally, the communication between the RPC’s and the central administrative employees about desired changes on their intranet page prove to be complicated (contact, content). It is very difficult for RPC’s to pinpoint at what position on their site they wanted to change something. The most straightforward solution used for this in the game is faxing a copy of the old intranet page and indicating with pen where the adjustments need to be made. Final conclusions In conclusion can be said that the regional project game is more valuable from a design interest perspective than from a training perspective. The most important lessons learned are that regional projects and neighborhood teams agree on the concern for quality, but do not agree on the concern for speed. Furthermore the core group is a desirable actor, but can be counterproductive when commitment is not addressed. Contact can be established effectively in various ways, and media choice can thus be optional to suit situational and personal needs. Finally addressing content explicitly clearly improves productivity. These lessons learned have resulted in numerous recommendations for the Amsterdam Police Force (Laere, 2002a; 2002b). Reflections on the action research studies are presented in the next chapter. 177

8 Reflection In this chapter we reflect on our research in three ways. First we discuss the application of our design approach in the two action research studies described in chapter 7. Some issues with regard to the usability and usefulness of the different elements of our design approach (coordination framework, paradoxical guidelines and simulation-games) are discussed. Secondly, we reflect on our stay in the host organization, the Amsterdam Police Force. The experiences during our case studies and action research studies and views on some other existent coordination issues in the police force are integrated into a summary of coordination challenges and avenues for change for the Amsterdam Police Force. Thirdly, we widen the scope of the discussion to a reflection on our research objective and elaborate on the question whether our design approach is also useful for other service organizations. We discuss six cases from literature. These cases are service organizations that are redesigning their coordination of distributed work. We dwell upon hypothetical contributions of our design approach for these cases, and show similarities between coordination challenges in these service organizations and in the Amsterdam Police Force.

8.1 Reflection on the application of the design approach In chapter 7 is shown how we applied the design approach in two action research studies. The main results produced during each step of our design approach are summarized, concluding with some recommendations for implementation of new ways of coordinating in the police force. Here, we show the value of the different elements of our design approach, as experienced during these two action research studies, by discussing some noticeable observations. The coordination framework structures the design discussions The coordination framework discerns between aspects of coordination on the organizational level and on the process level on four dimensions: composition/ competence, commitment/cohesiveness, contact/connection and content/context. This conception proves to be helpful to circumvent misunderstandings around the ambiguous notion of coordination. As already observed in chapter 1 and chapter 4, people have different conceptions of what coordination is. This is also the case in the design teams where policemen and researchers collaborate during the action research studies. 179

COORDINATING DISTRIBUTED WORK Some participants tend to focus heavily on composition aspects of coordination; some of them stress contact or content. Commitment issues are often implicitly mentioned, but are most of the time overlooked when coordination is defined explicitly. The framework helps to create a common understanding of the notion of coordination and forces the participants to pay equal attention to all 8 aspects of coordination in the different design phases. This is experienced as valuable to arrive at robust coordination alternatives and implementation proposals. Similarly, the coordination framework helps to structure the briefing and debriefing sessions of the simulation-games. During briefing a common view on coordination can easily be reached by presenting and explaining the framework. Game participants are encouraged to pay attention to all 4 coordination dimensions when playing the simulation-game. Finally during debriefing sessions discussions are often started with questions like “What are your experiences considering the dimension of contact/connection?”. As such the coordination framework helps to address all aspects of coordination in a comprehensible overview and prevents that issues are overlooked during evaluation. The framework helps to identify opportunities, threats, benefits and costs Given the general opportunities and threats listed for each coordination element in chapter 5, opportunities for the coordination issue under study can easily be generated. Also, possible costs and benefits related to each opportunity can quickly be identified. Thereby the coordination framework helps to gather the relevant information in an early stage of the design process. With the overview of opportunities, costs and benefits (see Table 48 and Table 54) alternative coordination mechanisms can be developed. The overview prevents that possible coordination mechanisms are overlooked, or that their value is overestimated or underestimated when certain costs or benefits are not taken into account. Guidelines can be tailored to the coordination issue under study During the action research studies it proves that the general guidelines for designing coordination of distributed work become more recognizable when they are reformulated in terms of the coordination issue under study. Also different general guidelines can be integrated into one tailored guideline. For example, in the first action research the direction for coordination “Let situational needs or “knowledge gaps” of policemen in the primary process guide group composition and information sharing” reflects issues touched upon in Guideline 1 (few/many group members) and Guideline 10 (little/much content). In addition, it proves that not all guidelines are always as important, although most guidelines are used in both action research studies. For instance, in the first action research study guidelines related to commitment are of minor relevance, whereas in the second action research study there is less need to address guidelines related to contact.

180

CHAPTER 8: REFLECTION Positive IT impact for composition/contact, negative for commitment/content In general the tension between the opportunities of composition and contact versus the threats of a lack of commitment and poorly understood content is apparent in both action research studies. For instance it is easy to contact an expert to assist in problems solving, but it is more difficult to gain his commitment to really share responsibility for solving the problem. Also, information on intranet may be continuously available, but locating the needed information without assistance by an expert or applying the gained knowledge without further hints how to do it in this particular case is experienced to be more difficult. In the second action research study the manageability and guaranteed commitment of a single actor (the RPC) is often weighed against the continuous availability and cumulated expertise of a group of actors (the core group). Furthermore the dilemma of quality (verifying content) versus speed (quick contact) is especially apparent in the moderation of the bulletin board. Gaming-simulation provides insight whether/when benefits outweigh costs Where the coordination framework helps to identify costs and benefits, it does not make clear whether benefits of a developed coordination mechanism do outweigh the costs or not, or when they do and when they do not. In the games the interplay of organizational, technological, social and behavioral issues becomes tangible, and consequently the combined impact of unlike costs and benefits are observable. Some coordination mechanisms are judged to be invaluable because costs outweigh benefits most of the time. For example, policemen are reluctant to register their availability, as unexpected changes are often occurring and overviews would easily become erratic. In contrast, the bulletin board is also assessed negatively in the first game, but here comments are given by participants under what conditions benefits would outweigh costs. Despite content and commitment problems, the developed coordination mechanisms for coordinating distributed work are judged as valuable in both studies. Participating policemen clearly conclude that neighborhood teams should more often consult distributed expertise than currently is done, and that in regional projects some tasks currently done by the RPC should be assigned to a core group of distributed experts. Unilateral guidelines are not adequate for guiding coordination Especially during design in the first action research it proves to be tempting to choose for unilateral guidelines. In that study the current way of working can be characterized as strongly relying on own expertise. As such the logical guideline would be: always consult intranet or an expert. However, policemen in the first simulation-game comment that reliance on own expertise is sometimes more productive than consulting an external source (although figures from the Access database shown in Table 52 suggest that intranet and consultation significantly outperform reliance on own expertise). Consequently game participants agree that they should more often consult external sources, but they refute 181

COORDINATING DISTRIBUTED WORK unilateral directions for coordination like “always consult intranet or an expert”, “always do a house search when a suspect has a fire arm in possession”. During the action research studies the importance of dilemmas and paradoxes becomes more and more clear. In the second action research the guidelines are more applied as balancing guidelines than as unilateral guidelines. This is a learning experience for the designers and the researcher. The issue of paradox in improving coordination of distributed work is not stressed at the start of the games. The participants are provided with the dimensions of the coordination framework and a rough overview of costs and benefits of possible avenues they can explore in the simulation-game. During the debriefings, when the participants have experienced the avenues of improvement, the question is raised what avenues are perceived as worthwhile and what not. The discussions show that there is no “best avenue” for all neighborhood teams / regional projects but that several ways of coordination can all produce satisfying results. For example, in the first game different neighborhood teams solve the same problem in various ways (relying on own expertise, consulting intranet, consulting expert) with comparable results. In the second game the regional projects of FA and AM are both successful in managing their intranet site, although they choose for quite different labor divisions. Thus, the issue is not to look for the best coordination mechanism and rely on it permanently, but rather to critically monitor the costs and benefits involved in a coordination mechanism, and switch to or add other mechanisms when benefits are too low or when costs are too high. Sometimes it is wise to involve more people to create innovative viewpoints (composition), sometimes it is better to aim for a small group and a quick decision (commitment). Sometimes it is beneficial to quickly consult a standard procedure on intranet (contact), sometimes it is needed to take more time for tailoring a best practice to the current problem (context). Consequently it is not possible to give hard indicators for productive coordination like “the amount of contacts per week” or so. Such figures can be helpful in a particular situation for a certain time (like when there are few interactions due to a lack of commitment and initiatives to increase commitment need to be evaluated). In another case a distributed work group may benefit from decreasing its interactions if that would lead to more efficient coordination. The game administrators who lead the debriefing emphasize this awareness for paradox as a clarification and as something to be remembered around the issue of coordination of distributed work. Policemen also argue that different coordination mechanisms should exist concurrently to be able to profit from both their benefits. For example, they suggest that centralized publishing and decentralized publishing can coexist. This is an example of a ‘dilemmaresistant-arrangement’. Comments during simulation-games are more specific than during design A striking observation during the debriefing discussions of the games is that game participants comment much more specifically on the new coordination mechanisms than 182

CHAPTER 8: REFLECTION the members of the design teams. For example, game participants stress the need for uniform navigation structures on intranet by giving examples of information they cannot find because different names are used for similar issues in different regional projects. These examples are much more compelling than the more general need for uniform navigation structures. Also, the regional project coordinators, who acknowledge the importance of balance between quality and speed during design, are much more emphasizing quality control after playing the game, thereby referring to answers given by non-experts that are judged as false. Finally it appears during the game that a lot of prejudices with regard to the bulletin boards are still present, whereas the design teams are rather enthusiastic about this option. In both action research studies game participants underpin their opinion with numerous practical experiences from the simulation-game, whereas the members of the design teams are much more speculating about possible effects. This shows the value of practicing the new coordination mechanisms, rather than only discussing them. In combination with the conclusion that the simulation-games are a valid representation of problem solving and knowledge supply in daily police work we can conclude that the outcomes of the simulation-games are valuable to inform the implementation of the new coordination mechanisms. Multiple goals are achieved in the simulation-games The main objective of our simulation-games is to gather information how our design of future coordination of distributed work processes can be improved. The aim of practicing and evaluating the new coordination mechanisms is to fine-tune them, and to eliminate design flaws before implementation of the work process and supporting communication media. Numerous practical implications for the design are collected during the action research studies (see conclusions at the end of sections 7.2 and 7.3). Examples are the need to specify skills of experts more detailed in the overview of skills, the need for uniform navigation structures on intranet, the need for standard procedures as well as illustrations of best practices on intranet, the need to guarantee quality and speeds of answers on the bulletin board, the need to appoint core groups of decentralized experts and the need to stimulate neighborhood team members to comment on inadequate and missing information on the intranet sites. The games do not only provide insights for the designers but also have a learning-impact on the participants. Many policemen develop or improve skills like locating expertise, giving feedback, effective searching and publishing on the intranet and coping with feedback issues in asynchronous interactions. Furthermore several policemen express that they are planning to make use of the new coordination mechanisms in daily practice (like more often consulting the intranet on the neighborhood team, or providing for a electronic bulletin board for their regional project). We have not conducted any additional interviews or surveys to measure post-participation effects. First this is not the primary focus of our research (the game is mainly an instrument to feed and improve the design, rather than to train policemen). Secondly we anticipate that the absence of some prerequisites for change (IT infrastructure, integral rather than local control) will still strongly impede adoption of the lessons learned in daily behavior. 183

COORDINATING DISTRIBUTED WORK From a research perspective the games also help to reflect on our way of thinking. Outcomes of the evaluative discussions provide examples to improve the phrasing and underpin the value of the paradoxical guidelines. We observe that these multiple goals (informing design, training policemen, reflecting on research) do not interfere. Computer use in the simulation-game makes logistics and analysis efficient The use of computers in gaming-simulation is strongly discouraged by experienced game designers (Elgood, 1988; Caluwe, 1996) Computers are perceived as less flexible and may therefore inhibit participant’s desire to change game processes. Also, intensive focus on computer use can limit social interaction between the participants. Finally there is the danger of computer failure which may disturb the simulation-game or may make continuation impossible. However, in our case the elements e-mail communication, intranet consultation and bulletin board discussions are crucial elements of the new working method that require computer use. Consequently computer use in our simulationgame is inevitable. In addition we choose to use the computer for distributing assignments, gathering answers, distributing assessments and calculation of scores. We observed that receiving assignments and assessments and sending solutions costs less effort for game participants and game administrators because of computer support. Moreover the computer support provides big benefits for score calculation, immediacy of feedback and data collection for research purposes. Whether computers inhibit flexibility of the simulation-game depends strongly on the further design of the game. Although teams work in separate rooms, it is easy to cross the hall and visit other teams. The game administrators do neither stimulate nor prohibit this. In general participants stay in their own team to simulate the distance between them. Occasionally, participants visit the jury to discuss judgments they doubt, or they visit a team that does not respond to mail and whose telephone number is engaged. Also participants prove to be inventive to insert a fax opportunity in the game (printing a page and walking to another team to deliver it without further conversation). Limiting the chance on computer failures can be achieved by elaborate testing of the simulation-game beforehand. Furthermore it is wise to have all assignments on paper to be able to switch to a paper-based game if needed. In our case this would however not help as intranet and the bulletin board are key elements of the simulation-game. We conclude that computer support for assignment logistics and score calculation can be encouraged rather than discouraged. Of course potential drawbacks of computer use need to be taken into account. Elements of the first game are reusable in the second one Finally we observe that there are strong similarities between the two simulation-games considering their structure. Both games make use of six to seven workstations, key information shared are assignments, judgements and scores. Also communication technologies like telephones, the police-intranet, the bulletin board and e-mail are similar in both games. We expect that it is quite easy to reconfigure the simulation-game for another service organization. Basically the structure of the game can stay as it is. Only the assignments and standard-judgments and the actors involved in the game need to be reformulated. Note that this does not imply that it is easy to conduct a complete action-

184

CHAPTER 8: REFLECTION research in another service organization. The identification of coordination problems and opportunities and the design of coordination mechanisms that need to be explored during the simulation-game (distribution of actors over the workstations, their possible interactions, contextual conditions like management control) may require thorough research before redesign of the simulation-game.

8.2 Reflection on coordination in the Amsterdam Police Force During our research we spend over 4 years in the Amsterdam Police Force. Besides our research activities we participate in several informal and formal discussions on the future of police work and related organizational dilemmas. Here we reflect on our research in the police force, discuss the major challenges the police organization is facing and give suggestions how they can be tackled. Understanding coordination problems: from a rational to an integrated view Our analysis of coordination problems in the Amsterdam Police Force starts with the question how the performance of the regional projects can be improved. When we arrive the police force’s coordination problems are formulated as a lack of communication between the distributed professionals. In terms of our coordination framework, earlier studies into the performance of the regional projects have analyzed these problems from a composition perspective (Boes, 1997) and a connection perspective (Wal, 1997). From these rational perspectives ICT is seen as a promising way to increase communication and thus to improve coordination between the distributed professionals. We start to look for the underlying reasons why horizontal communication is so minimal and why recommendations of the earlier studies have not been implemented successfully. With the help of theories from organization science we illustrate that the matrix structure of the police organization is a complex but inevitable way of organizing the police force. The police force wants to manage multiple competing dimensions. They want to be customer focused (close to the civilian), they want to be professional on multiple disciplines (youth, firearms, burglary and so forth), and they want to act as a regionally united police force against area-exceeding criminality. Answering these competing forces at once with a matrix structure means organizing conflict and complexity. To prevent that the police force cannot act instantly due to these continuous management conflicts, general management has chosen to give the local neighborhood teams (which represent the interest of the local civilian) relatively more power. A major problem is that the complex challenge of the matrix organization is not fully adopted by all organizational members. Therefore a lot of the neighborhood teams tend to act in their own interests, and forget to utilize the strengths of the matrix organization. They hardly consult the knowledge of the regional projects during daily problem solving, and they do not always think of consequences of their local policies for the overall regional safety and livability. In contrast, some professionals keep arguing that they just need to spend more time on their topic. So they also aim at serving their own interests, at the expense of the other demands. As such we discover a new set of important factors around coordination, those involving commitment/cohesiveness. 185

COORDINATING DISTRIBUTED WORK The fourth dimension in our coordination framework, content, becomes prominent in observing and participating in the daily work process. The group of regional project coordinators that have to think of ways to fight youth annoyance suffer from difficulties to frame the problem in an integrated picture, rather than in pigeon holes (each of them making sense of the problem from their own frame of reference). Secondly, we observe how policemen have to combine the languages of legal law, governmental policy, and civilian's concerns in their daily work. They have the ungrateful job of explaining the civilian why a problem is outside the scope of a law or outside the priorities of current local government. Understanding before design The value of our first two inductive case studies is that they result in extension of our view on coordination. Commitment and content are more consciously addressed in analysis and design, in addition to composition and contact. In transforming from a hierarchical organization (aiming at catching criminals and primarily reacting on incidents) to a matrix organization (that protects safety and livability) the police force faces some complex challenges. The coordination framework and related paradoxical guidelines give clues how these challenges can be handled. Vagueness of duties Boundaries between the police organization and other organizations that take care of society are becoming vague. A policeman teaches kids in school about norms, values and dangers around guns and drugs. A doctor visits an arrested person in the cell to determine whether he is mentally disturbed or needs to be prosecuted. Civilian-teams watch over their properties at night and catch possible burglars. The question of composition (who does what) becomes more and more complex and is answered differently from situation to situation. Experiments are needed to stimulate other organizations to participate in safety and livability provision and to explore promising alternatives in a time of lacking capacity. However, at some point choices need to be made to turn experimental settings into clear and efficient work processes. Boundaries between what is police work and what is not need to be explicated. Otherwise continuous change and uncertainty harm efficiency. The issues of the day versus developing long term relationships Heading for everything (general duty of neighborhood team) means making temporal choices in priorities. A neighborhood team cannot have maximum attention for the disciplines of 20 regional projects at once. They only invest heavily in those topics that are overlapping with their current problems and priorities (Terpstra, 2002). However, there is a long way between not paying maximum attention or not paying any attention. There should be a reasonable, minimum amount of attention for “topics in rest” to be able to intensify relations with this network when needed. Similar commitment problems exist with external partners. There should be enough long term cohesion to start up committed short-term activities. Currently choices are too black-and-white. Relations between neighborhood teams and internal and external parties seem either intense or dead.

186

CHAPTER 8: REFLECTION Rephrasing the problem: balancing contacts instead of maximizing contacts The problem of intensifying contacts (which was first the central problem) is re-phrased as a question of first looking at composition and commitment, and then determining the desired amount of interaction given these needs. As composition and commitment fluctuate highly depending on temporal and geographical preferences, the amount of interaction and the modes of interaction will fluctuate similarly. Consequently a broad set of media needs to be available. Here a first challenge is to introduce new media and teach organizational members how to use them. Too often it is a choice of either regular time consuming face-to-face meetings (low efficiency), or no communication at all (low effectiveness). Options in between (keeping in touch regularly with phone or e-mail), with a possibly more attractive cost/benefit ratio (more effective than no communication, more efficient than face-to-face meetings), need to be included as common alternatives. A second related challenge is that the police force needs to aim at more fluent interfaces with external organizations. With the most important partners standard media need to be developed collaboratively, without making the relationship so tight that adaptation to local needs is impossible. Sharing intranets and free (protected) e-mail traffic between for instance police and justice could improve their mutual communication (as long as there is some guidance which part of the information is relevant for one another). Police performance: assessing quality of work in addition to output Current discussions on police performance are heavily stressing the amount of work to be performed. Less attention seems to be paid to the quality of work. Society asks for more policemen on the street, but the question what policemen should do and whether they do it adequately is seldom addressed. Our perception is that much can be earned by more investing in proper diagnosis of the source of problems. That is a content issue. It is not surprising that the police force wants to pay more attention to prevention of criminality. They want to lower the need for the huge amount of crime reports, rather than just solve more crime reports. Their best performance is making themselves superfluous. Balancing repression and prevention is a continuous challenge and clear measures need to be developed that show the results of both intervention strategies (Bruijn, 2001; 2002). Furthermore it is important to look at chain-performance, the combined performance of the police force and external parties like justice, judiciary and (local) government. Different languages (content) and goals (commitment) result in communication problems (contact) and a lack of concerted action (composition). The trap in solving these problems is to keep focusing on the symptoms (composition and contact) and not simultaneously addressing the sources of the coordination problems (commitment and content). Our coordination framework should prevent falling in that trap. Decentralization of control requires skilled employees In general we observe that organizational design activities are moving from the drawing board of general management to the hands of organizational members involved in daily work. For management remains the duty to provide prerequisites in the areas of competence, cohesion, connection and context. Policemen experience more and more freedom in managing composition, commitment, contact and content in their daily work. 187

COORDINATING DISTRIBUTED WORK Our study has shown that this increased freedom can have beneficial but also harmful consequences for organizational performance. Therefore it is very important to train organizational members in managing these four dimensions of composition, commitment, contact and content. The simulation-games, developed from a design interest in our research, can also have value as training instruments for the two work processes we have studied. The Amsterdam police force is planning to use an adapted version of the first game to train all policemen how to use general ICT like e-mail, intranet and bulletin boards for knowledge sharing in daily work processes. In addition it is important to address the issues of composition, commitment, contact en content more explicitly in evaluation of daily work. Then the learning experience in the simulation-game is also embedded in daily work practice. Keep discussing and explaining why we organize the way we do On the organizational level “sense making” around the concept of the matrix organization needs to be widened from a discussion on organizational policy between managers, to an organization wide discussion between all policemen on the question “how do we coordinate in the matrix?”. Often good ideas are not or partly implemented because they are lost somewhere between the policy plan and daily practice. It is very important to keep these levels attuned, to keep discussing on the work practice level why you work the way you do. One of the regional coordinators once told that he was sometimes surprised that the main issue in presentations for districts and neighborhood teams is not to motivate them, but rather to explain them what the connection is between all the different organizational change projects currently running. When people lose overview, they also lose motivation to contribute to all the organizational change initiatives. This observation recalls the danger of delineation as a possible outcome of the paradox of organizing (Kastelein, 1990). The coordination framework might be a way to quickly picture the way things are currently done. This might be of help to keep track of the differences in “how we coordinate in the matrix” in time and between different regions. It might also help to reinvolve organizational members who somewhere lost the thread of the discussion around the transformation. We think that our framework and design approach can play a role in supporting the discussion on the continuing improvement of the matrix organization. The matrix organization is still a valuable concept for hybrid organizations In the four years we have been involved in the Amsterdam Police Force much has been done, and our general impression is that still much needs to be done. Looking back, the matrix is maturing. Although in organization literature some leading authors have abandoned this concept as too complex and conflicting, we still see it as a promising concept for hybrid organizations like the police force. They can never completely turn into network organizations. They need to keep some elements from their old bureaucratic hierarchy. Also, it is not possible to separate reactive and proactive work and assign it to two different organizations. Reactive and proactive work are too interdependent. We conclude that the organizational concept of combining organic and mechanic features might be closer to the matrix organization than to other idealistic concepts.

188

CHAPTER 8: REFLECTION

8.3 Value of the design approach for service organizations The aim of our research is to develop a design approach to improve coordination of distributed work in large service organizations. Because our research strategy aims at theory building rather than theory testing and because we only have a limited amount of research resources, we are not able to apply the design approach in other organizations than the Amsterdam Police Force. However, to answer the question whether our design approach is valuable for large service organizations in general, we do not need to conduct a large number of case studies. Our goal is not to find statistical significance, but representative power (Vreede, 1995; Yin, 1989). This means that the extent to which we allege that our findings have significance beyond the context of the case situation, is determined by the fit of our case organization (the Amsterdam Police Force) with the application area (large service organizations). To illustrate this fit six cases from literature are described. We present the coordination problems observed and how they have been tackled. Subsequently we elaborate on the value of our design approach in these cases. Next we summarize the analogies between the Amsterdam Police Force and these cases. From that analysis we conclude that our approach is valuable for the application area of large service organizations. Computer support for clinical work Symon, Long and Ellis (1996) and Bardram (1997a) both describe an analysis of the coordination between ward rooms and the radiology department in a hospital. As Bardram notes the challenge in coordinating the work of these distributed professionals is to minimize mutual articulation (divide, allocate, coordinate schedule etc.) between them (contact) without jeopardizing the motive of treating the patient (commitment). The main coordination artifacts are (paper based) order forms that indicate what the patient is to be examined for and why he needs to be examined. However (Symon et al., 1996) describe that forms are often incomplete (content), but still accepted to prevent delaying the process for the sake of the patient (commitment). Also specialists tend to visit the radiology department personally, either to discuss the treatment prior to filling in the form (contact, content) or to speed up the processing (contact, commitment). In addition Bardram (1997a) elaborates on the crucial role of the secretary in managing temporal constraints. In addition to sending the order forms (content) she bargains by phone (contact) about requesting appropriate time-slots for them (commitment). Often this means making deals with other wards or postponing one of her own patients. The forms were not adequate for reserving time-slots as there was no direct feedback on commitment or rejection of the request. Both studies stress the importance of these informal work practices in addition to the formal work flow, that are needed to meet the situated nature and different values of at first sight similar activities. Symon et al. (1996) even state that most breakdowns (coordination failures) occur in standard (low status) examinations. Specialistic (high status) examinations are more interesting, involve higher status staff, and are more thoroughly processed administratively. They conclude that both the value placed on the work activity (commitment) and the status of the people involved (composition) have 189

COORDINATING DISTRIBUTED WORK impacts on coordination quality. Not surprisingly, both studies warn for CSCW-systems that rely too much on formal work procedures and do not support the informal work practices. Bardram (1997a) explains how different actors (composition) in the whole system require different overviews (content). The information of a prescribed examination is registered in the patient’s medical record, in the nurse’s examination program and in the secretary’s day-plan. He illustrates how functions in a currently used information system are partly used because they do not support these different requirements. For instance, the system does support the sharing of order forms between ward and radiology, but does not support the prioritization that is partly influenced by radiology and partly by the ward (secretary calling). Also, at radiology some cancer patients require planning of treatments for periods of 10 weeks. The system could not provide this long-term overview, and thus a manual planning board was maintained next to the system. The main reason for doubling all information in the information system was accountability purposes of government. Bardram (1996, 2000) describes how an alternative information system (Patient Scheduler) is designed and how two organizational prototyping sessions have been organized as part of this design process. Bardram defines organizational prototyping as “a dual process of both adapting the tool to the organization and adapting the work practice to conditions of the tool”. He discusses two examples where both the Patient Scheduler and the work process have been adapted. The first is the dilemma of service departments (like radiology) to share their resources (enable wards to book examinations themselves), but at the same time stay in control of their use to prevent inefficient planning and escalation of their expenses. The agreed on solution was to include access rights in Patient Scheduler for who could reserve what kind of examination at a particular time on a particular resource. An immediate result of the workshop was that the time allocations agreed on were already implemented in current paper coordination procedures. The second involved planning of surgeries. Here it proved that planning a surgery involved considerable communication between the secretary planning the surgery and the surgeon and nurses involved. The communication module of Patient Scheduler was extended to support this kind of asynchronous communication. To improve the link between planning and daily operation the head nurse of the surgical theaters became involved in the weekly planning meeting, and the secretary moved to an office closer to the surgical theaters where she could see how the plans were realized. Alpha en Zeta company: consultancy and after sales service Orlikowski (1992; 1995) describes two cases of Lotus-Notes implementation and use in commercial service organizations. In both cases Lotus-Notes is expected to enable sharing of work experiences and thereby improving overall quality of service. However, in the first one, the consulting firm Alpha, two factors inhibit the productive use of LotusNotes. The technology is introduced quickly to as many employees as possible (to create critical mass). But poor communication and training causes that the employees develop weak technological frames of the new technology. If people have poor or inappropriate understanding of the unique and different features of a technology they may resist using it, or may not integrate it appropriately into their work practices (gaming). Secondly it 190

CHAPTER 8: REFLECTION proved that the existing norms, policies and rewards were a significant inhibitor to using Notes (cohesion). Incentive schemes, evaluation criteria and culture and individualistic and competitive, whereas Lotus-Notes was intended to be used cooperatively. However, people feared loss of power, control, prestige and promotion opportunities. Because of these cognitive and structural elements groupware did not answer its expectations in this case. Different experiences are recorded at the software company Zeta (Orlikowski, 1995). Here the customers support department starts to use Lotus-Notes as a system for both real time documentation of their work process and knowledge sharing. Major problems with the prior system were that it was not real time, employees only registered their calls after they solved the problem (which could take several days) and kept track of their work in progress on paper. Management had no clue on workloads and performance. With the new system some anticipated and unanticipated changes occurred. First there was indeed more insight in calls in progress and total output (commitment). Also it proved to be much easier to search for solutions to similar problems (contact). Unexpectedly employees started to put much more effort in clearly explaining why they had chosen for specific solutions, in order to ensure reusability of their documented call (content). This increased the reflection on their work process (competence). Furthermore colleagues started to censor each other by commenting constructively on finished work of others and by learning from each other’s solutions. Moreover, they started to contribute proactively to each other’s problem solving in progress by suggesting solutions instead of waiting to be asked to help (composition). Management did not only get insight in call volumes, but also in quality of call handling (content). Whereas at Alpha access to personal information was experienced as dangerous, at Zeta it was accepted as part of the job. Management was trusted by the way they handled this information, they did not abuse it (commitment). Finally the department could be completely reorganized due to the technology: junior staff received all calls in the front office, senior employees were available for assistance and complex questions in the back office (composition). An unexpected backlog of the new way of working was the increasing dependence on technology. As (especially new) employees relied more and more on the existing prior answers in the system, they did not learn to think and solve problems on their own. A two-day breakdown of the system showed that some employees could not handle calls anymore without Lotus-Notes support. This was a reason to reinitiate training on solving problems by thinking and consulting handbooks, rather than copying from the system (competence). E-bank and Brightco: banking and after sales service Swan et al. (2000) describe two cases of multinational companies, Ebank and Brightco, that both aimed at providing a more integrated service across countries to global customers. In the case of Ebank a worldwide communications structure using intranet technology was developed (connection). However, instead of an increase of knowledge sharing across functional and geographical boundaries the project resulted in the development of all kinds of local intranets, which were hardly used for knowledge sharing, even at the local level. Two causes are highlighted. First the translation of the global vision was left to individuals working at the local level. Therefore they all 191

COORDINATING DISTRIBUTED WORK developed different intranet sites and hardly shared experiences on development successes and failures. Secondly only IT specialists were involved. As a result some intranets proved to be technically very sophisticated, but offered little in the way of business relevant innovation. There was a lack of attention for organizational culture and people management (commitment). Also there was no clear vision on what information was relevant to share (content), only the telephone directory and bus timetable were used frequently. Furthermore it is astonishing that the whole case speaks about information sharing, but that composition opportunities are not discussed at all. In comparison Brightco followed a more balanced approach. In developing a global ERP system they involved both IT specialists (connection) and business managers (composition, commitment, content). Furthermore the project teams responsible for development relied both on involving local staff (to be able to adjust the design to local circumstances) as well as central coordination between the different project teams. For the latter purpose the teams arranged to work at the local sites, but met once in three weeks for several days at the company headquarters (contact, commitment). Furthermore, they frequently used e-mails (contact, content) and telephone (contact, commitment) to share development experiences and other strategic concerns in between these meetings. In the end the project teams succeeded in developing a common business information system that was used by most of the employees. It was too early to say whether performance has been improved. The value of our design approach for the literature cases The hospital cases stress the situated nature of coordination choices. Informal work practices are needed in addition to the formal workflow. Underlying dilemmas like quality (incomplete forms) versus speed (delaying the treatment) and efficiency (minimize contacts) versus effectiveness (good treatment) require balancing of situational costs and benefits. In the other cases the situated nature of coordination is not explicitly addressed, but can still be recognized. For instance in Zeta junior staff need to decide for each assignment whether they need to consult a senior consultant or not. Also the degree of detail in registering calls is a clear trade-off that may differ from situation to situation. Similar examples could be given for Alpha. In both E-bank and Brightco the dilemma of prescribing corporate standards versus tailoring the system to local situations is apparent. Although the description of the Brightco case is rather superficial it seems that they succeeded in balancing this dilemma by holding to the corporate standard for some occasions and tailoring to local desires for others. In conclusion the quality of coordination in all cases does not depend on identifying and implementing the “one-best solution”. It rather depends on understanding the various costs and benefits involved in coordination processes and on developing the skill to quickly balance these costs and benefits and make a situated coordination choice. We argue that our design approach focuses on picturing possible coordination choices and on supporting organizational members to explore when what choices are beneficial or not. In addition our coordination framework and accompanying guidelines stress the existence of dilemmas that need to be coped with. Whereas the detailed registration of calls on Zeta company is presented as a benefit (easier reuse, reflection on work process), it can also become a backlog (slows down performance, entries too long to be read quickly). Our design approach more 192

CHAPTER 8: REFLECTION quickly points at the dangers of unilateral solutions and stimulates discussions aiming at identifying dilemma-resistant-solutions. In all cases the elements of our coordination framework are recognizable. This suggests that the framework can be of value in detecting problems and developing solutions. In the action research studies at the Amsterdam Police Force the framework helps to provide structure and overview in discussions on problems and solutions. We argue that clustering the problems on the dimensions of the coordination framework will most likely clarify the analysis. Furthermore we expect that using the framework prevents that opportunities and solutions are identified in a limited direction. For instance, the E-bank case disregards composition, commitment and content issues. The use of our coordination framework would have indicated this in an early stage. The work of Bardram confirms the value of gaming-simulation related techniques in developing useful ICT embedded in a sound socio-technical work setting. He gives clear examples how valuable it can be to provide future users with an integrated picture of new ICT embedded in daily work processes. Still organizational context factors and personal traits are less addressed than in our games because the medical staff involved “discussed” future work alternatives, rather than actually executing them like in our simulation-games. The coordination framework as well as gaming-simulation could have revealed the problems experienced at Alpha and E-bank far before implementation. Similarly, a thoughtful discussion of the elements of the coordination framework and exploration of the new work process in a game could have sped up the learning experiences at Zeta. Unexpected opportunities and problems could have been anticipated before introduction of the new technology. The Brightco case lacks details to say whether the involvement of local users has revealed enough implementation opportunities and difficulties, or whether gaming-simulation could have been of additional value. Possibly gaming-simulation could have shown the feasibility of the increase in performance in an earlier stage, which would have motivated users to adopt the system. Analogies between the literature cases and the Amsterdam Police Force The various service organizations presented here face similar coordination problems as the Amsterdam Police Force. Recognizable problems in the hospital context are balancing the interests of functional departments versus the patient’s concern, balancing the value of formal work procedures versus informal work practices and coping with ICT that both facilitate and complicate work performance. The introduction of ICT without good training in Alpha company, and the misfit between ICT functionality and company culture are also partly transferable to the Amsterdam Police Force. Also, the decision whether a junior employee asks a senior for assistance and the degree of detail in call registrations in Zeta company are reflected in similar problems during problems solving in the neighborhood teams of the police force. Finally, the balance between corporate standards and tailored systems in E-bank and Brightco resemble discussions on intranet structure and content at the Amsterdam Police Force. Schoemaker (1998) defines “border-organizations” as organizations where modern and postmodern forms of organizing and working coexist: “Hierarchical controlled production 193

COORDINATING DISTRIBUTED WORK or service processes are present next to temporary customer focused consultancy functions and knowledge and talent-intensive service projects”. As examples he mentions banks, insurance companies, privatized research organizations, privatized government organizations and privatized health services. Similarly Mintzberg (1983) describes hybrid organizations as organizations that do not design their organization consistent to one of the six configurations, but instead combine features of different configurations. Sometimes these hybrids are dysfunctional; sometimes they are inevitable. As examples Mintzberg mentions police forces and educational institutes, where standardization of skills and standardization of work processes interfere. We argue that the Amsterdam Police Force can be seen as an example of a hybrid or border-organization. Not only forces of standardization of skills (autonomous trained policemen) and standardization of work procedures (juridical laws and government policies) are interfering. Other forces are simultaneously valid: • standardization of output (government policies, control of geographical units) • direct supervision (anti-riot squad fighting during violations of public order) • a strong police culture (not showing weaknesses, solidarity with colleagues) • and mutual adjustment (temporary project teams that focus on preventing crime or inventing new collaboration with external parties like child welfare and the area health authority). Therefore the police force has been a suitable case to develop a design approach that guides coordination of distributed work in hybrid or border-organizations. Conclusion From the theoretical contributions of Schoemaker and Mintzberg and the various practical illustrations we conclude that the Amsterdam Police Force has been a representative case for the application area of large service organizations. As a result we hypothesize that our design approach is valuable for commercial and non-commercial service organizations that are facing problems in coordination of distributed work. We argue that these organizations may benefit from our design approach by better understanding the competing forces in coordination, and by identifying, thoughtfully developing and exploring new ways of coordination, in which ICT can help to integrate these competing forces. Banner and Gagne (1995) point at a quest for mature or intelligent organizations that are able to master the coordination paradox (Kastelein, 1990): they can combine bureaucratic and organic features and as a result exploit the benefits of both intense specialization and true integration. The Amsterdam Police Force may be a prototype-indevelopment for those kinds of large but still organic organizations.

194

CHAPTER 8: REFLECTION

195

9 Epilogue In our research we have explored the issue of coordinating distributed work. We started with the observation that IT has a dual impact on organizational coordination. On the one hand there are numerous opportunities to improve organizational coordination. On the other hand these opportunities prove to be difficult to utilize in practice: computer supported coordination of distributed work is often perceived as complex and ineffective. These observations prompted us to formulate the following research question and objective: • Question: How can groups of distributed professionals in large service organizations improve the coordination of their distributed work processes, given the opportunities and potential drawbacks of information- and communication technologies? • Objective: Formulate a design approach for improving the coordination of distributed work, that incorporates an integrated view on distributed work processes, coordination mechanisms, supporting ICT and skills for distributed coordination within an organizational context. We pursued this research objective by a literature search into the topic of coordination (chapter 3), by exploring the problem of coordinating distributed work in practice (chapter 4), by developing an integrated view on coordination of distributed work (chapter 5), by developing the desired design approach (chapter 6), by applying the design approach in two action research studies (chapter 7) and by evaluating the value of the design approach (chapter 8). In this final chapter we reflect upon the rigor and relevance of our research by summarizing the main research findings (section 9.1) and by discussing the strengths and weaknesses of our research approach (section 9.2). In section 9.3 we conclude with an agenda for future research.

9.1 Research findings In this section we present the main lessons learned. We answer our research questions by summarizing the main conclusions and subsequently discussing the implications of these conclusions for science and society.

197

COORDINATING DISTRIBUTED WORK Research question 1: How to conceptualize coordination? •

•

Current coordination theories have limited views on some of the identified aspects of coordination (process level, organizational level, work arrangements, dependencies, coordination mechanisms, behavioral issues, ICT support, situational-, organizational- and environmental context). The coordination framework provides an integrated view on coordination by discriminating between coordination choices on four dimensions on the process level (composition, commitment, contact, content) and coordination choices on four related dimensions on the organizational level (competence, cohesion, connection, context).

To be able to study and discuss ICT impact on coordination and improvement of coordination a first prerequisite is to understand what is meant with coordination itself. This proves to be more problematic than expected. Malone and Crowston (1990) state that we all have an intuitive sense what coordination means, but that it is hard to define coordination precisely. We fully endorse that observation. Both in literature and in practice we encounter a wide diversity of definitions and interpretations, which complicates the straightforward study, support and improvement of coordination. As the analyzed existing views on coordination lack an integrated understanding of coordination on all identified aspects, we propose a view on coordination that better reflects daily coordination practice in large service organizations. Key conclusion is that managers (on the organizational level) and employees (on the process level) constantly make coordination choices. To be able to cope with the complexity and dynamism of the multiple duties of these organizations, managers do not prescribe how coordination should occur, but rather provide preconditions and different coordination options for their employees. Thereby, employees are enabled to choose to some extent which coordination mechanisms to apply in a situation they encounter. As a result interplay develops between coordination choices of managers and employees. Deliberations about experiences with earlier use of coordination mechanisms, ICT opportunities and threats and adjustment to situational, organizational and environmental context are embedded in all choices. The coordination framework, that embodies our view on coordination, serves multiple purposes. First, it helps to reach shared understanding amongst participants in the organizational redesign process about what is meant with coordination. Discussing about coordination means addressing all 8 areas of coordination choices. Secondly, the framework supports analysis of current coordination practice. What coordination choices get most attention or are overlooked currently? To what extent does management provide or restrict room for coordination choices on each dimension? To what extent do employees make use of the room provided? Thirdly, the framework shows multiple ways of improving coordination. Focus can be shifted from one coordination dimension to another. Managers can provide more or less room on different dimensions. Employees can increase variance in their choices on a particular dimension, or opt for more consistency in their choices. Although not addressed in this research, we expect that the coordination framework can also be valuable as framework for individual reflection and

198

CHAPTER 9: EPILOGUE learning in daily practice. What choices prove to be beneficial under particular circumstances? What choices tend to be overlooked by this person? How much room is desirable for each coordination dimension in this work process? Managers and employees can critically assess their choices over time and use their learning experience to initiate a new redesign process. Both in science and society our framework broadens the view on coordination. We argue that in this case a broader view is a richer view. In analysis the coordination framework enables better understanding of coordination, in improvement it sketches more directions for improvement. As such our framework can be used to mirror and reflect upon currently adopted coordination perspectives in science and society. Research question 2: What is the impact of ICT support on coordination? •

•

ICT has a dual impact on coordination. A multitude of factors (technological, social, psychological and work related) can hamper utilization of opportunities. Consequently, the added results of ICT support on coordination as a whole are unclear. The coordination framework helps to untangle the chaotic interaction of these multiple factors. It proves that ICT has a positive impact on the dimensions of composition and contact and a negative impact on the dimensions of commitment and content.

Our research findings confirm our initial observations in literature and practice that ICT has a dual impact on coordination in distributed work settings. In addition, our coordination framework clarifies where impact is positive and where it is negative. This understanding is crucial to be able to both seize the opportunities and circumvent the threats of ICT support in coordinating distributed work. As shown in the tables in chapter 5 and in the action research studies in chapter 7 listing and addressing both opportunities and threats results in robust design of coordination in distributed work processes. Both researchers and practitioners tend to either advocate or discourage ICT to support coordination. For instance, rational views on coordination (that focus on composition and contact issues) are enthusiastic about ICT use, whereas social views (that emphasize commitment and content issues) are rather negative. We argue that such a black-andwhite discussion does not help organizations to overcome their adaptation problems. Both high hopes and prejudiced skepticism with respect to ICT support hamper improvement of organizational coordination. It is important that managers and employees in organizations start with adopting a neutral attitude towards ICT support. They need to be convinced that the success or failure of ICT supported coordination largely depends on how they apply and use ICT. Do they seize the opportunities of contributing to more processes, involving more colleagues and intensifying contacts in important relations? Do they circumvent the threats of uncommitted participation, vague goals, information overload and lack of shared understanding? Identifying and discussing potential opportunities and threats is a first step

199

COORDINATING DISTRIBUTED WORK in seizing and circumventing them in design and implementation. The coordination framework structures these overviews and discussions and prevents that certain opportunities or threats are overlooked. Researchers can encourage such balanced views by shifting their attention from questions like ‘Is one medium superior to another in a particular situation?’ to questions like ‘How can different media reinforce one another in a particular work context?’ or ‘How can coordination and communication be supported adequately with the media available’. We argue that answers to the latter two questions are only marginally a matter of media choice. Equally important are the organization of the work process, the skills and motivation of the actors involved and the relevance of the information exchanged. Our design approach emphasizes a neutral but critical attitude towards ICT support. It promotes addressing the diversity of impacts of ICT support on coordination, and it stresses that ICT support is only one of the many aspects of coordinating distributed work. The latter complies with an old lesson learned regarding ICT support and organizational change: “First organize, then automate”. Research question 3: How to guide improvement of coordination? • •

Improving coordination requires addressing all four dimensions of the coordination framework on the organizational and process level. Paradoxical guidelines are more valuable than unilateral guidelines due to the multitude of interfering factors, the competing values existent in performance yardsticks and the consequential situated nature of coordination choices.

The concept of coordination choices, which are a key element in our understanding of coordination, provides a clear point of departure: good coordination requires that managers and employees take good coordination choices. In each organization and work process coordination choices are present on all four dimensions. Good coordination requires that coordination choices on all these dimensions are continuously addressed. Good or bad coordination choices are reflected in the productivity of a work process. Improving coordination should lead to an increased quality of service, less efforts (time, money), or both. However, most of the time interventions have dual impacts. An increase in quality goes together with an increase in efforts. The net effect on productivity (impact per unit of additional effort) is hard to determine because especially the quality difference is hard to measure and hard to compare with the increase in effort. Furthermore effectiveness and efficiency are not predetermined and objectively defined constructs. Instead, they are socially constructed and may vary from situation to situation. Sometimes it is wise to involve more people to create innovative viewpoints; sometimes it is better to aim for a small group and a quick decision. Sometimes it is beneficial to quickly consult a standard procedure on intranet; sometimes it requires more time to tailor a best practice to the current problem. Similarly, socializing may be perceived as necessary for commitment and cohesion or as a waste of precious throughput time. In conclusion we are confronted with dynamic and conflicting performance indicators.

200

CHAPTER 9: EPILOGUE Aiming at good coordination is problematic when both the goals and the impact of different interventions depend on situational circumstances like the problem to be tackled, the actors involved, the media currently available, their historical relations and earlier coordination choices. We argue that coordination theories like the information processing and contingency perspective cannot cope with this complexity. It is impossible to select a fixed coordination mechanism for a dependency beforehand. In our research and design approach we deal with this complexity by formulating paradoxical guidelines. Paradoxical guidelines force actors involved in a distributed work process to pay attention to two or more opposing issues simultaneously (for instance quality and throughput time). It stresses the self-responsibility of the distributed work group: whether an issue becomes a benefit or a drawback is more a consequence of more or less heedfully balancing the paradox than that it is a given feature of the distributed work situation. The group members need to reconsider the most convenient way of coordination from situation to situation. Handy (1994) argues that the main requirement for handling paradoxes is recognizing and understanding them. Our design approach supports this in various ways. First the coordination framework and the related general guidelines make practitioners aware of the opposing requirements in a distributed work setting. Secondly, managers and employees are stimulated to formulate their own paradoxical guidelines that are particularly valid for their distributed work setting. Thirdly, they experience how to handle the identified dilemmas and apply their paradoxical guidelines in a game and in their daily work setting. To improve coordination in distributed work settings managers and employees need to recognize and understand paradoxes to be able to handle them in their daily work. For researchers it remains a major challenge to further explore the thin line between successful and unsuccessful handling of paradoxes and to make the dual impacts of paradoxical interventions visible and measurable. Furthermore, there may be more paradoxical guidelines to be identified than the 12 we found. An important aim for further research is the search for more guidelines and dillemma-resistant-arrangements. Research question 4: What is the added value of gaming? •

Gaming-simulation is a promising technique for organizational design because the interplay of organizational, technological, social and behavioral issues becomes tangible and consequently the combined impact of unlike costs and benefits are observable in an early stage of the design process.

Most existing simulation-games have a motivation (unfreeze/freeze) or training (freeze) interest. Other known applications are the use as support instruments for learning in assessment, research and policy-making (Caluwe et al., 1996). In this research we show that games can also be used during the “move” phase of an organizational change process, characterized as a process of “unfreeze-move-freeze” by Lewin (1951). We apply gaming-simulation before the organizational design is fixed and needs to be

201

COORDINATING DISTRIBUTED WORK communicated towards organizational members. We use a simulation-game as a tool to facilitate interactive design rather than as a support tool for the final implementation stage of a linear design (Laere et al., 2000a; 2000b). The value of a game as a design instrument is illustrated by the numerous design recommendations that follow from the debriefings of the simulation-games. The comments in these evaluative discussions are much more detailed than in the earlier redesign discussions and the introductory briefings. Game participants gain insight in what coordination mechanisms are feasible under what circumstances, whether benefits outweigh the costs and when they do and when they do not. We argue that the use of simulation-games in Business Engineering is promising (Laere et al., 2000a; 2000b). They may prevent many of the implementation failures that are caused by a negligence of social, behavioral or contextual issues. The main implication for society is that practitioners are enabled to address these issues in coherence in an early stage of the design and change process. Furthermore developing a realistic simulationgame requires heavy participation of stakeholders. Their participation does not only improve the design of the game, but is also an important learning experience for the people that need to guide the implementation and change process later on (Caluwe et al., 1996). As such not only playing, but also developing the simulation-game is a valuable experience in the collaborative design and change process. A pitfall that may arise for radical, interactive and innovative redesign processes or interventions like these games is the danger of being perceived as a pleasure trip or becoming a ritual (Edelenbos and Monnikhof, 2001). Therefore the facilitator of the redesign and change process should stress a strong relationship between the games played and the implementation and change process afterwards. Although simulation-games can be costly in development, they are often repeatable and offer a high quality learning environment. Overall the benefits seem to outweigh the costs. Nevertheless an ongoing interest for researchers is to show more clearly the deliverables and to shorten the development time by developing flexible frame games that can easily be attuned to the situation at hand. Research objective: design approach that guides and supports improvement •

The design approach offers an integrated view on coordination and gives guidance and support for improving the coordination of distributed work in a large service organization that faces dual demands

We conclude from the answers on our four research questions that we have reached our research objective. Development and implementation of new ways of coordination in a distributed setting is a complex design process. ICT raise opportunities for improving coordination among distributed actors and groups in an organization, but on the other hand numerous problems can occur when trying to utilize these opportunities. In the end the effect on

202

CHAPTER 9: EPILOGUE actual coordination may be unclear or negative. Besides features of the supporting technology, factors on work, organizational, social and psychological level can hamper the utilization of ICT opportunities. Business engineers and actors involved in distributed work processes need to be supported with methods and tools that are capable of handling this complexity. We conclude that existing design approaches are not adequately supporting the improvement of coordination of distributed work. They are either overemphasizing the opportunities of new ICT and neglecting the threats (e.g. Business Process Redesign), or focusing on identifying and listing problems and lacking prescriptive guidelines (e.g. Adaptive Structuration Theory), or limited to change management issues without addressing the question what needs to be changed in what direction (e.g. Technology Use Mediation). In this research we have developed a design approach for improving coordination of distributed work in large service organizations that answers these needs. The main distinctive elements of this design approach are: • the coordination framework that provides an integrated view on coordination; • the paradoxical guidelines that guide improvement in a situation of dynamic and conflicting performance indicators; • and the use of gaming-simulation that shows the interplay between organizational, technological, social and behavioral issues of the design. As the design approach stresses the interplay between coordination on the organizational and process level it aims at supporting a redesign process in which managers, employees and facilitators collaboratively explore new constellations of coordination. The approach can be aimed at one work process in its organizational context, or at multiple linked work processes in an organization. From our analysis in section 8.3 we conclude that it is likely that the design approach is relevant for large service organizations (e.g. hospitals or banks) that face the challenge to comply to conflicting performance indicators like skill development, customer needs and service integration all at once. They need to become mature or intelligent organizations that can combine bureaucratic and organic features (Banner and Gagne, 1995; Quinn, 1992; Qureshi, 2002; Weick, 1998) and are able to master this coordination paradox (Kastelein, 1990). Their employees need to be close to colleague experts to share knowledge an experience, close to the customers to know their demands and close to colleagues with complementary knowledge to provide a set of integrated services. Consequently, employees play multiple roles in multiple processes. Our design approach guides how managers and employees of modern service organizations can utilize ICT to answer these dual demands.

203

COORDINATING DISTRIBUTED WORK

9.2 Research approach As argued in chapter 2 we chose to apply an inductive-hypothetic research strategy and to make use of literature search, case studies and action research studies as research instruments. The most important strength of our research approach is, in our view, the continuous interaction between theory and practice in the descriptive and prescriptive parts of the research. In the initial literature studies we develop some awareness for the main issues and questions surrounding coordination of distributed work. The exploratory case studies at the Amsterdam Police Force are very helpful to expand, deepen, reformulate and refine these initial views. For example, the case study observations guide us how existing views on coordination can be joined into an integrated coordination framework. The questions that arise from our case study observations (like how to guide improvement when indicators are dynamic and conflicting?) give directions for further literature study. Later on, the action research studies support experiencing and reflecting upon the application of our design approach. As such our initial design approach is repeatedly improved into a robust and coherent way of thinking, working, modeling and controlling. These shifts from practice to theory and from theory to practice are strongly incorporated in the inductive-hypothetic research strategy. In addition, the clearly recognizable research phases of the research strategy prevent the researcher from getting lost in either theory or practice or somewhere in between. We spent approximately 4 years in the Amsterdam Police organization. This intensive investigation proves to be very valuable. During the action research studies we could quickly and easily understand and interpret facts, factors and stories that were contributed by policemen participating in the design process. This is beneficial for both the quality and the throughput time of the action research. Furthermore our long term involvement led to a high level of trust between the researchers and employees of the organization under study. This also eased and improved mutual communication and understanding. Due to limited time and resources it has not been in our reach to apply the design approach in another organization. However, in section 8.3 we have elaborated on the representative power of our case situation for the area of large service organizations. From the similarities in coordination problems and redesign difficulties with cases in for instance clinical work and banking enterprises we conclude that it is likely that our design approach is valuable for large service organizations in general. Finally we would like to dwell upon our choice to keep our object of study rather broad. Although it is hard to keep an eye on the multiple factors that have an impact on coordination choices, we argue that this broad scope is necessary to be able to understand the practice of coordination in distributed work processes. When we had limited our focus to one factor like for instance trust, awareness or media choice we would not have been able to explain the different coordination choices policemen make in daily work. In practice multiple fatcors simultaneously influence coordination choices. Therefore theories should also explain how these multiple factors interact. We recommend that more studies in our field should adopt a broad scope and include a large set of factors in their

204

CHAPTER 9: EPILOGUE analysis. The challenge is to sustain reasonable rigor in the presence of high degrees of relevance (Davison et al., 2000). Possible avenues are triangulation of data collection methods, longitudinal investigations of cases as well as research designs that include multiple cases. When these conditions are not achievable within one study they should be secured in research programs. Of course, studies focusing on the impact of one isolated factor remain valuable, but when investigated in the field they should take into account possible side effects from the many other factors involved.

9.3 Future research In the discussion of our research findings we already touched upon some issues that need further exploration. In this final section we discuss these and some other future research topics. We elaborate on some improvements of the design approach and we suggest some research avenues for the field of computer supported cooperative work. Evolution of the design approach First, it has not been in our reach to carry out the final implementation step of the design approach. In our research we only carry out the steps of (1) analyze current work process, (2) identify opportunities, (3) design coordination and (4) develop and play simulationgame. A first recommendation is to include the fifth implementation step in the next opportunity for evaluating and refining the design approach. Secondly, the design approach should be applied in a variety of service organizations to see whether our expectation that the design approach is valuable for large service organizations in general comes true. That would also enable further exploration whether important differences exist between these types of organizations with respect to coordinating distributed work. Thirdly, the four dimensions of the coordination framework can be analyzed more in depth by more elaborately identifying possible problems, opportunities, costs and benefits. For instance, insights from the disciplines of organization science and group dynamics can be used to further specify the dimension of composition/competence. Similarly, organizational psychology and group dynamics can contribute to the dimension of commitment/cohesion, the discipline of IT to the dimension of contact/connection and the disciplines of discursive psychology and speech-act theory to the dimension of content/context. Although we unconsciously promote mono-disciplinary contributions to our framework by linking disciplines to dimensions, the integration of these various contributions into the framework needs to be a continuous point of interest. In this way new dillemas, like for instance orality verus literacy (Woerkum, 2002; 2003), can also be explored and integrated in the coordination framework and design approach. Similarly, further research can aim at identifying more paradoxical guidelines and dilemma-resistant arrangements. With regard to gaming-simulation an interesting issue for further exploration is the simulation of time-duration. Although we introduced several game-elements to simulate 205

COORDINATING DISTRIBUTED WORK unavailability of actors and asynchronous communication, the time elapses between different contacts are relatively small compared to the normal situation. To study for instance the impact of more time for interpretation between contacts we would recommend to do some experiments with simulation-games that last several days and require only a few short-lasting moments a day to participate. In our research we focus on the way of thinking and the way of modeling of the design approach. Thereby the emphasis has been on answering the question what the design of the new work processes should be like. Less attention has been paid to the decision making and change process of approving and implementing the new work methods. Venrooy (2002) shows how elements from a content oriented business engineering approach and elements from a process oriented process management approach can be merged into a tailor-made design approach for public service delivery. After the process management approach has guided the selection and mobilization of actors the business engineering approach helps to structure and facilitate discussion and vision creation amongst those actors. In turn, more elaborated envisioning of problems and opportunities helps to create a sense of urgency for mobilization of actors and change of the work situations. In line with these conclusions a final avenue for further research is to explore how elements from the process management approach can be more explicitly addressed in our adaptive and incremental way of controlling. Furthermore it is a challenge to develop and explore management and change methodologies that are even less focusing on control and more on experimentation (Woerkum, 2003). The field of computer supported cooperative work Given the numerous definitions of coordination we encountered, we strongly advocate that a scientific discussion starts to summarize this confusing abundance of definitions into some integrated concepts. To enhance the depth of this discussion studies could be undertaken where some researchers who each use a different coordination theory or model analyze the same phenomenon in practice and contrast their findings. Such analyses would give insights to what extent different models lead to different insights. A major challenge for researchers advocating the importance of paradoxes in managing organizational work is to provide accurate performance indicators with regard to process and output. These measures should be able to reflect the positive and negative consequences of not correctly balancing the dilemmas involved. So, in some occasions they should point out that many contacts are valuable, whereas in other occasions they should indicate that the same amount of contacts is inefficient and distracting. Without such measures it is hard to discriminate between successful and unsuccessful balancing of conflicting demands. Although there is wide agreement amongst researchers and professionals that participation of stakeholders is required to ensure quality of business engineering projects, there are still lots of questions how collaborative business engineering needs to be carried out (Bockstael-Blok, Mayer and Valentin, 2003; Hengst-Bruggeling and Vreede, 2003; Maghnouij et al., 2001). Should stakeholders be involved as early in the process as possible, or should they react on some initial models developed by 206

CHAPTER 9: EPILOGUE researchers/professionals? How can their ideas and contributions be inserted quickly and adequately? How can large groups of stakeholders participate in model building? How do we prevent strategic behavior of stakeholders? Answers to these kinds of questions need to be explored to improve the usefulness and effectiveness of design approaches for business engineering like the one we developed in this research. We observe that much research in computer supported cooperative work focuses on comparing the performance of two media with one another, or on the identification of beneficial features of ICT for specific uses and situations. In this research we underline that beneficial features of ICT are potential (Leeuwis and Lamerichs, 1999). It depends on the use of ICT in social behavior and daily work practices whether these benefits are utilized and whether a particular medium outperforms another (Qureshi and Zigurs, 2001). Therefore we argue that the core research questions should shift from a comparing face to face with computer mediated coordination, to comparing good work design with bad work design. The choice of the medium is only one of the many aspects of good work design in both co-located and distributed work groups. Often, a less suited medium can still lead to adequate performance, if other factors (division of labor, commitment, mutual understanding) are compensating for the inferior media choice. Also, it is often not a matter of choosing between two media, but using them both to reinforce one another. We conclude that revealing the difference between good and bad work designs and work practices is more valuable than only revealing the differences between good and bad computer support. We hope that these suggestions for future research will motivate researchers in this area to continue the exploration that we have just started. After having been engaged in this research for five years we still see the ‘heedful’ integration of ICT in daily work practices as a major research challenge for the coming years.

207

References ABEL, M. (1990). Experiences in exploratory distributed organisation. In: Galegher, J., Kraut, R.E., Egido, C. (eds), Intellectual teamwork, social and technological foundations of cooperative work. Erlbraum publishers, Hillsdale, NJ, USA. ACKERMAN, M. S. (1994). Augmenting the organizational memory: a field study of answer garden. Proceedings of the ACM conference on computer supported cooperative Work (CSCW’94), pp. 243-252. ACKERMAN, M. S. (1998). Augmenting the organizational memory: a field study of answer garden. ACM Transactions on Information Systems, vol. 16, no. 3, pp. 203-224. ACKOF, R.L. (1978). The art of problem solving. John Wiley & Sons, New York. ACKOF R.L., AND EMERY, F.E. (1972). On purposeful systems. Aldine, Atherton, Chicago and New York. AKEN, J.E., VAN (1994). Strategievorming en organisatiestructurering : organisatiekunde vanuit ontwerpperspectief. Kluwer bedrijfsinformatie, Deventer, the Netherlands. (in Dutch) AKEN, J.E. (2003). Management research based on the paradigm of the design sciences: the quest for field-tested and grounded technological rules. Forthcoming in the Journal of management studies. ALLEN, T. J. (1977). Managing the Flow of Technology: Technology Transfer and the Dissemination of Technological Information within the R&D Organization. MIT Press, Cambridge, MA, USA. ALTER, S. (1996). Information systems, a Management perspective. Benjamin/Cummings Publishing Company, 2nd edition, Menlo Park, USA. AMELSVOORT, P.J. VAN (1992). Het vergroten van de bestuurbaarheid van productieorganisaties. Dissertatie, ST-groep, Oss, the Netherlands. (in Dutch) AMERICAN HERITAGE DICTIONARY (1981). Houghton Mifflin, Boston. ANDRIESSEN, J.H.E. (2002). Working with Groupware. Understanding and Evaluating Collaboration Technology. Springer Verlag, London. ARTMAN, H. AND WÆRN, Y. (1995). Theoretical perspectives on and pragmatic concerns for Situated Action in relation to Information Technology. In: Y. Wærn (Ed.) Prerequisites and Consequences of the Use of Information Technology. HSFR, Brytpunkt, Stockholm, pp. 73-87. ASHBY, W. R. (1969). Self regulation and requisite variety. In: Emery, F.E., Systems Thinking. Penguin Books, New York. BANDOW, D. (1997). Geographically distributed work groups and IT: A case study of working relationships and IS professionals. Proceedings of the 1997 Conference on Computer Personnel Research. SIGCPR 1997, pp. 87-92. BANNER, D.K. AND GAGNE, T.E (1995). Designing effective organizations; traditional and transformational views. Sage, Thousand Oaks, USA. BARDRAM, J. E. (1996). Organisational Prototyping: Adopting CSCW Applications in Organisations. Scandinavian Journal of Information Systems, vol. 8, no. 1, pp. 69-88. BARDRAM, J. E. (1997a). Plans as Situated Action: An Activity Theory Approach to Workflow Systems. Proceedings of the 5th European Conference on Computer Supported Cooperative Work. Kluwer Academic Publishers, Lancaster, UK. BARDRAM, J. E. (1997b). “I Love the System - I just don’t use it!”. Proceedings of GROUP'97 Conference. Phoenix, Arizona, USA. BARDRAM, J. E. (2000): Temporal Coordination: On Time and Coordination of Collaborative Activities at a Surgical Department. Journal of Computer Supported Cooperative Work: the journal of collaborative computing, vol. 9, no. 2, pp. 157-187. BINDT, H., Rapportage enquête netwerk taakaccenthouders vuurwapens, Regiopolitie Amsterdam-Amstelland, Amsterdam. (in Dutch)

209

COORDINATING DISTRIBUTED WORK BENBASAT, I., GOLDSTEIN, D.K., AND MEAD, M. (1987). The case research strategy in studies of information systems. MIS Quarterly, vol. 11, no. 3, pp. 369-386.. A. (1998). Telewerk, omwenteling van tijdruimtelijke arbeidstructuren. BENSCHOP, http://www.pscw.uva.nl/SOCIOSITE/TELEWERK/index.html (in Dutch) BLOM, P. VAN DER AND KORF, J. (2001). Beter uitwisselen van vakkennis binnen het regiokorps AmsterdamAmstelland: Het ontwerpen van werkprocessen om vakkennis beschikbaar te maken voor de korpsleden met behulp van intranet en vraagbaak. Faculteit Techniek, Bestuur en Management, Technische Universiteit Delft, Delft, the Netherlands. (in Dutch) BOCKSTAEL-BLOK, W. (2001). Chains and Networks in Multimodal Passenger Transport; Exploring a design approach. Doctoral dissertation, Delft University of Technology, DUP Science, Delft, the Netherlands. BOCKSTAEL-BLOK, W., MAYER, I. AND VALENTIN, E. (2003). Supporting the design of an inland container terminal through visualization and gaming-simulation. Proceedings Hawaii International Conference on System Sciences (HICSS-36). BOES, L. (1997). Regionale projecten bij de politie Amsterdam-Amstelland: de sterke arm in de matrix. Technische Universiteit Delft, Delft, the Netherlands. (in Dutch) BOSMAN, A. (1995). Werkverdeling en organiseren. Afscheidsrede, Universiteit Groningen. Grongingen, the Netherlands. (in Dutch) BOTS, P.W.G. (1989). An environment to support problem solving. Doctoral dissertation, Delft University of Technology, Delft, the Netherlands. BOTS, P.W.G. AND BRUIJN, J.A. DE (2002). Effective Knowledge Management in Professional Organizations: Going by the rules. Proceedings of the 35th Hawaii International Conference on System Sciences (HICSS-35), January 7-10, 2002, Hawaii (CD-ROM), no. DDOML10.doc, Institute of Electrical and Electronic Engineers (IEEE), 15 pages BOTS, P.W.G. AND SOL, H.G. (1988). Shaping organizational information systems through co-ordination support. In: Lee, R.M., McCosh, A.M., Migliarese (eds.), Organizational Decision Support Systems, Elsevier publishers, Amsterdam, the Netherlands, pp. 139-154. BOUDREAU, M.-C. AND ROBEY, D. (1996). Coping with Contradictions in Business Process Re-engineering. Information Technology & People. Vol. 9, no. 4, pp. 40-57. BRIGGS, R.O. (1994). The Focus Theory of Group Productivity and its Application to the Design, Development, and Testing of Electronic Group Support Technology. Doctoral dissertation, University of Arizona, Tucson, Arizona, USA. BRIGHTMAN, H. J. (1980). Problem solving: a logical and creative approach.Georgia State University, Atlanta, GA, USA. BRUIJN, J.A. DE (2001). Prestatiemeting in de publieke sector. Tussen professie en verantwoording. Lemma BV, Utrecht, the Netherlands. (in Dutch) BRUIJN, J.A. DE (2002). Outputsturing in publieke organisaties. Over het gebruik van een product- en procesbenadering. Management & Organisatie (M&O), vol. 56, nr. 3, pp. 5-21. (in Dutch) BRUIJN, J.A. DE, AND HEUVELHOF, E.F. TEN (1995). Netwerkmanagement. Strategieën, instrumenten en normen. Lemma, Utrecht, the Netherlands. (in Dutch) BRUIJN, J.A. DE, HEUVELHOF, E.F. TEN, AND VELD, R.J. IN ‘T (1998). Procesmanagement, over procesontwerp en besluitvorming. Academic Service, Amsterdam, the Netherlands. (in Dutch) BRUIJN, J.A. DE, AND DE NERÉE TOT BABBERICH, C. (2000). Opposites Attracts, an essay on competing values in knowledge management. Lemma, Utrecht, the Netherlands. CALUWE, L. DE, GEURTS, J., BUIS, D., AND STOPPELENBURG, A. (1996). Gaming: organisatieverandering met spelsimulaties. Delwel, ’s Gravenhage, Twijnstra Gudde, Amersfoort, the Netherlands. (in Dutch) CALUWE, L. DE (1997). Veranderen moet je leren; over de opzet van een grootschalige cultuurinterventie met behulp van een spelsimulatie. Delwel, ’s Gravenhage, the Netherlands. (in Dutch) CAMPION, M. A., MEDSKER, G. J., AND HIGGS, A. C. (1993). Relations between work group characteristics and effectiveness: Implications for designing effective work groups. Personnel Psychology, vol. 46, pp. 823–850. CHECKLAND, P.B (1981) Systems Thinking, Systems Practice. John Wiley and Sons, Chichester, UK. CHECKLAND, P.B. (1985). From optimizing to learning. A development of systems thinking for the 1990s. Journal of the Operational Research Society, vol. 36, pp. 757-767. CHIDAMBARAM, L., AND BOSTROM, R. (1996). Group Development (I): A Review and Synthesis of Development Models. Group Decision and Negotiation, vol. 6, no. 2, pp. 159-187.

210

REFERENCES CHIDAMBARAM, L., AND BOSTROM, R. (1997). Group Development (II): Implications for GSS Research and Practice, Group Decision and Negotiation, vol. 6, no. 3, pp. 231-254. CHING, C., HOLSAPPLE, C.W. AND WHINSTON, A.B. (1992). Reputation, Learning and Coordination in Distributed Decision-Making Contexts. Organization Science, vol. 3, no. 2, pp. 275-297. CHURCHMAN, C. W. (1971). The Design of Inquiring Systems: Basic Concepts of Systems and Organization. Basic Books. New York. CIBORRA, C. (1996), What does groupware mean for the organisations hosting it? In: Ciborra, C. (ed.), Groupware and teamwork, John Wiley, New York, pp. 1-19 COHEN, M.D., MARCH, J.G. AND OLSEN, J.P. (1972). A Garbage Can Model of Organizational Change. Administrative Science Quarterly, vol. 17, no. 1, pp.1-25. COLE, P. (1995). The impact of group context on patterns of groupware use: a study of computer conferencing as a medium of work group communication and coordination. Working paper #182, http://ccs.mit.edu/papers/CCSWP182.html. University of Michigan, Center for Coordination Science at the Massachusetts Institute of Technology, Michigan, USA COLEMAN, D. (1997). Groupware: collaborative strategies for corporate LANs and intranets. Prentice Hall, Upper Saddle River, USA. COLLINS, J.C., AND PORRAS, J.I. (1997). Built to last: Successful habits of visionary companies, Harper Busines, New York, USA. COOPERRIDER, D.L., SORENSEN P.F. JR., WHITNEY D. AND YAEGER, T.F. (2000). Appreciative Inquiry. Rethinking organization toward a positive theory of change. Stipes Publishing, Champaign, Illinois, USA. COPPES, R., GROOT, F. DE AND SHEERAZI, A., (1997), Politie en criminaliteit van Marokkaanse jongens, Gouda Quint, Deventer and Willem Pompe Instituut, Utrecht (in Dutch). CORBEIL, P. (1999). Learning From the Children: Practical and Theoretical Reflections on Playing and Learning. Simulation and Gaming. Vol. 30, no. 2, pp. 163-180. COZIJNSEN, A.J. (1989). Het innovatievermogen van politie-organisaties; onderzoek naar de mogelijkheden van een non-profit organisatie om complexe vernieuwingen succesvol door te voeren. Kluwer, Deventer, the Netherlands. (in Dutch) CROWSTON, K. (1994). A taxonomy of organizational dependencies and coordination mechanisms. Working paper #174, http://ccs.mit.edu/papers/CCSWP174.html. University of Michigan, Center for Coordination Science at the Massachusetts Institute of Technology, Michigan, USA. CULNAN, M. (1985). The dimensions of perceived accessibility to information: Implications for the delivery of information systems and services. Journal of the American Society for Information Science. Vol. 36, no. 5, pp. 302-308. CULNAN, M.J. AND MARKUS, M.L. (1987). Information technologies. In: Jablin, F.M., Putnam, L.L., Roberts, K.H. and Porter, L.W. (Eds.), Handbook of Organizational Communication: An Interdisciplinary Perspective. Sage Publications, Newbury Park, CA, USA, pp. 420-443. CUMMINGS, T. (1981). Designing effective work-groups. In: Nystrom, P. and Starbuck, W. (Eds.). Handbook of Organizational Design. Oxford University Press, New York. CURTIS, B. (1989). Modeling coordination from field experiments. In: Proceedings of the Conference on Organizational Computing, Coordination and Collaboration: Theories and Technologies for Computer-Supported Work. CYERT, R.M. AND MARCH, J.G. (1963). A Behavioral Theory of the Firm. Englewood Cliffs, NJ, USA. DAFT, R. L. AND LENGEL, R. H. (1984). Information richness: a new approach to managerial behavior and organizational design. In B.Staw and L.L. Cummings, Research in Organizational Behavior. Vol. 6, pp. 191-233. DAFT, R.L., AND LENGEL, R.H. (1986). Organisational information requirements, media richness and structural design. Management Science. Vol. 32, no. 5, pp. 554-571. DAFT, R. AND WEICK, K.E. (1984). Toward a model of organizations as interpretation systems. Academy of management review, vol. 9, no. 2, pp. 284-295. Reprinted in: Weick, K.E. (2001). Making sense of the organization. Blackwell, Oxford, chapter 10, pp. 241-258. DAVENPORT, T.H. AND SHORT, J.E, (1990). The new industrial engineering: Information technology and business process redesign. Sloan Management Review, Summer, vol. 31, no. 4, pp 11-27. DAVENPORT, T.H. (1993). Process Innovation: Reengineering Work Through Information Technology. Harvard Business School Press, Boston, Massachusetts.

211

COORDINATING DISTRIBUTED WORK DAVISON, R.M., QURESHI, S., DE VREEDE, G.J., VOGEL, D.R. AND JONES, A.N. (2000). Group Support Systems Through the Lens of Action Research: Experiences in Organisations. Journal of Global IT Management, vol.3, no. 4, pp. 6-23. DENNIS, A. R., VALACICH, J. S., SPEIER, C., AND MORRIS, M.G. (1998). Beyond Media Richness: An Empirical Test of Media Synchronicity Theory. Proceedings of the 31st Hawaii International Conference on System Sciences (pp.48-57). Los Alamitos, CA: IEEE Computer Society. DESANCTIS G., AND GALLUPE, R.B. (1987). A foundation for the study of GDSS. Management Science. Vol. 33, no. 5, pp. 589-609. DESANCTIS, G., AND POOLE, M.S. (1994). Capturing the complexity in advanced technology use: Adaptive structuration theory, Organization Science, vol. 5, no. 2, pp. 121-147. DONALDSON, L. (1996). The Normal Science of Structural Contingency Theory. In: Clegg, S.R., Hardy, C., Nord, W.R., Handbook of Organization Studies. Sage Publications, London, pp. 57-76. DUMAINE, B. (1994). The trouble with teams. Fortune, vol. 131, no. 5, pp. 86-92. DRUCKER, P.F. (1988). The coming of the new organisation. Harvard Business Review, Vol. 66, no. 1, pp. 4553. DRUSKAT, V.U. AND WOLFF, S.B. (2001). Building the emotional intelligence of groups. Harvard Business Review, vol 79, no. 3, pp. 80-90. DUKE, R.D. (1974), Gaming: The future's language. New York: Sage Publications. DUNCKER, K. (1945). On problem solving. Psychological Monographs, vol. 58, no. 270. DUR, R. C. J. (1992). Business Reengineering in Information Intensive Organizations. Doctoral dissertation, Delft University of Technology, Delft, the Netherlands. EASON, K. (1988). Information technology and organisational change. London, Taylor & Francis. EDELENBOS, J. AND MONNIKHOF, R. (2001), Lokale interactieve beleidsvorming. Een vergelijkend onderzoek naar de consequenties van interactieve beleidsvorming voor het functioneren van de lokale democratie, Lemma, Utrecht, the Netherlands. (in Dutch) EIJCK D.T.T. VAN (1996). Designing Organizational Coordination. Doctoral dissertation, Delft University of Technology, Delft, the Netherlands. ELGOOD, C. (1988). Handbook of management games. Aldershot, Gower, 4th edition. ELLIS, C.A., GIBBS, S.J., AND REIN, G.L. (1991). Groupware: Some issues and experiences. Communications of the ACM. Vol. 34, no. 1, pp. 38-58. EMERY, F.E. AND EMERY, M. (1974). Participative design. Canberra: Centre for continuing education, Australian national university. FINHOLT, T., SPROUL, L., AND KIESLER, S. (1990). Communication and ad hoc task groups. In: Galegher, J., Kraut, R.E., Egido, C. (eds), Intellectual teamwork, social and technological foundations of cooperative work. Erlbraum publishers, Hillsdale, NJ, USA, pp. 291-326. FJERMESTAD, J., AND HILTZ, S. (2000). Case and field studies of group support systems: an empirical assessment, Proceedings of the 33rd Hawaii International Conference on System Sciences (HICSS33), January 4-7, 2000, Hawaii (CD-ROM), Institute of Electrical and Electronic Engineers (IEEE), 10 pages FLORES, F., GRAVES, M., HARTFIELD, B., AND WINOGRAD, T. (1988). Computer systems and the design of organizational interaction. ACM Transactions on Office Information Systems, vol. 6, no. 2, pp. 153172. FULK, J. (1993). Social construction of communication technology. Academy of Management Journal, Volume 36, Number 5, pp. 921-950. FULK, J. AND DESANCTIS, G. (1995) Electronic Communication and Changing Organizational Forms. Organization Science, Volume 6, Number 4, pp. 337-349. FULK J. AND RYU D. (1990). Perceiving electronic mail systems: A partial test of the social information processing. Paper presented to a meeting of the International Communication Association, Dublin. FULK, J., STEINFIELD, C. W., SCHMITZ, J. AND POWER, J. G. (1987). A social information processing model of media use in organizations. Communication Research. Vol. 14, no. 5, pp. 529-552. GALBRAITH, J.R. (1973). Designing complex organisations. Addison Wesley, Reading, USA. GALBRAITH, J.R. (1974). Organization Design: An Information Processing View. Interfaces. Vol. 4, no. 3, pp. 28-36. GALBRAITH, J. R. (1993). Organizing for the Future: The New Logic for Managing Complex Organizations. San Francisco, CA: Jossey-Bass Publishers. GALBRAITH, J.R. (1994). Competing with flexible lateral organizations. Addison Wesley, Reading, USA.

212

REFERENCES GALLIERS, R.D. (1991). Choosing appropriate information systems research approaches: a revised taxonomy. In: Nissen, H. E., Klein, H.K. and Hirscheim, R. (eds.), Information systems research: Contemporary approaches and emergent traditions. Elsevier Science Publishers, North-Holland, pp. 327-345. GEORGE, A.L. (1980), Presidential decisionmaking in foreign policy: The effective use of information and advice, Boulder, Westview Press. GEURTS, J.L. (1995). Opportunities and challenges for Gaming/Simulation. In: Crookall, D. and Arai, K. (eds.), Simulation and Gaming across disciplines and cultures. Sage Publications Inc., Thousands Oaks, California. GOODMAN, P.S., RAVLIN, E., AND SCHMINKE, M. (1987). Understanding groups in organizations. In: L.L. Cummings and B.M. Staw (eds.), Research in Organizational Behavior, 1987: An Annual Series of Analytical Essays and Critical Reviews. Stamford, CT: JAI Press, pp. 121-173. GREENBLAT, C.S. (1988). Designing games and simulations: An illustrated handbook. Sage Publications Inc., Newbury Park, CA. GREIF, I (1988). Computer-Supported Cooperative Work: A Book of Readings. Morgan Kaufman publishers, San Mateo, California. GRUDIN, J. (1988). Why CSCW Applications Fail. Proceedings of the ACM conference on computer supported cooperative Work (CSCW’88), pp. 85-93. GRUDIN, J. (1994). Groupware and Social Dynamics. Communications of the ACM. Vol. 37, no. 1, pp. 93-105. GULICK, L. AND URWICK L. (1937). Papers on the Science of Administration. Institute of public administration, New York. HACKMAN, J.R. (1987). The design of workteams. In: Lorsch, J.W. (ed.), Handbook of organization behavior. Prentice Hall, Englewood Cliffs, NJ. HAMMER, M. (1990), Reengineering work: Don’t automate, obliterate. Harvard Business Review, vol. 68, no. 4, pp. 104-112 HAMMER, M. AND CHAMPY J. (1993). Reengineering the Corporation. Harper Business, New York. HANDY, C. (1994). The age of paradox. Harvard Business School Press, Boston. HANDY, C. (1995). Trust and the virtual organization. Harvard Business Review, Vol. 73, No. 3, pp.40-50. HANSEN, M.T., NOHRIA, N., AND TIERNEY, T. (1999). What’s your strategy for managing knowledge? Harvard Business Review, vol. 77, no. 2, pp. 106-116. HAYWOOD, M. (1998). Managing virtual teams: practical techniques for high-technology project managers. Artech House, London. HAERENI T. AND SIGHOLT E. (1995) A perspective on IT and knowledge transfer. In Proceedings of the Third European Conference on Information Systems (ECIS 1995), pp. 701-720, Athens, Greece. HENGST-BRUGGELING, M. DEN (1999). Interorganizational Coordination in Container Transport; a chain management design. Doctoral dissertation, Delft University of Technology, Delft, the Netherlands. HENGST-BRUGGELING, M. DEN AND VREEDE, G.J. DE (2003).Vision Support Systems: A Framework for Research. Proceedings of the Thirty-Sixth Annual Hawaii International Conference on System Sciences (HICSS36). HERZBERG, F., MAUSNER, B., AND SNYDERMAN, B.B. (1959). The motivation to work. John Wiley, New York. HERZBERG, F. (1968). One more time: How do you motivate employees? Harvard Business Review, vol. 46, no.1, pp. 53-62. HETTINGA, M. (2002). Understanding evolutionary use of groupware. Doctoral dissertation Delft University of Technology. Telematica institute, Enschede, the Netherlands. HILLS, M. (1997). Intranet as groupware. Wiley, New-York. HILTZ S.R., DUFNER, D., FJERMESTAD, J., KIM, Y., OCKER, R., RANA A., AND TUROFF, M. (1996). Distributed groups support systems: theory development and experimentation. In: Olsen B.M., Smith, J.B., Malone T., Coordination theory and collaboration technology, Lawrence Erlbaum associates, Hillsdale NJ, USA. See also: http://eies.njit.edu/~hiltz/CRProject/DGSS/dgsschap.html HINSSEN, J.H. (1998). What difference does it make? The use of groupware in small groups. Doctoral dissertation Erasmus University Rotterdam. Telematica institute, Enschede, the Netherlands. HOLMSTRÖM J (1999). It’s as if Somebody Slammed the Door, On the Limits of Organisational Change around Information Technology. In: Proceedings of the Seventh European Conference on Information Systems (ECIS 1999), pp. 320-329, Copenhagen Business School, Copenhagen. HOLT, C.C. (1991). Conceptual environment for organization support systems: information technology in organizations. In: Sol, H.G., Vecseny (eds.), Environments for supporting decision processes, pp. 215-238. Elsevier Science publishers (Noth-Holland), Amsterdam, the Netherlands

213

COORDINATING DISTRIBUTED WORK HOOFF, B.J. VAN DEN (1997). Incorporating Electronic Mail. Adoption, use and effects of electronic mail in organizations. Doctoral dissertation University of Amsterdam. Otto Cramwinckel, Amsterdam, the Netherlands. HOWCROFT, D. AND MITEV, N.N. (1999) Doctors on the Net: are GPS ready for the Information Superhighway? In: Proceedings of the Seventh European Conference on Information Systems (ECIS 1999), pp. 866879, Copenhagen Business School, Copenhagen. HUBER, G.P. AND MCDANIEL, R. (1986) The decision making paradigm of organizational design. Management Science, vol. 32, no. 5, pp. 572-589. JANSSEN, M.F.W.H.A. (2001). Designing Electronic Intermediaries. An agent-based approach for designing interorganizational coordination mechanisms. Doctoral Dissertation, Delft University of Technology, Delft, The Netherlands. JONES, M. (1995). The contradictions of business process re-engineering. In: Burke, G. and Peppard, J. (Eds), Examining Business Process Re-engineering , The Cranfield Management Series, London, pp. 43-59. KASTELEIN, J. (1990). Modulair organiseren, tussen autonomie en centrale beheersing. Wolter Noordhof, Groningen. (in Dutch) KATZENBACH, J.R. AND SMITH, D.K. (1993). The wisdom of teams: Creating the high performance organization. Harvard Business School Press, Boston. KEEN, P.G.W. (1980). Adaptive design for DSS. Database, vol. 12, no. 1-2, pp. 15-25. KICKERT, W.J.M. (1979). Organisation of decision making: A systems-theoretical approach. Doctoral dissertation, Eindhoven University of Technology, Eindhoven, the Netherlands. KIESER A., AND KUBICEK, H. (1977). Organisation. Walter de Gruyter, Berlin/New York. (in German) KINNEY, S.T., AND PANKO, R.R. (1996). Project teams: profiles and member perceptions. Proceedings of 25thHawaiian International conference on System Sciences (HICSS 25), IEEE, Computer Society Press, Los Alamitos, CA, USA. KOCK, N. (1999). Process improvement and organizational learning: the role of collaboration technologies. Idea, Hershey. KOCK, N. (2001). Asynchronous and distributed process improvement: the role of collaborative technologies, Information Systems Journal, Vol. 11, No. 1, pp. 87-110. KOLB, D.A. (1984). Experimental learning. Prentice Hall, Englewood Cliffs, NJ. KUIPERS, H. AND AMELSVOORT, P. VAN (1992). Slagvaardig organiseren: inleiding in de sociotechniek als integrale ontwerpleer. Kluwer Bedrijfswetenschappen, Deventer, the Netherlands. (in Dutch) KYDD, C.T. AND FERRY, D.L. (1992). Electronic Mail, Social Presence and Information Richness. Proceedings of the 25th Hawaiian International Conference on System Sciences (HICSS-25), pp. 380- 391. KYDD C.T., AND FERRY, D.L. (1995). Electronic Mail and New Methods for Measuring Media Richness, on: http://hsb.baylor.edu/ramsower/acis/papers/kydd.htm LAERE, J. VAN (1998a). Het ondersteunen van regionale samenwerking bij de Regiopolitie AmsterdamAmstelland, met als doel bevordering van professionalisering en innovatie, Faculteit Techniek, Bestuur en Management, Technische Universiteit Delft, Delft, the Netherlands. (in Dutch) LAERE, J. VAN (1998b). Supporting distributed co-ordination: improving the ‘metastructuring’ of computer supported co-ordination mechanisms by adopting a holistic situated co-ordination perspective. Research proposal. Faculty of Technology, Policy and Management, Delft University of technology, Delft, the Netherlands. LAERE, J. VAN (1999a). Coördinatie tussen verspreid werkende groepen bij de Regiopolitie AmsterdamAmstelland: een verkennend onderzoek binnen het Regionaal Project Vuurwapens naar de mogelijkheden van ondersteuning van coördinatie op afstand met informatie- en commununicatietechnologie. Faculteit Techniek, Bestuur en Management, Technische Universiteit Delft, Delft, the Netherlands. (in Dutch) LAERE, J. VAN (1999b). Supporting distributed coordination. Contribution to the Doctoral Consortium at the 7th European conference on information systems (ECIS 1999), Copenhagen, June 20th -22nd , 1999. LAERE, J. VAN (2000). Het spel poliprof: het verbeteren van de coördinatie van vakkennis en expertise tussen wijkteams van de regiopolitie Amsterdam-Amstelland. DITSE, Delft, the Netherlands. (in Dutch) LAERE, J. VAN, VREEDE, G.J., AND SOL, H.G. (2000a). Supporting intra-organisational distributed coordination at the Amsterdam Police Force. In: Proceedings of the 33rd Hawaii international conference on system science (cd-rom), 4-7 januari 2000, IEEE, Los Alamitos, CA, pp. 1-10. LAERE, J. VAN, VREEDE, G.J., AND SOL, H.G. (2000b). Designing intra-organizational distributed coordination at the Amsterdam police force: The application of gaming as a tool to facilitate technology use mediation. Journal of information technology cases and applications. Vol. 2, no. 4, pp. 3-26.

214

REFERENCES LAERE, J. VAN (2002a). Coördinatie tussen verspreid werkende professionals bij de Regiopolitie AmsterdamAmstelland. Hoe wijkteams en regioprojecten samen problemen (kunnen) oplossen binnen de context van een matrixorganisatie. DITSE, Delft, the Netherlands. (in Dutch) LAERE, J. VAN (2002b). Coördinatie op afstand tussen wijkteams en regioprojecten binnen de matrixorganisatie. Presentation for the Management Board of the Amsterdam Police Force, April 18th, 2002. DITSE, Delft, the Netherlands. (in Dutch) LANE, C., AND BACHMAN, R. (1998). Trust within and between organizations; conceptual issues and empirical applications. Oxford: Oxford University Press LAVE, J., AND WENGER, E. (1991). Situated learning: Legitimate peripheral participation. Cambridge University Press, Cambridge, UK. LEAVITT, H.J. (1965). Applying orgnisational change in industry: structural, technological and humanistic approaches. In: March, J.G., Handbook of organizations. RandMcNally, Chicago, Illnois. LEEUW, A.C.J. DE (1990). Organisaties, Management, Analyse, Ontwerp en Verandering. Van Gorcum, Assen/Maastricht. (in Dutch) LEEUW, A.C.J. DE (1994). Besturen van veranderingsprocessen: fundamenteel en praktijkgericht management van veranderingsprocessen. Van Gorcum, Assen. (in Dutch) LEEUWIS, C. AND LAMERICHS, J. (1999). Internet en communicatieve interventie. In: Woerkum, C.M.J. van and Meegeren, R.C.F. van (eds.), Basisboek communicatie en verandering, pp. 195-223. Boom, Amsterdam/Meppel. (in Dutch) LEWIN, K. (1951). Field theory in social science, Harper & Row, New York. LEWIN, A. Y., AND STEPHENS, C. U. (1993). Designing postindustrial organizations: Combining theory and practice. In: Huber, G. and Glick, W. (eds.), Organizational change and redesign: combining ideas and insights for improving managerial performance, pp. 393-407. Oxford University Press, New York. LIAO, Z., AND LANDRY JR., R. (2000) An empirical study on organisational acceptance of new information systems in a commercial bank environment. Proceedings of the 33rd Hawaii International Conference on System Sciences (HICSS-33). January 4-7, 2000, Hawaii (CD-ROM). Institute of Electrical and Electronic Engineers, (IEEE), 10 pages. LOHMAN, F.A.B. (1999). The effectiveness of management information: a design approach to contribute to organizational control. Doctoral dissertation, Delft University of Technology. Dordrecht, the Netherlands. MAGHNOUJI, R., VREEDE, G.J. DE , VERBRAECK, A., AND SOL, H.G. (2001). Collaborative Simulation Modeling: Experiences and Lessons Learned. Proceedings of the 34th Hawaiian International Conference on Systems Sciences (cd-rom proceedings, 10 pages). IEEE Computer Society Press, Los Alamitos, CA. MAIDIQUE, M.A. AND HAYES, R.H. (1984). The Art of High-Technology Management. Sloan Management Review, vol. 25, no. 2, pp.17-31. MALONE, T. W., AND CROWSTON, K. (1990). What is coordination theory and how can it help design cooperative systems? In: Proceedings of the conference of computer supported cooperative work (CSCW '90), ACM, New York. MALONE, T.W., AND CROWSTON, K. (1994). The interdisciplinary study of coordination. ACM Computing Surveys. Vol. 26, no. 1, pp. 87-119. MALONE, T.W. AND ROCKART, J.F. (1991). Computer, networks and the corporation. Scientific American, vol. 265. No. 3, pp.92-99 MARCH, J.G., AND OLSEN, J.P. (1976). Ambiguity and Choice in Organizations. Universitatsforlaget, Bergen, Norway. MARCH, J.G., AND SIMON, H.A. (1958). Organizations. John Wiley & Sons, Inc., New York. MARKUS, M.L., AND CONNOLLY, T. (1990). Why CSCW Applications Fail: Problems in the Adoption of Interdependent Work Tools, Proceedings of the ACM conference on computer supported cooperative Work (CSCW’90), pp. 371-380. MCGRATH, J.E. (1984). Groups: Interaction and Performance. Prentice-Hall, Englewood Cliffs, N.J., USA. MCGRATH, J.E. (1990). Time matters in groups. In: In: Galegher, J., Kraut, R.E., Egido, C. (eds), Intellectual teamwork, social and technological foundations of cooperative work. Erlbraum publishers, Hillsdale, NJ, USA. MCGRATH, J.E. (1991). Time, interaction, and performance (TIP): A theory of groups. Small group research, vol. 22 no. 2, pp. 147-174

215

COORDINATING DISTRIBUTED WORK MCGRATH, J.E. AND HOLLINGSHEAD, A.B. (1993). Putting the group back in group support systems. In: Jessup, L. and Valacich, J. (Eds.). Group Support Systems: New Perspectives. Macmillan, New York, pp. 7896. MCGRATH, J.E., AND HOLLINGSHEAD, A.B (1994). Groups interacting with technology: ideas, evidence, issues and an agenda. Sage Publications, London. MEEL, J.W. VAN (1994). Dynamics in business engineering. Doctoral dissertation, Delft University of Technology, Delft, the Netherlands. MEINSMA, R.R. (1997). Decision support in business environments. Doctoral dissertation, Delft University of Technology, Delft, the Netherlands. MENNEKE, B.E., VALACICH, J.S. AND WHEELER, B.C. (2000). The effects of media and task on user performance: A test of the task media fit hypotheses. Group decision and negotiation, vol. 9, pp.507529. MILES, M. (1979). Qualitative Data as an Attractive Nuisance: The Problem of Analysis. Administrative Science Quarterly, vol. 24, no. 4, pp. 590-601. MINTZBERG, H. (1979a). The structuring of organisations. Prentice-Hall, Engelwood Cliffs, NJ, USA. MINTZBERG, H. (1979b). An Emerging Strategy of 'Direct' Research. Administrative Science Quarterly, vol. 24, pp. 580-589 MINTZBERG, H. (1983). Structure in fives, designing effective organisations. Prentice Hall International Editions, NJ, USA. MITROFF I.I., BETZ F., PONDLY L.R., AND SAGASTY F. (1974). On Managing Science in the Systems Age: Two Schemas for the Study of Science as a Whole Systems Phenomenon. TIMS Interfaces, vol. 4, no. 3, pp. 46-58. MONGE, P. R. AND EISENBERG, E. M. (1987). Emergent communication networks. In: Jablin, F.M., Putnam, L.L., Roberts, K.H. and Porter, L.W. (Eds.), Handbook of Organizational Communication: An Interdisciplinary Perspective. Sage Publications, Newbury Park, CA, USA, pp. 304-342 MOORHEAD,G. AND GRIFFIN, R.W. (1989). Organizational Behavior. Houghton Mifflin Co., Boston. MORRIS, D. (1993). De Naakte Aap, A.W. Bruna Uitgevers B.V., Utrecht. (in Dutch) MORTON, M.S.S. (1992). Reshaping the corporation for the future, Managing the organizational challenge successfully. Revue des systèmes de décision, vol 1, no. 4, pp. 341-352. MUMFORD, E. (1983). Designing Human Systems for New Technology: The ETHICS Method. Manchester: Manchester Business School. MUMFORD, E. (1995). Book review of Hammer, M., Champy, J., Reengineering the corporation: a manifesto for business revolution. European journal of information systems, no. 4, pp. 116-120. MUNKVOLD, B.E. (2000). Tracing the roots: the influence of socio-technical principles on modern organizational change practices. In: Coakes, Willis and Lloyd-Jones (eds.), The new sociotech: grafitti on the long wall. Springer-Verlag, London, pp. 13-25. NAHAVANDI, A., AND ARANDA, E. (1994). Restructuring teams for the re-engineered organisation. Academy of Management Executive. Vol. 8, no. 4, pp. 58-65. NGWENYAMA, O.K., AND LEE, A.S. (1997). Communication Richness in Electronic mail: Critical Social Theory and the Contextuality of meaning. MIS Quarterly, pp. 145-165. NIJHUIS, R.J.H. (1996). Working apart together. Working paper, Delft University of Technology, Delft, the Netherlands. NIJHUIS, R.J.H., AND DE VREEDE, G.J. (1996). A research agenda for distributed groups support en coordination. Proceedings of the 7th Euro GDSS Workshop. Grenoble, March 28-29, 1996. NUNAMAKER, J. F., DENNIS, A. R. VALACICH, J. S. VOGEL, D. R. AND GEORGE, J. F. (1991). Electronic meeting systems to support group work. Communications of the ACM. Vol. 34, no. 7, pp. 40-61. O’BRIEN, M. (1995). Who’s got the ball? Jossey Bass, San Francisco, CA, USA. OKAMURA, K., FUJIMOTO, M., ORLIKOWSKI, W.J., AND YATES, J. (1994) Helping CSCW applications succeed: The role of mediators in the context of use. In: Proceedings of the Conference on Computer Supported Cooperative Work (CSCW’94), ACM/SIGCHI & SIGOIS, NY, p. 55-65. B. (2002). Blauwe bureaucraten. Intermediair, vol. 38, no. 17, OOSTERHOUT, www.intermediair.nl/Weekblad/weekblad/inhoud0217.html (in Dutch) ORLIKOWSKI, W.J. (1992). Learning form notes, organisational issues in groupware implementation. Proceedings of the ACM conference on computer supported cooperative Work (CSCW’92), pp. 362269.

216

REFERENCES ORLIKOWSKI W.J., YATES, J., OKAMURA, K., AND FUJIMOTO, M. (1995). Shaping electronic communication: the metastructuring of technology in the context of use. Organization Science. Vol. 6, no. 4, pp. 423444. ORLIKOWSKI, W.J. AND BARLEY, S.R. (2001). Technology and institutions: What can research on information technology and research on organizations learn from each other? MIS Quarterly, vol. 25, no. 2, pp. 145-165. ORR, J. (1990). Sharing Knowledge, Celebrating Identity: War Stories and Community Memory in a Service Culture. In: Middleton, P. and Edwards, D. (Eds.) Collective Remembering: Memory in Society. Sage Publications, London, pp. 169-189. OXFORD DICTIONARY (1982). The concise Oxford Dictionary. Oxford university press, Oxford. PARKER FOLLET, M. (1937). The process of control. In: Gulick, L. and Urwick L., Papers on the Science of Administration. Institute of public administration, New York. PERROW, C. (1984). Normal accidents. Basic Books, New York. PINSONNEAULT, A., AND KRAEMER, K.L. (1989). The Impact of Technological Support on Groups: An Assessment of the Empirical Research. Decision Support Systems. Special Issue on Group Decision Support Systems, vol. 5, pp. 197-216. PUGH, D.S. AND HICKSON, D.J. (1989). Writers on organizations. Penguin, London, 4th ed.. PUTNAM, L.L., PHILLIPS, N. AND CHAPMAN, P. (1996). Methapors of communication and organization. In: Clegg, S.R., Hardy, C., Nord, W.R., Handbook of Organization Studies. Sage Publications, London, pp. 375-408. QUINN, J. B. (1992). Intelligent Enterprise. The Free Press, New York. QUINN, R.E. (1988). Beyond rational management: Mastering the paradoxes and competing demands of high performance. W.H. Freeman, San Francisco. QUINN, R.E., FAERMAN, S.R., THOMPSON, M.P. AND MCGRATH, M.R. (1996). Becoming a master manager: a competency framework. John Wiley & Sons, New York. QUINN, R.E. AND ROHRBAUGH, J. (1983). A Spatial Model of Effectiveness Criteria: Towards a Competing Values Approach to Organizational Analysis. Management Science, vol. 29, no. 1, pp. 363-377. QURESHI, S. (1995). Organizations and Networks: Theoretical Considerations and a Case Study of Networking across Organizations. Doctoral dissertation. London School of Economics, University of London, London. QURESHI, S., AND VOGEL, D. (1999). Organisational adaptiveness in virtual teams. In: Vreede, G.J., Ackermann, F. (ed.), Proceeedings of the 10th EuroGDSS workshop, Copenhagen. QURESHI, S. AND ZIGURS, I. (2001). Paradoxes and Prerogatives in Global Virtual Collaboration. Communications of the ACM, vol 44, no. 12, pp. 85-88. QURESHI, S., VAN DER VAART, A., KAULINGFREEKS, G., DE VREEDE, G.J., BRIGGS, R.O., AND NUNAMAKER, J. (2002). What does it mean for an Organisation to be Intelligent? Measuring Intellectual Bandwidth for Value Creation. In: Proceedings of the 35th Hawaii International Conference on System Sciences (HICSS-35), January 7-10, 2002, Hawaii (CD-ROM), no. CLUSR14.doc, Institute of Electrical and Electronic Engineers (IEEE). RESNICK, L.B. (1991). Shared Cognition: Thinking as Social Practice. In: Resnick, L., Levine, J. and Teasley, S.D. (eds.). Perspectives on Socially Shared Cognition. American Psychological Association, Washington, DC, pp. 1-22. RICE, R., AND SHOOK, D. (1988). Access to, usage of, and outcomes from an electronic messaging system. ACM: Transactions on Office Information Systems, vol. 6, no. 3, pp. 255-276. RICE, R.E. (1994). Relating electronic mail use and network structure to R&D work. Journal of Management and Information Services, vol. 11, no. 1, pp. 9-30. ROBERTSON, M., SORENSEN, C., AND SWAN, J. (2000). Facilitating knowledge creation with Groupware: a case study of a knowledge intensive firm. Proceedings of the 33rd Hawaii International Conference on System Sciences (HICSS-33), January 4-7, 2000, Hawaii (CD-ROM). Institute of Electrical and Electronic Engineers, (IEEE), 10 pages. ROGERS, E. M. AND AGARWALA-ROGERS, R. (1976). Communication in Organizations. The Free Press, NewYork. SAKAMOTO, Y., AND KUWANA, E. (1993). Toward Integrated Support of Synchronous and Asynchronous Communication in Cooperative Work: An Empirical Study of Real Group Communication. Conference on Organizational Computing Systems (COOCS’93), ACM press, pp. 90-97. SALANCIK, G.R. AND PFEFFER, J. (1978). A social information processing approach to job attitudes and task design. Administrative Science quarterly, vol. 23, pp.224-253.

217

COORDINATING DISTRIBUTED WORK SCHOEMAKER, M.J.R. (1998). Tussen slavernij en anarchie; een verkenning naar veranderingen in de contractuele configuratie in laat-moderne organisaties. Doctoral dissertation, University of Nijmegen. Kluwer Bedrijfsinformatie, Deventer. SCHMIDT, K. AND BANNON, L. (1992). Taking CSCW seriously: Supporting Articulation Work. Computer Supported Cooperative Work: the journal of collaborative computing. Vol. 1, no. 1, pp. 7-40. SCHMIDT, K., AND SIMONE, C. (1996). Coordination mechanisms: Towards a conceptual foundation of CSCW system design. Computer Supported Cooperative Work: the journal of collaborative computing. Vol. 5, pp. 155-200. SHANNON, R.E. (1975). Systems Simulation: The art and science. Prentice Hall, Englewood Cliffs, New Jersey SHANNON, C.E., AND WEAVER, W. (1949). The Mathematical Theory of Communication. The University of Illinois Press, Urbana. SHAW, M. (1981). Group Dynamics: The Psychology of Small Group Behavior. McGraw-Hill, New York, 3rd Edition. SHEOMBAR, H. S. (1995). Understanding Logistics Coordination: A Foundation For Using EDI In Operational (Re)Design Of Dyadical Value-Adding Partnerships. Doctoral dissertation, Tilburg University, Tilburg, The Netherlands. SHORT, J., WILLIAMS, E. AND CHRISTIE, B. (1976). The social psychology of telecommunications. Wiley, New York. SHULMAN A.D. (1996). Puttings group information technology in its place: communication and good work group performance. In: Clegg, S.R., Hardy, C., Nord, W.R., Handbook of Organization Studies. Sage Publications, London, pp. 357-374. SIMON, H.A. (1945). Administrative behavior. Macmillan, New-York. SIMON, H.A. (1960). The new science of management decision. Harper and Row, New York. SIMON, H.A. (1973). The structure of ill-structured problems. Artificial Intelligence, vol. 4, pp. 181-201. SITTER, L.U. DE (1982). Op weg naar nieuwe fabrieken en kantoren. Kluwer, Deventer. (in Dutch) SMAGT, A.G.M. VAN (1997). Situationeel handelen in organisaties: het einde van het information processing paradigma. In: Verslag van de achtste NOBO onderzoeksdag: Dynamiek in organisatie en bedrijfsvoering, Eindhoven, 18 november 1997, pp. 157-166. (in Dutch) SMITH, A. (1910). The Wealth of Nations. Dent, London. SOL, H.G. (1982). Simulation in information systems development. Doctoral dissertation, University of Groningen, Groningen, the Netherlands. SOL, H.G. (1990). Information systems development, a problem solving approach. In: Proceedings of the international symposium on system development technologies, Atlanta, USA. SOL, H.G. (1992). Shifting boundaries in systems engineering and policy analysis. Inaugurall address, Delft University of Technology, Delft, the Netherlands. SOL, H.G. AND BOTS, P.W.G. (1993). Information Systems to Support Decision Processes: From decision support to networks of inquiry systems. In: C.W. Holsapple, A.B. Whinston, (eds.), Recent Developments in Decision Support Systems, Springer-Verlag, Berlin, pp. 23-34. STEINFIELD, C., JANG, C., AND PFAFF, B. (1999). Supporting virtual team collaboration: The TeamSCOPE system. Proceedings of the Group99 Conference, ACM press, New York. STRAUSS, S.G., AND MCGRATH, J.E. (1994) Does the medium matter? The interaction of task type and technology on group performance and member actions. Journal of Applied Psychology, vol. 79, no. 1, pp. 87-97. SUCHMAN, L. (1987). Plans and situated action: the problem of human machine communication. Cambridge University Press, Cambridge, UK. SUCHMAN, L. (1994). Do categories have politics?, Computer Supported Cooperative Work: the journal of collaborative computing. Vol. 3, no. 2, pp. 177-190. SUTTON, R.L. AND LUIS, M.R. (1987). How selecting and socializing newcomers influences insiders. Human resources management, vol. 26, pp. 347-361. SWAN, J., NEWELL, S. AND ROBERTSON, M. (2000). Limits of IT-Driven Knowledge Management Initiatives for Interactive Innovation Processes: Towards a Community-based Approach. Proceedings of the 33rd Hawaii International Conference on System Sciences (HICSS-33), January 4-7, 2000, Hawaii (CDROM). Institute of Electrical and Electronic Engineers, (IEEE), 10 pages. SYMON, G., LONG, K. AND ELLIS, J. (1996). The coordination of work activities: Cooperation and conflict in a hospital context. Computer Supported Cooperative Work: the journal of collaborative computing. Vol. 5, pp. 1-31 TAYLOR, F.W. (1947). Scientific management. Harper & Row, New York.

218

REFERENCES TERPSTRA, J. (2002). De lokale aansturing van politie en politiewerk. Een verkennend onderzoek tegen de achtergrond van een veranderende sturingscontext en sturingsstijl. Politie & Wetenschap, Apeldoorn. Uitgeverij Kerckebosch, Zeist. (in Dutch) THOMPSON, J.D. (1967). Organisations in action. McGraw-Hill, New York. TOFFLER, A. (1970). Future shock. Random House, New York. TREVINO, L. K., LENGEL, R. H. AND DAFT, R. L. (1987). Media symbolism, media richness, and media choice in organizations: A symbolic interactionist perspective. Communication Research. Vol. 14, no. 5, pp. 553-574. TRIST, E. AND BAMFORTH, K. (1951). Some Social and Psychological Consequences of the Longwall Method of Coal Getting. Human Relations, vol. 4, no. 1, pp. 3-38 TRIST, E. (1981), The Evolution of Socio-technical Systems as a Conceptual Framework and an Action Research Tool. In: Van de Ven, A. and Joyce, W. (eds.), Perspectives on Organizational Design and Behavior. Wiley, New York TUCKMAN, B.W. (1965). Developmental sequence in small groups. Psychological bulletin, Vol. 63, no. 6, pp. 384-389. TUCKMAN, B. W., AND JENSEN, M. A. C. (1977). Stages in small group development revisited. Group and Organizational Studies, vol. 2, pp. 419-427. TURNER, W.S. (1980). Project auditing methodology. North Holland, Amsterdam. TUROFF, M., AND HILTZ, S.R. (1991). An overview of research activities in computer mediated communications from 1976 to 1991 conducted by the Computerised Conferencing and communications center at NJIT. On: http://eies.njit.edu/~turoff/Administrative/ccc.htm TWIST, M.J.W. VAN, EDELENBOS, J. AND BROEK, M.G. VAN DER (1998). In Dilemma’s Durven Denken, M&O, Tijdschrift voor Management en Organisatie, vol. 52, no. 5, pp.7-23. (in Dutch) UIJLENBROEK, J. J.M. (1997). Designing Electronic Document Infrastructures, Doctoral Dissertation; Delft University of Technology, Delft, the Netherlands. VAN ALSTYNE, M. (1997). The State of Network Organization: A Survey in Three Frameworks. Journal of Organizational Computing, 1997, Vol. 7, no. 3. VAUGHAN, D. (1996). The challenger launch decision. University of Chicago Press, Chicago. VAN DE VEN, A.H., DELBECQ, A.L., AND KOENIG, R., JR. (1976). Determinants of coordination modes within organizations. American Sociological Review, vol. 41, pp. 322-338. VELDE, R. TE (2002). Kennis maken en kennis hebben. In: Baalen, P. van, Weggeman, M, Witteveen, A., Kennis en management, pp. 358-379, Scriptum, Schiedam, the Netherlands. (in Dutch) VENROOY, A. VAN (2002). Nieuwe vormen van interorganisationele publieke dienstverlening: de ontwikkeling en verkenning van een ontwerpaanpak. Doctoral Dissertation; Delft University of Technology, Delft, the Netherlands. VISSER, C. (2001). Succesvol verandermanagement door Appreciative Inquiry: Afscheid van probleemgerichte interventies?, M@N@GEMENT, see http://www.managementsite.net/content/articles/295/295.asp VREEDE, G.J. DE (1995). Facilitating organisational change, the participative application of dynamic modelling. Doctoral dissertation, Delft University of Technology, Delft, the Netherlands. VREEDE, G.J. DE (1997). Collaborative Support for Design: animated electronic meetings. Journal of MIS. Vol. 14, no. 3, pp. 141-164. VREEDE, G.J. DE, DAVISON, R.M. AND BRIGGS, R.O. (2002). How a Silver Bullet May Lose Its Shine. Communications of the ACM. Vol. 46, no. 8, pp. 96-101. VROLIJK, C. (2000) POLITEAM: de ontwikkeling van een management game voor de Regiopolitie AmsterdamAmstelland om te leren coördineren m.b.v. ICT. Faculteit Techniek, Bestuur en Management, Technische Universiteit Delft, Delft, the Netherlands. (in Dutch) WAL, A. VAN DER (1997). Van Kakafonie naar Symfonie, een onderzoek naar de interne communicatie van de regionale projecten, Regiopolitie Amsterdam-Amstelland / Universiteit van Amsterdam, Amsterdam, the Netherlands. (in Dutch) WALSH, J.P. AND UNGSON, G.R. (1991). Organizational memory. Academy of management review, vol.16, pp. 57-91. WEBER, M. (1947). The theory of social and economic organization. Free Press, New York. WEBSTER, J. (1998). Desktop Videoconferencing: Experiences of Complete Users, Wary Users, and Non-Users, MIS Quarterly, vol. 22, pp. 257-286. WEICK, K. E. (1969). The social psychology of organizing. Addison-Wesley, Reading, MA, USA.

219

COORDINATING DISTRIBUTED WORK WEICK, K.E. (1985a). Sources of order in underorganized systems. Themes in recent organizational theory. In: Lincoln, I. (ed.), Organizational theory and inquiry: the paradigm revolution, Sage publications, Beverly Hills, CA. Reprinted in: Weick, K.E. (2001). Making sense of the organization. Blackwell, Oxford, chapter 2, pp. 32-56. WEICK, K.E. (1985b). Cosmos vs. Chaos: Sense and nonsense in electronic contexts. Organizational Dynamics, vol. 14, no. 2, pp. 50-65. Reprinted in: Weick, K.E. (2001). Making sense of the organization. Blackwell, Oxford, chapter 20, pp. 444-457. WEICK, K.E. (1993). Organizational redesign as improvisation. In: Huber, G.P. and Glick, W.H. (eds.), Organizational change and redesign: Ideas and insights for improving performance. Oxford university press, Oxford, UK. Reprinted in: Weick, K.E. (2001). Making sense of the organization. Blackwell, Oxford, chapter 3, pp. 57-92. WEICK, K.E. (1998). The attitude of wisdom: Ambivalence as the optimal compromise. In: Srivaste, S. And Cooperrider, D.L. (eds.), Organizational wisdom and executive courage, Lexington, San Franciso, CA. Reprinted in: Weick, K.E. (2001). Making sense of the organization. Blackwell, Oxford, chapter 16, pp. 361-379. WEICK, K.E. (2001). Making sense of the organization. Blackwell, Oxford. WEICK, K.E. AND BROWNING, L. (1986). Arguments and narration in organizational communication. Jorunal of management, vol. 12, pp. 243-259. WEICK, K.E. AND ROBERTS, K. (1993). Collective mind in organizations: Heedful interrelating on flight decks. Administrative science quarterly, vol. 38, no. 3, pp. 357-381. Reprinted in: Weick, K.E. (2001). Making sense of the organization. Blackwell, Oxford, chapter 11, pp. 259-283. WENGER, E. (1998). Communities of practice, Learning, Meaning and Identity. Cambridge university Press, Cambridge, UK. WICKS, A.C. AND FREEMAN, R.E., (1998), Organization studies and the new pragmatism: positivism, Antipositivism, and the search for ethics. Organization Science, vol. 9, no. 2, pp. 123-140. WIGAND, R.T., PICOT, A., AND REICHWALD, R. (1997). Information, Organisation and Management: Expanding Markets and Corporate Boundaries. Wiley & Sons, New York. WINOGRAD, T. AND FLORES, F. (1986). Understanding Computers and Cognition: A New Foundation for Design, Ablex Publishing Corporation, Norwood, NJ. WOERKUM, C. VAN AND MEEGEREN, P. VAN (1999). Communicatie en verandering een introductie. In: Woerkum, C. van, Meegeren, P. van, Basisboek communicatie en verandering, Uitgeverij Boom, Amsterdam, the Netherlands, pp. 11-28. (in Dutch) WOERKUM, C. VAN (2002). Orality in environmental planning. European Environment, vol. 12, pp. 160 – 172. WOERKUM, C.M.J. VAN (2003). Organisaties en hun biotoop: over communicatie van organisaties. Inaugurall address, Wageningen University, June 5th, 2003. (in Dutch) WOMACK, J, AND JONES D. (1994). From lean production to the lean enterprise. Harvard Business Review, vol. 72, no. 2, pp. 93-103. YIN, R.K. (1989). Case study research: Design and methods (Revised Edition). Sage, London.

220

REFERENCES

221

Summary Coordinating distributed work Exploring situated coordination with gaming-simulation December 4th, 2003 Joeri van Laere This dissertation sheds some more light on the heavily studied but still largely unexplored field of coordinating distributed work. The main deliverables are an integrated coordination framework to support analysis, a set of paradoxical guidelines for improving coordination of distributed work, and the demonstration of simulation-games as valuable design instruments. Research problem Through the ages organizations have faced the challenge of organizing (distributed) work and coordinating the efforts of (distributed) organizational members. The increasing rate of development of information- and communication technologies (ICT) in the past 20 years has raised some new challenges around this old issue. At first sight there seem to be numerous opportunities to solve well-known coordination problems more easily. By adopting the new technology coordination, collaboration and the accomplishment of work itself can become independent of time and place. This independence leads to more flexibility for individual group members in accomplishing their tasks and may thereby improve individual performance, work group performance as well as organizational performance. But in practice it proves to be quite difficult to utilize these ICT opportunities. Most organizations fail to fully utilize the opportunities for improving their coordination. They tend to stick to their usual ways of working and rely on traditional communication media. Mediated collaboration and coordination in distributed work settings is perceived to be more complex and less effective than coordinating or collaborating face-to-face So, ICT brings benefits as well as new problems and the added results may be unclear. Researchers as well as practitioners struggle with integrating ICT, tasks, structures and human behavior in organizations. In addition, we observe some shortcomings in contemporary research and understanding. First, there is a lack of integration of the findings of various research disciplines, like organization science, information technology 223

COORDINATING DISTRIBUTED WORK and group dynamics. Secondly, limited conceptions of coordination and communication prevail, that do not address issues like meaning generation, relational aspects, persuasive appeals and social and political context. Thirdly, there is a domination of experimental studies while group work proves to be socially, organizationally and temporarily embedded. Finally, design approaches aiming at improving coordination of distributed work picture human and behavioral aspects as an implementation problem, rather than as a part of the socio-technical system to be redesigned. Research questions and objective We argue that there is a need for more integrative views and approaches to the problem of coordinating distributed work. Our research question, How can actors in large service organizations improve the coordination of their distributed work processes, given the opportunities and potential drawbacks of information- and communication technologies?, is answered by addressing the following four subquestions: • How can we conceptualize coordination? • What is the impact of ICT support on coordination? • How can we guide improvement of coordination? • What is the added value of gaming-simulation as design instrument? The ultimate goal of this research is to develop a design approach that guides actors in large service organizations how to improve the coordination of their distributed work processes, based on an integrated view on coordination. We pursue this research objective by a literature search into the topics of distributed work (chapter 1) and coordination (chapter 3), by exploring the problem of coordinating distributed work in practice (chapter 4), by developing an integrated view on coordination of distributed work (chapter 5), by developing the desired design approach (chapter 6), by applying the design approach in two action research studies (chapter 7) and by evaluating the value of the design approach (chapter 8). Research design Ontologically, our research follows an interpretivist philosophy and inductive reasoning, subjectively interpreting observations of reality. We apply an inductive-hypothetical research strategy that is suitable to support theory building and to study ill-structured problems. This strategy consists of five steps. We start with a literature exploration that results in a rough initial theoretical framework based on existing theories of coordination. This initial theory guides our observations in two participatory case studies that aim at increasing our understanding of what coordination means and at identifying what the experienced opportunities and challenges of ICT in supporting coordination of distributed work processes are (Initiation). The case study observations and the results of some additional literature search result in some requirements for our view on coordination and for our design approach (Abstraction). Subsequently, we develop an integrated view on coordination that incorporates the factors that promote and hamper improvement of coordination of distributed work. Next we formulate a design approach by describing the way of thinking, modeling, working and controlling with a particular emphasis on guidelines for improving coordination and on gaming-simulation as a design instrument 224

SUMMARY (Theory Formulation). We conduct two action research studies to show and evaluate the value of our design approach as a ‘proof by demonstration’ (Implementation). Finally, we reflect upon our experiences in the action research studies (Evaluation). In our research case studies are the primary research instruments during the initiation and implementation phases. We have chosen to conduct all 4 case studies in one organization because it is important to be long enough in an organization to be able to understand what is going on. This organization is the Amsterdam Police Force. The police force struggles with facilitating collaboration and knowledge sharing between their geographically distributed neighborhood teams. Communication and coordination in the regional projects (a kind of communities of practice that connect policemen from different neighborhood teams on a specific topic) is both promising as well as disappointing. During all case studies both quantitative and qualitative data from various sources are collected to enable a rich representation of the phenomena under investigation and to permit comparison and contrast of the collected data (multiple sources of evidence). Coordination theory As we aim to arrive at an integrated view on coordination we explore how approaches from various research disciplines look at coordination. We limit our exploration to the following approaches: • the information processing paradigm, contingency theories, managing task- and actor dependencies and socio-technical systems theory from the discipline of organization science. • Business Process Redesign approaches from the discipline of IT research. • theories and approaches on group performance and group design in face-to-face settings and computer supported settings from the discipline of group dynamics. • theories of enactment and sensemaking and situated action or situated coordination that have roots in the disciplines of cognitive psychology, human computer interaction, and social psychology. From this analysis we conclude that coordination starts with a division of labor that results in a certain work arrangement. This work arrangement creates dependencies between tasks and actors that need to be managed by coordination mechanisms. Coordination occurs on the organizational level, the group level and the individual level. Furthermore aspects of coordination that require attention are behavioral, ICT support and (situational-, organizational- en environmental-) context. We observe that no theory or approach covers all identified aspects of coordination. Coordination problems can result from uncertainty (dominant in rational views on coordination) and ambiguity (underlined in social views on coordination). In addition, behavioral approaches emphasize the problematic nature of communication and decision making. Due to the complex nature of coordination the studied theories and approaches are rather vague about guiding and measuring improvement. The conception of coordination and coordination problems heavily influences the directions for improvement.

225

COORDINATING DISTRIBUTED WORK The main implications for our case studies are that we need to find out whether all identified elements, aspects and levels of coordination and sources of coordination problems are recognizable in practice. Furthermore practice may give clues how the different conceptions of the analyzed approaches can be integrated. Finally, it is interesting to look how uncertainty and ambiguity problems are solved in daily coordination in the Amsterdam Police Force. Exploratory case studies at the Amsterdam Police Force Two different distributed work processes within the Amsterdam Police Force are analyzed in order to find out what types of coordination mechanisms, coordination problems, ICT opportunities and promoting and hampering factors can be discerned. These are respectively the work group Youth Annoyance and the Regional Project (Fire)Arms. The work group Youth Annoyance is a work group consisting of six regional coordinators who need to give an advice how to deal with Youth Annoyance problems in District 6. We analyze the internal coordination within the work group as well as the individual and group interactions with (people from) District 6. Most of the internal coordination occurs in the 10 group meetings that are organized approximately once a week. The regional project Fire-Arms consists of all policemen in the organization who are involved in tracing arms en firearm-suspects, handling arms and suspects and information analysis and regional knowledge sharing considering Fire-Arms. These policemen work in various departments. One regional coordinator stimulates knowledge exchange and makes sure that the different organizational units that are involved work together smoothly and that their efforts are attuned to outside parties like nation wide departments and the administration of justice. In both cases we observe that the distinction between work arrangements, dependencies and coordination mechanisms is recognizable. However, due to the multitude of dependencies and coordination mechanisms present in the police organization, there is not a clear and fixed coordination mechanism for each dependency to be managed. Instead of a fixed design that is executed, we observe that policemen have considerable freedom to emphasize particular coordination mechanisms more than other ones. These so-called coordination choices are needed to deal with the complexity and dynamism of the organizational environment. By making different coordination choices policemen can adapt the coordination of their work to situational circumstances. In addition, we observe that uncertainty and ambiguity problems coexist in practice and require joint solutions. The impact of ICT is diverse. ICT supported coordination mechanisms coexist with their traditional counterparts (work procedure in handbook and on intranet), or make formerly unrealistic coordination mechanisms feasible (visiting versus calling or mailing an expert at the other side of town). As such, ICT results in an even larger set of coordination mechanisms to choose from and consequently it complicates the coordination choices to be made. ICT is particularly useful to solve uncertainty problems, but can also contribute to solving ambiguity problems. Furthermore ICT can be useful for distributed as well as co-located groups. Unbalanced use of ICT may lead to counterproductive effects. Both 226

SUMMARY for traditional and ICT supported coordination mechanisms holds that a multitude of interfering technological, social, organizational and personal factors can hamper the successful appliance of the coordination mechanism, which worsens rather than improves coordination. The observed coordination choices prove to be the main starting point for improvement of coordination. Matching currently emphasized coordination mechanisms to situational circumstances and to the currently prevalent dependencies is more useful than trying to define one fixed coordination mechanism to be executed. In practice coordination choices are based on a comparison of the potential costs and benefits of different coordination mechanisms. These costs and benefits cannot accurately be determined, but are roughly estimated. They refer to the match with the dependency to be managed, to situational circumstances and to the expected difficulties of applying the coordination mechanism (e.g. hampering technological, social, organizational and personal factors). A first requirements for our view on coordination and our design approach that follows from this analysis is that ‘coordination choices’ should be a central issue in the conception of coordination: better coordination choices result in improved coordination. Furthermore, the conception of coordination should pay attention to uncertainty problems and ambiguity problems. Finally, our design approach should not only enable identification of ICT opportunities, but it should also explain how these opportunities can be seized successfully, by means of showing potential costs and benefits. The multitude of technological, social, organizational and personal factors that influence ICT adoption need to be addressed in these costs and benefits. Coordination framework From the literature exploration and the case study observations we develop an integrated view on coordination that is articulated in the coordination framework (see Figure 4, page 98). The framework illustrates that work execution produces results for a certain assignment and requires coordination choices on the group level on 4 dimensions (made by employees) and 4 related coordination choices on the organizational level on each dimension (made by managers). Consequently, coordination is the combination of: • composing a group of employees with appropriate organizational competencies, • ensuring commitment among this group supported by an overall organizational cohesiveness, • producing contacts between the group members enabled by available organizational connections, • and sharing content with each other interpreted in an organizational context, in order to produce the desired results for a given assignment. We observe that uncertainty issues (composition, contact) are connected with the amount of relations and the amount of accommodations. Ambiguity issues (commitment, content) appear to be related to the quality of relations and the quality of accommodations. From our overview of costs, benefits, opportunities and threats we conclude that ICT impact is generally positive for solving uncertainty issues and generally negative for ambiguity

227

COORDINATING DISTRIBUTED WORK issues. However, whether ICT reveals as an opportunity or as a threat depends heavily on how the employees in the organization apply the technologies in their daily work and coordination, and on the situational circumstances. When employees are aware of the ICT opportunities the chance rises that they will be turned into real advantages. Similarly knowing the threats helps circumventing them. Our design approach guides how the right balance in the coordination choices can be found. The design approach The design approach is characterized by its mode of thought (Weltanschauung), its modeling constructs, its working method and its management approach. We refer to these characteristics as the way of thinking, the way of modeling, the way of working and the way of controlling. The main purpose of the way of thinking in our design approach is to guide how opposing issues in coordination of distributed work, like uncertainty and ambiguity problems and opportunities and threats of ICT support can be balanced. Coordination choices have both costs and benefits and the overall impact is unclear and differs from situation to situation. A choice between two alternatives that both incur severe problems or that both are equally attractive is called a dilemma. A dilemma is by definition unsolvable in theory, but can be dealt with in practice by accommodation. This means understanding the paradox and consolidating the learning experiences in the quest for ‘dilemma-resistantarrangements’. From the competing values school we learn that guidelines, which stress balancing of simultaneous opposing issues, are often formulated as paradoxes. We formulate 12 guidelines that clarify how opposing demands in coordination of distributed work can be managed. These are (see also Table 44, page 118): COMPOSITION / COMPETENCE 1. Consult distributed expertise extensively, but keep groups small, focused and productive 2. Select employees by standard competencies present and by unique competencies needed 3. Deploy experts to specialist and generalist tasks COMMITMENT / COHESION 4. Stimulate both social communication and formulation of clear goals early in process 5. Rotate leadership among group members 6. True support requires expressing a critical attitude CONTACT / CONNECTION 7. Contact regularly to show progress and evaluate, but do not interrupt productive action unnecessarily 8. Let fast information sharing give room to interactive interpretation 9. Customized (unpredictable) media choice requires (predictable) communication standards for all media CONTENT / CONTEXT 10. Be to the point but do not omit relevant details 11. Add missing context and hide distracting context 12. Let experts assist consultation of knowledge systems

For our way of modeling we use two modeling techniques. Task-actor models fulfill the need of providing a quick overview of current work processes and coordination mechanisms as well as sketching possible future alternatives. In addition gamingsimulation fulfills the need of experiencing new work processes, new coordination mechanisms, new ICT to support coordination, new behavioral and social norms and future organizational context as one meaningful whole. By enacting the future situation organizational employees can both evaluate the new coordination alternatives more 228

SUMMARY thoroughly as well as gain experience how to deal with the coordination framework and the paradoxical guidelines. The way of working involves 5 steps, namely: Analyze current work processes, identify promising opportunities for distributed work, design coordination of distributed work, develop and play simulation-game, and implement the new work processes. To facilitate learning the way of controlling needs to be “middle out” (focusing on a small but critical part or the organization that is recognizable) and “incremental” (quick delivery of an initial model that can be revised and enhanced in a number of cycles). Thereby quick feedback is gained, which is not only beneficial for learning and the quality of the design, but also helps to strengthen management and employee support for the change process. Action research studies One action research study focuses on the involvement of distributed expertise in daily problem solving in neighborhood teams. The second action research study involves the coordination of filling and maintaining the regional project intranet site and answering and verifying answers on the message boards of the regional projects. In both studies we carry out the first four steps of our design approach (analyze current work process, identify opportunities, design coordination, develop and play game). In the identification of opportunities, in the design process and during playing and evaluation of the games the paradoxical guidelines are applied. In the first study the developed game is played nine times with policemen from the Amsterdam Police Force. In the second study the game is played two times. Both studies result in numerous learning experiences and practical recommendations for the Amsterdam Police Force. For instance, in the first study policemen prove to rely too much on their own expertise, while consulting other souces (like intranet or other experts) lead to better performance. The intranet contains too much general and longwinded information which is hard to apply on practical problems. The breadth and depth of experts’ knowledge and expertise differ highly, whereas little information is available to locate the most suited expert to help solving the problem at hand. Finally, the bulletin board will only be used intensively when speed and quality of answers can be guaranteed. In the second study it shows that regional projects aim at quality rather than at speed, whereas neighborhood teams (their customers) ask for both quality and speed. Furthermore is concluded that appointing both central and local webmasters balances the needs for a standard navigation structure versus local adaptations and quality control. Finally, the benefits of a small, distributed core group of experts (continuous occupation, larger contribution of expertise) outweigh the costs of more complicated coordination and the danger of lacking commitment. Reflection With regard to the usability and usefulness of the different elements of our design approach (coordination framework, paradoxical guidelines and simulation-games) in the action research studies we conclude the following: • the coordination framework structures the design discussions; • the framework helps to identify opportunities, threats, benefits and costs; • guidelines can be tailored to the coordination issue under study;

229

COORDINATING DISTRIBUTED WORK • • • • • • •

in general there is a positive IT impact for composition/contact, and a negative IT impact for commitment/content; gaming-simulation provides insight whether/when benefits outweigh costs; unilateral guidelines are not adequate for guiding coordination: policemen confirm the need for paradoxical guidelines to meet conflicting demands in their work; comments during simulation-games are more specific than during design due to applying the guidelines on practical cases rather than talking about them in general; multiple goals are achieved in the simulation-games, they serve a design as well as a training interest; whereas gaming experts advise against computer use in games, computer use makes logistics and analysis efficient in our game without having negative side effects; elements of the first game are reusable in the second one, which could imply that it is useful to develop a kind of frame-game.

The main contribution of our research for the Amsterdam Police Force is that we have brought about a shift in the discussion on coordination in the regional projects. Commitment and content are more consciously addressed in analysis and design of coordination, in addition to composition and contact which were the dominant issues when we arrived. Furthermore the coordination framework and related paradoxical guidelines can be helpful in managing the challenges related to the transformation from a hierarchical organization to a matrix organization. Examples of these challenges are dealing with the fact that boundaries between the police organization and other societal service organizations are becoming vague and the problem of assessing quality of work in addition to output quantities. To answer the question whether our design approach is valuable for large service organizations in general, we aim at representative power rather than statistical evidence. This means that we do not have to carry out large numbers of case studies, needed to demonstrate the fit of our case organization (the Amsterdam Police Force) with the application area (large service organizations). From our analysis of 6 cases in literature (considering large service organizations involved in clinical work, banking and consultancy) we conclude that the Amsterdam Police Force has been a representative case for the application area of large service organizations. As a result it is likely that our design approach is valuable for service organizations in commercial and non-commercial fields that are facing problems in coordination of distributed work. We argue that these organizations may benefit from our design approach by better understanding the competing forces in coordination, and by identifying, thoughtfully developing and exploring new ways of coordination, in which ICT can help to integrate competing forces. Conclusions The main conclusions of our research are: • Current coordination theories have limited views on some of the identified aspects of coordination (process level, organizational level, work arrangements, dependencies, coordination mechanisms, behavioral issues, ICT support, situational-, organizational- and environmental context). 230

SUMMARY •

• • • • •

The coordination framework provides an integrated view on coordination by discriminating between coordination choices on four dimensions on the process level (composition, commitment, contact, content) and choices on four related dimensions on the organizational level (competence, cohesion, connection, context). ICT has a dual impact on coordination. A multitude of factors (technological, social, psychological and work related) can hamper utilization of opportunities. Consequently, the added results of ICT support on coordination are unclear. The coordination framework helps to untangle the interaction of these multiple factors. It proves that ICT has a positive impact on the dimensions of composition and contact and a negative impact on the dimensions of commitment and content Improving coordination requires addressing all four dimensions of the coordination framework on the organizational and process level. Better choices lead to better coordination. Paradoxical guidelines are more valuable than unilateral guidelines due to the multitude of interfering factors, the competing values existent in performance yardsticks and the consequential situated nature of coordination choices Gaming-simulation is a promising technique for organizational design because the interplay of organizational, technological, social and behavioral issues becomes tangible and consequently the combined impact of unlike costs and benefits are observable in an early stage of the design process.

The most important strength of our research approach is, in our view, the continuous interaction between theory and practice in the descriptive and prescriptive parts of the research. Furthermore our long stay and intense investigation in the police force eased interpretation of data and led to a high level of trust between the researchers and the policemen. We experience that the broad scope of our research is necessary to fully grasp and understand the interaction of all factors influencing coordination of distributed work. When others in our field would adopt such a broad scope, the challenge is to sustain reasonable rigor in the presence of high degrees of relevance. Recommendations The design approach can be further improved by conducting studies in other organizations, by including the fifth implementation step in those explorations, by further elaboration of dilemmas and paradoxical guidelines present in the dimensions of the coordination framework, by developing simulation-games that last several days to better simulate elapsed time and by exploring how creativity and lessons from process management could enrich our way of controlling. In the field of computer supported cooperative work some important issues to be addressed are: trying to integrate or clarify the differences between the multiple definitions of coordination; developing performance indicators that discriminate between successful and unsuccessful balancing of conflicting demands; developing guidelines for productive collaborative business engineering; and shifting research questions from comparing face to face with computer mediated coordination, to comparing good work design with bad work design.

231

Samenvatting Coördinatie van gedistribueerd werk Het verkennen van situationele coördinatie met spelsimulatie 4 december 2003 Joeri van Laere Deze dissertatie laat nieuw licht schijnen op het intensief bestudeerde maar nog grotendeels onontgonnen terrein van coördinatie van gedistribueerd werk. De belangrijkste onderzoeksuitkomsten zijn een integraal coördinatie raamwerk ten behoeve van analyse, een aantal paradoxale richtlijnen voor de verbetering van coördinatie van gedistribueerd werk en het demonstreren van simulatiespelen als waardevolle ontwerpinstrumenten. Onderzoeksprobleem Door de eeuwen heen hebben organisaties te maken gehad met de uitdaging om (gedistribueerd) werk te organiseren en de inspanningen van (verspreide) organisatieleden te coördineren. De groter wordende ontwikkelingssnelheid van informatie- en communicatietechnologie (ICT) in de afgelopen 20 jaar heeft nieuwe uitdagingen gecreëerd rond dit thema. Op het eerste gezicht lijken er allerlei kansen te ontstaan om bekende coördinatieproblemen eenvoudiger op te lossen. Door de nieuwe technologie te benutten kunnen coördinatie, samenwerking en het werk zelf onafhankelijk van tijd en plaats uitgevoerd worden. Deze onafhankelijkheid leidt tot meer flexibiliteit voor individuele werknemers in het uitvoeren van hun taken, en kan daarmee de individuele prestatie, de groepsprestatie en de prestaties van de organisatie verbeteren. In de praktijk blijkt het echter moeilijk te zijn om deze ICT-kansen te benutten. De meeste organisaties lukt het niet om de kansen om hun coördinatie te verbeteren volledig te benutten. Zij hebben de neiging om zich vast te houden aan hun gebruikelijke manier van werken en op de traditionele communicatiemedia te vertrouwen. Computerondersteunde samenwerking en coördinatie in gedistribueerde werksituaties wordt gezien als ingewikkelder en minder effectief dan face-to-face coördineren en samenwerken. ICT brengt dus voordelen en problemen met zich mee en de uitkomst van die twee samen kan onduidelijk zijn. Onderzoekers en managers in de organisatiepraktijk worstelen met het integreren van ICT, taken, structuren en menselijk gedrag in organisaties. Bovendien 233

COORDINATING DISTRIBUTED WORK constateren we een aantal tekortkomingen in het huidige onderzoek en de huidige begripsvorming. Allereerst is er een gebrek aan integratie van de bevindingen van diverse onderzoeksdisciplines, zoals organisatiekunde, informatiesysteem onderzoek en groepsdynamica. Ten tweede worden onderwerpen zoals het genereren van betekenissen, relationele aspecten, aantrekkingskracht en sociale en politieke context onderbelicht in de conceptualisatie van coördinatie en communicatie. Ten derde is er een dominantie van experimentele onderzoeken terwijl groepswerk sociaal, organisationeel en in de tijd ingebed is. Tenslotte zien ontwerpbenaderingen, die de coördinatie van gedistribueerd werk willen verbeteren, menselijke gedragsfactoren als implementatieprobleem, in plaats van als een deel van het socio-technische systeem dat herontworpen moet worden. Onderzoeksvragen en doelstelling Wij stellen dat er behoefte is aan meer geïntegreerde visies en benaderingen voor het coördineren van gedistribueerd werk. Onze onderzoeksvraag, Hoe actoren in grote dienstverlenende organisaties de coördinatie van hun gedistribueerde werkprocessen kunnen organiseren, gegeven de kansen en potentiële nadelen van informatie- en communicatietechnologie?, wordt beantwoord door aandacht te besteden aan de volgende vier subvragen: • Hoe kunnen we coördinatie conceptualiseren? • Wat is het effect van ICT-ondersteuning op coördinatie? • Hoe kunnen we de verbetering van coördinatie sturen? • Wat is de toegevoegde waarde van spelsimulatie als ontwerpinstrument? Het uiteindelijke doel van dit onderzoek is om een ontwerpaanpak te ontwikkelen die actoren in grote dienstverlenende organisaties begeleidt in het verbeteren van de coördinatie van hun gedistribueerde werkprocessen, gebaseerd op een geïntegreerde visie op coördinatie. We streven deze doelstelling na door literatuuronderzoek te doen naar de onderwerpen ‘gedistribueerd werk’ (hoofdstuk 1) en ‘coördinatie’ (hoofdstuk 3), door coördinatie van gedistribueerd werk in de praktijk te bestuderen (hoofdstuk 4), door een geïntegreerde visie op coördinatie van gedistribueerd werk te ontwikkelen (hoofdstuk 5), door de gewenste ontwerpaanpak te ontwikkelen (hoofdstuk 6), door de ontwerpaanpak toe te passen in twee actieonderzoeken (hoofdstuk 7) en door de waarde van de ontwerpaanpak te evalueren (hoofdstuk 8). Onderzoeksontwerp Ontologisch volgt ons onderzoek een interpretatieve filosofie en inductief redeneren, waarbij waarnemingen in de realiteit subjectief geïnterpreteerd worden. Wij zijn in dit onderzoek te werk gegaan volgens een inductief-hypothetische onderzoeksstrategie die geschikt is voor theorie ontwikkeling en het bestuderen van slecht gestructureerde problemen. Deze strategie bestaat uit vijf stappen. We starten met een literatuurverkenning die resulteert in een grof initieel theoretisch raamwerk gebaseerd op bestaande coördinatietheorieën. Deze initiële theorie geeft richting aan onze observaties in de twee participatieve case studies. Deze case studies hebben als doel ons begrip te vergroten wat coördinatie inhoudt en te identificeren wat de ervaren kansen en uitdagingen voor ICT zijn om gedistribueerde werkprocessen te ondersteunen (Initiatie). 234

SAMENVATTING De case studie observaties en de resultaten van het aanvullende literatuuronderzoek resulteren in een aantal eisen voor onze visie op coördinatie en voor onze ontwerpaanpak (abstractie). Vervolgens ontwikkelen we een geïntegreerde visie op coördinatie waarin factoren die de verbetering van coördinatie van gedistribueerd werk bevorderen of belemmeren opgenomen zijn. Daarna formuleren we een ontwerpaanpak door de wijze van denken, modelleren, werken en beheersen te beschrijven, waarbij de nadruk ligt op de richtlijnen voor het verbeteren van coördinatie en op spelsimulatie als ontwerpinstrument (theorie formulering). We voeren twee actieonderzoeken uit om de waarde van de ontwerpaanpak te demonstreren en te kunnen evalueren (implementatie). Tenslotte reflecteren we op onze ervaringen in de actieonderzoeken (evaluatie). In ons onderzoek zijn case studies de primaire onderzoeksinstrumenten tijdens de initiatie en implementatie fase. We hebben ervoor gekozen om alle case studies in één organisatie uit te voeren omdat het belangrijk is om lang genoeg in een organisatie aanwezig te zijn om te kunnen begrijpen wat er zich afspeelt. Deze organisatie is de Regiopolitie Amsterdam-Amstelland. De politieorganisatie worstelt met het faciliteren van samenwerking en kennisuitwisseling tussen geografisch gedistribueerde wijkteams. Communicatie en coördinatie in de regionale projecten (een soort communities of practice die politiemensen uit verschillende wijkteams met een gemeenschappelijk onderwerp verbinden) is zowel veelbelovend als teleurstellend. Tijdens alle case studies worden zowel kwantitatieve als kwalitatieve data verzameld om een rijke representatie van de onderzochte fenomenen, en het vergelijken en contrasteren van de verzamelde data mogelijk te maken (meerdere bronnen van bewijslast). Coördinatie theorie Aangezien we er naar streven om bij een geïntegreerde visie op coördinatie uit te komen verkennen we hoe benaderingen uit verschillende wetenschapsdisciplines naar coördinatie kijken. We beperken onze verkenning tot de volgende benaderingen: • Het informatieverwerkingsparadigma, contingentie theorieën, het managen van taakactor afhankelijkheden en de socio-technische systeemtheorie van de discipline organisatiekunde; • ‘Business process redesign’ benaderingen van de discipline informatiesysteem onderzoek; • theorieën en benaderingen op het gebied van groepsprestatie en -ontwerp in face-toface situaties en computerondersteunde situaties van de discipline groepsdynamica; • en theorieën als ‘enactment’, betekenisgeving en situationele actie en coördinatie die hun wortels hebben in de disciplines van cognitieve psychologie, mens-computer interactie en sociale psychologie. Uit deze analyse concluderen we dat coördinatie start met werkverdeling die resulteert in bepaalde ordeningen van werk. Deze werkarrangementen creëren afhankelijkheden tussen taken en actoren die gemanaged moeten worden door coördinatiemechanismen. Coördinatie vindt plaats op organisatieniveau, groepsniveau en individueel niveau. Verder zijn er aspecten van coördinatie die aandacht verdienen, zoals gedrag, ICT-ondersteuning en (situationele, organisationele en omgevings-) context. Wij concluderen dat geen enkele theorie of benadering al deze aspecten meeneemt.

235

COORDINATING DISTRIBUTED WORK Coördinatieproblemen kunnen voortkomen uit onzekerheid (dominant in rationele visies op coördinatie) en ambiguïteit (benadrukt in sociaal psychologische visies op coördinatie). Bovendien is het zo dat de sociaal psychologische benaderingen het problematische karakter van communicatie en besluitvorming benadrukken. Ten gevolge van het complexe karakter van coördinatie zijn de onderzochte theorieën en benaderingen vrij vaag over het sturen en meten van verbetering. De conceptualisering van coördinatie en coördinatieproblemen beïnvloedt sterk de richting van verbetering. De belangrijkste consequenties voor onze case studies zijn dat we moeten nagaan of al de geïdentificeerde elementen, aspecten en niveaus van coördinatie en bronnen van coördinatieproblemen herkenbaar zijn in de praktijk. Verder kunnen in de praktijk aanwijzingen gevonden worden hoe de verschillende opvattingen over coördinatie geïntegreerd kunnen worden. Tenslotte is het interessant om te kijken hoe problemen met onzekerheid en ambiguïteit in de dagelijkse coördinatie bij de Regiopolitie AmsterdamAmstelland opgelost worden. Verkennende case studies bij de Regiopolitie Amsterdam-Amstelland Twee verschillende gedistribueerde werkprocessen in de Amsterdamse politieorganisatie worden geanalyseerd om te ontdekken welke typen coördinatiemechanismen, coördinatieproblemen, ICT-kansen en bevorderende en belemmerende factoren onderscheiden kunnen worden. Dit zijn respectievelijk de werkgroep Jeugdoverlast en het regionaal project Vuurwapens. De werkgroep Jeugdoverlast is een groep die bestaat uit 6 regionaal coördinatoren die een advies moeten geven hoe er met Jeugdoverlastproblemen in District 6 omgegaan moet worden. Wij analyseren de interne coördinatie in de werkgroep en de individuele en groepsinteracties met mensen uit District 6. De meeste interne coördinatie treedt op in 10 groepsbijeenkomsten die ongeveer 1 keer per week georganiseerd worden. Het regionaal project Vuurwapens bestaat uit alle politieagenten in de organisatie die betrokken zijn bij het opsporen van wapens en vuurwapenverdachten, het afhandelen van wapens en vuurwapenverdachten en de informatieanalyse en regionale kennisuitwisseling met betrekking tot vuurwapens. Deze politiemensen werken in verschillende afdelingen. Eén regionaal coördinator stimuleert de kennisuitwisseling en zorgt dat de verschillende organisatieonderdelen die betrokken zijn soepel samenwerken en dat hun inspanningen afgestemd zijn op die van externe partijen zoals landelijke afdelingen en justitie. In beide gevallen zien we dat het onderscheid tussen werkarrangementen, afhankelijkheden en coördinatiemechanismen herkenbaar is. Echter, door het veelvoud aan afhankelijkheden en coördinatiemechanismen die aanwezig zijn in de politieorganisatie is er niet een duidelijk en vast coördinatiemechanisme aanwijsbaar voor elke afhankelijkheid die gemanaged moet worden. In plaats van een vastgelegd ontwerp dat uitgevoerd wordt, zien we dat er redelijk wat vrijheid is voor politiemensen om bepaalde coördinatiemechanismen meer te benadrukken dan andere. Deze coördinatiekeuzes zijn nodig om met de complexiteit en dynamiek van de organisatieomgeving om te kunnen gaan. Door verschillende coördinatiekeuzes te maken 236

SAMENVATTING kunnen politiemensen hun coördinatie aanpassen aan situationele omstandigheden. Bovendien zien we dat onzekerheids- en ambiguïteitsproblemen in de praktijk samen voorkomen en dat daarom gecombineerde oplossingen nodig zijn. De invloed van ICT is divers. ICT-ondersteunde coördinatiemechanismen bestaan samen met hun traditionele tegenhangers (werkprocedures in een handboek en op intranet), of ze maken voorheen niet realistische coördinatiemechanismen nu haalbaar (bezoeken versus opbellen of mailen van een expert aan de andere kant van de stad). Op die manier resulteert ICT in een nog grotere reeks van coördinatiemechanismen om uit te kiezen. ICT is voornamelijk geschikt om onzekerheidsproblemen op te lossen, maar kan ook bijdragen aan het oplossen van ambiguïteitsproblemen. Verder kan ICT zowel bruikbaar zijn voor gedistribueerde groepen als voor groepen die op één locatie samenwerken. Ongebalanceerd gebruik van ICT kan leiden tot averechtse effecten. Voor traditionele en ICT-ondersteunde coördinatiemechanismen geldt dat een veelvoud aan interfererende technologische, sociale, organisationele en persoonlijke factoren de succesvolle toepassing van coördinatiemechanismen kan belemmeren, wat de coördinatie verslechtert in plaats van verbetert. De waargenomen coördinatiekeuzes blijken het voornaamste startpunt voor verbetering van coördinatie te zijn. Het in overeenstemming brengen van de huidige benadrukte coördinatiemechanismen met de situationele omstandigheden en met de nu overheersende afhankelijkheden is nuttiger dan te proberen om één constant coördinatiemechanisme vast te stellen dat uitgevoerd moet worden. In de praktijk zijn coördinatiekeuzes gebaseerd op het vergelijken van potentiële kosten en baten van verschillende coördinatiemechanismen. Deze kosten en baten kunnen niet exact bepaald worden, maar worden grof geschat. Ze hebben betrekking op de afstemming met de afhankelijkheden die gemanaged moeten worden, op de situationele factoren en op verwachte moeilijkheden om de coördinatiemechanismen toe te passen (bijvoorbeeld de belemmerende technologische, sociale, organisationele en persoonlijke factoren). Een eerste eis voor onze visie op coördinatie en onze ontwerpaanpak die volgt uit deze analyse is dat coördinatiekeuzes een centraal aandachtspunt moeten zijn in de conceptualisatie van coördinatie: betere coördinatiekeuzes leiden tot betere coördinatie. Verder moet de conceptualisatie van coördinatie aandacht besteden aan onzekerheidsproblemen en ambiguïteitsproblemen. Tenslotte moet onze ontwerpaanpak niet alleen het identificeren van ICT-kansen mogelijk maken, maar moet de aanpak ook duidelijk maken hoe deze kansen succesvol benut kunnen worden, door middel van het inzichtelijk maken van potentiële kosten en baten. De invloed van de vele technologische, sociale, organisationele en persoonlijke factoren op ICT-adoptie moeten in deze kosten en baten meegenomen worden. Coördinatie raamwerk Met behulp van de literatuurverkenning en de case studie observaties ontwikkelen we een geïntegreerde visie op coördinatie die geëxpliciteerd is in het coördinatieraamwerk (zie figuur 4, pagina 98). Het raamwerk illustreert dat uitvoering van werk resultaten oplevert voor een bepaalde opdracht. Daarbij worden coördinatiekeuzes gemaakt op groepsniveau 237

COORDINATING DISTRIBUTED WORK op vier dimensies (door werknemers) en vier gerelateerde coördinatiekeuzes gemaakt op organisatieniveau voor elke dimensie (door managers). Dientengevolge is coördinatie de combinatie van: • het samenstellen van een groep van medewerkers met de geschikte organisationele competenties; • het verzekeren van commitment binnen deze groep ondersteund door een algemene organisatiecohesie; • het maken van contacten tussen groepsleden mogelijk gemaakt door aanwezige organisationele connecties; • en het uitwisselen van inhoud met elkaar geïnterpreteerd in een organisationele context; met als doel het produceren van het gewenste resultaat voor een gegeven opdracht. We zien dat onzekerheidsgerelateerde dimensies (compositie en commitment) verbonden zijn met het aantal relaties en het aantal afstemmingen. Ambiguïteitsgerelateerde dimensies (commitment, inhoud) zijn verbonden met de kwaliteit van relaties en de kwaliteit van de afstemmingen. Uit ons overzicht van kosten, baten, kansen en bedreigingen concluderen we dat de invloed van ICT over het algemeen positief is voor het oplossen van onzekerheidsproblemen en negatief voor het aanpakken van ambiguïteitsproblemen. Of ICT echter als een kans of bedreiging uitpakt hangt in hoge mate af van hoe de medewerkers in de organisatie de technologie toepassen in hun dagelijkse werk en coördinatie, en van de situationele omstandigheden. Als medewerkers zich bewust zijn van de ICT-mogelijkheden dan wordt de kans groter dat ze omgezet worden in daadwerkelijke voordelen. Op dezelfde manier helpt het kennen van de bedreigingen hen die bedreigingen te vermijden. Onze ontwerpaanpak geeft aan hoe de goede balans in de coördinatiekeuzes gevonden kan worden. De ontwerpaanpak De ontwerpaanpak wordt gekarakteriseerd door zijn denkwijze, zijn modelleer concepten, zijn manier van werken en zijn beheerswijze. Het belangrijkste doel van de denkwijze in onze ontwerpaanpak is richting te geven hoe tegengestelde krachten in de coördinatie van gedistribueerd werk, zoals onzekerheids- en ambiguïteitsproblemen en kansen en bedreigingen van ICT-ondersteuning gebalanceerd kunnen worden. Coördinatiekeuzes hebben kosten en baten en de totale uitkomst is onduidelijk en wisselt van situatie tot situatie. Een keuze tussen twee alternatieven die beide vervelende gevolgen hebben of die beiden even attractief zijn wordt een dilemma genoemd. Een dilemma is per definitie in theorie niet op te lossen, maar kan in de praktijk gehanteerd worden door er mee om te leren gaan. Dit betekent dat de paradox begrepen moet worden en dat de leerervaringen bijeengebracht moeten worden in de zoektocht naar dilemma-resistente-arrangementen. Van de concurrerende waarden benadering leren we dat richtlijnen, die het balanceren van gelijktijdig tegengestelde zaken benadrukken, vaak geformuleerd worden als paradoxen. Wij formuleren 12 richtlijnen die duidelijk maken hoe tegengestelde eisen in de coördinatie van gedistribueerd werk gemanaged kunnen worden.

238

SAMENVATTING Dit zijn (zie ook tabel 44, pagina 118): Compositie / competentie 1. Raadpleeg gedistribueerde expertise uitgebreid, maar houd groepen klein, gefocust en productief 2. Selecteer medewerkers op aanwezige standaardcompetenties en op benodigde unieke competenties 3. Zet experts in op specialistische en algemene taken Commitment / cohesie 4. Stimuleer sociale communicatie en het formuleren van duidelijke doelen vroeg in het proces 5. Rouleer het leiderschap onder groepsleden 6. Echte steun vereist een kritische houding Contact / connectie 7. Maak regelmatig contact om voortgang te tonen en te evalueren, maar onderbreek productieve actie niet 8. Laat snelle informatie-uitwisseling tijd bieden voor interactieve interpretatie 9. Situationele (onvoorspelbare) mediakeuze vereist standaard communicatie regels voor alle media Inhoud / context 10. Wees kort en krachtig maar laat niet de relevante details achterwege 11. Voeg ontbrekende context toe en laat afleidende context weg 12. Laat experts de consultatie van kennissystemen begeleiden

Voor onze modelleerwijze gebruiken we twee modelleringstechnieken. Taak-actor modellen voldoen aan de behoefte om een snel een overzicht te creëren van bestaande werkprocessen en coördinatiemechanismen en van schetsen van mogelijke toekomstalternatieven. Daarnaast biedt spelsimulatie de mogelijkheid om nieuwe werkprocessen, nieuwe coördinatiemechanismen, nieuwe ICT, nieuwe sociale- en gedragsnormen en toekomstige organisatiecontext als een betekenisvol geheel te ervaren. Door de toekomstige situatie te beleven kunnen organisatiemedewerkers de nieuwe coördinatiealternatieven grondiger evalueren en ervaring opdoen hoe om te gaan met het coördinatieraamwerk en de paradoxale richtlijnen. De manier van werken bestaat uit vijf stappen, namelijk: het analyseren van de huidige werkprocessen, het identificeren van veelbelovende kansen voor gedistribueerd werken, het ontwerpen van de coördinatie van gedistribueerd werken, het ontwikkelen en spelen van simulatiespelen en het implementeren van nieuwe werkprocessen. Om het leren te faciliteren moet de manier van beheersen ‘middle out’ (met de focus op een klein maar kritiek deel van de organisatie dat herkenbaar is) en incrementeel (snelle oplevering van een initieel model dat gereviseerd en verbeterd kan worden in een aantal cycli) zijn. Op die manier kan snel feedback verzameld worden, die niet alleen goed is voor het leren en voor de kwaliteit van het ontwerp, maar die ook helpt om de steun van management en medewerkers voor het veranderingsproces te krijgen. Actie onderzoek Het eerste actieonderzoek richt zich op het gebruik van gedistribueerde expertise in het dagelijkse probleemoplossingsproces in wijkteams. Het tweede actieonderzoek gaat over de coördinatie van het vullen en onderhouden van de ‘regionale projecten intranet sites’ en het beantwoorden en verifiëren van antwoorden op de discussielijsten van de regionale projecten. In beide studies passen we de eerste 4 stappen van onze ontwerpaanpak toe (analyseren van huidige werkprocessen, identificeren van kansen, ontwerp van coördinatie en ontwikkelen en spelen van een simulatiespel). Tijdens de identificatie van kansen, tijdens het ontwerpproces en tijdens het spelen en evalueren van de simulatiespelen worden de paradoxale richtlijnen gebruikt. In het eerste actieonderzoek 239

COORDINATING DISTRIBUTED WORK wordt het simulatiespel negen keer gespeeld met politiemensen van de Regiopolitie Amsterdam-Amstelland. In het tweede actieonderzoek wordt het simulatiespel twee keer gespeeld. Beide studies resulteren in talrijke leerervaringen voor de Regiopolitie Amsterdam-Amstelland. In de eerste studie blijkt bijvoorbeeld dat politiemedewerkers teveel op hun eigen expertise vertrouwen, terwijl het consulteren van andere bronnen (zoals intranet of andere experts) tot een betere prestatie leidt. Het intranet bevat teveel algemene en langdradige informatie die moeilijk toepasbaar is op praktijkproblemen. De breedte en diepte van de kennis van experts verschilt erg veel, terwijl weinig informatie voorhanden is om de meest geschikte expert te lokaliseren die met de oplossing van het actuele probleem kan helpen. Tenslotte zal de discussielijst alleen intensief gebruikt worden als de snelheid en kwaliteit van de antwoorden gegarandeerd kan worden. In de tweede studie blijkt dat de regionale projecten zich meer richten op kwaliteit dan op snelheid, terwijl de wijkteams (hun ‘klanten’) zowel kwaliteit als snelheid eisen. Ook wordt geconcludeerd dat het aanstellen van zowel centrale als lokale webmasters de behoefte aan een standaardnavigatiestructuur versus de behoefte aan lokale wijzigingen en kwaliteitscontrole balanceert. Tenslotte wegen de voordelen van een kleine gedistribueerde kerngroep van experts (continue bezetting, grotere bijdrage van expertise) zwaarder dan de kosten van de meer gecompliceerde coördinatie en het gevaar van ontbrekend commitment. Reflectie Met betrekking tot de bruikbaarheid en het nut van de verschillende elementen van onze ontwerpaanpak (coördinatieraamwerk, paradoxale richtlijnen en simulatiespelen) in de twee actieonderzoeken concluderen we het volgende: • Het coördinatieraamwerk structureert de ontwerpdiscussies; • het raamwerk helpt om kansen, bedreigingen, kosten en baten te identificeren; • richtlijnen kunnen aangepast worden aan het actuele coördinatievraagstuk; • in het algemeen is er een positieve ICT-invloed op de compositie en contact dimensies en een negatieve op de commitment en inhoudsdimensies; • spelsimulatie geeft inzicht of en wanneer de baten tegen de kosten opwegen; • ééndimensionale richtlijnen schieten tekort voor het verbeteren van coördinatie: politiemensen bevestigen de noodzaak van paradoxale richtlijnen om tegemoet te komen aan de conflicterende eisen in hun werk; • commentaar tijdens spelsimulaties is veel specifieker dan tijdens het ontwerp tengevolge van het daadwerkelijk toepassen van de richtlijnen in plaats van er alleen maar in het algemeen over te praten; • meerdere doelen worden behaald tijdens de simulatiespelen: ze dienen als ontwerpen als trainingsinstrument; • hoewel game-experts adviseren om computergebruik in simulatiespelen te beperken, maakt computergebruik wel de logistiek en analyse efficiënt in ons simulatiespel, zonder negatieve bijeffecten; • elementen van het eerste simulatiespel zijn herbruikbaar in het tweede, wat erop zou kunnen wijzen dat het zinvol is om een soort ‘frame-game’ te ontwerpen.

240

SAMENVATTING De belangrijkste bijdrage van ons onderzoek voor de regiopolitie Amsterdam-Amstelland is dat we een verschuiving in de discussie over coördinatie in de regionale projecten teweeg gebracht hebben. Er wordt nu meer bewust aandacht besteed aan commitment en inhoud in het analyseren en ontwerpen van coördinatie, naast compositie en contact welke de dominante onderwerpen waren toen wij ons onderzoek startten. Verder kunnen het coördinatieraamwerk en de gerelateerde paradoxale richtlijnen nuttig zijn bij het managen van uitdagingen bij de transformatie van een hiërarchische organisatie naar een matrixorganisatie. Voorbeelden van die uitdagingen zijn het omgaan met het feit dat grenzen tussen de politieorganisatie en andere maatschappelijke organisaties vervagen en het probleem om de kwaliteit van het werk te meten in aanvulling op de kwantitatieve outputcijfers. In antwoord op de vraag of onze ontwerpaanpak waardevol is voor grote serviceorganisaties in het algemeen streven we naar representatieve kracht van onze resultaten in plaats van statistische bewijslast. Dit betekent dat we niet een groot aantal case studies hoeven uit te voeren, maar dat we de overeenkomst tussen onze case (de regiopolitie Amsterdam-Amstelland) en het toepassingsgebied (grote dienstverlenende organisaties) moeten laten zien. Uit onze analyse van 6 cases uit de literatuur (betreffende grote dienstverlenende organisaties in klinisch werk, bankieren en organisatieadvisering) concluderen we dat de Regiopolitie Amsterdam-Amstelland een representatieve case is voor het gebied van grote dienstverlenende organisaties. Op basis hiervan formuleren we de hypothese dat onze ontwerpaanpak waardevol is voor grote dienstverlenende organisaties in de commerciële en niet-commerciële sector. We stellen dat deze organisaties van onze ontwerpaanpak kunnen profiteren door de concurrerende waarden met betrekking tot coördinatie te leren begrijpen en door nieuwe manieren van coördineren te identificeren, grondig te ontwerpen en te verkennen in simulatiespelen, waarbij ICT kan helpen om de concurrerende waarden te integreren. Conclusies De belangrijkste conclusies van ons onderzoek zijn: • Bestaande coördinatietheorieën hebben beperkte visies op sommige van de geïdentificeerde aspecten van coördinatie (procesniveau, organisatieniveau, werkarrangementen, afhankelijkheden, coördinatiemechanismen, gedragsaspecten, ICT-ondersteuning, situationele-, organisationele-, en omgevingscontext). • Het coördinatieraamwerk levert een geïntegreerd beeld op van coördinatie door onderscheid te maken tussen coördinatiekeuzes op vier dimensies op procesniveau (compositie, commitment, contact en inhoud) en coördinatiekeuzes op vier gerelateerde dimensies op organisatieniveau (competentie, cohesie, connectie en context). • ICT heeft een tweeledige invloed op coördinatie. Een veelvoud aan factoren (technologische, sociale, psychologische en werkgerelateerde) kan de benutting van kansen belemmeren. Dientengevolge zijn de samengestelde uitkomsten van ICTondersteuning bij coördinatie onduidelijk.

241

COORDINATING DISTRIBUTED WORK •

• •

•

Het coördinatieraamwerk helpt om helderheid te brengen in de chaotische interactie van deze vele factoren. Het blijkt dat ICT een positieve invloed heeft op de dimensies van compositie en contact en een negatieve invloed op de dimensies van commitment en inhoud. Voor het verbeteren van coördinatie moet aandacht besteed worden aan alle vier de dimensies van het coördinatieraamwerk op organisatieniveau en procesniveau. Betere keuzes leiden tot betere coördinatie. Paradoxale richtlijnen zijn waardevoller dan ééndimensionale richtlijnen ten gevolge van het veelvoud aan interfererende factoren, de concurrerende waarden die aanwezig zijn in prestatie indicatoren en het daaruit voortvloeiende situationele karakter van coördinatiekeuzes. Spelsimulatie is een veelbelovende techniek voor organisatieontwerp omdat de interactie van organisationele, technologische, sociale en gedragsaspecten concreet wordt en dientengevolge worden de gecombineerde invloed van de ongelijkwaardige kosten en baten zichtbaar in een vroeg stadium van het ontwerpproces.

De grootste kracht van onze onderzoeksaanpak is ons inziens de continue interactie tussen theorie en praktijk in de beschrijvende en voorschrijvende onderdelen van ons onderzoek. Verder heeft ons lange verblijf en intensieve onderzoek in de politieorganisatie de interpretatie van de data vergemakkelijkt en bijgedragen aan een hoog niveau van vertrouwen tussen de onderzoekers en de medewerkers van de politieorganisatie. We ervaren dat de ruime focus van ons onderzoek noodzakelijk is om de interactie tussen alle factoren die de coördinatie van gedistribueerd werk beïnvloeden volledig te kunnen begrijpen. Als andere onderzoekers in ons veld ook zo’n brede focus zouden hanteren dan is een belangrijke uitdaging voor hen om voldoende onderzoeksnauwkeurigheid te garanderen bij hoge niveaus van relevantie. Aanbevelingen De ontwerpaanpak kan verder verbeterd worden door hem toe te passen in andere organisaties, door de vijfde implementatiestap in deze verkenningen mee te nemen, door verdere uitwerking van de dilemma’s en paradoxale richtlijnen die aanwezig zijn in de dimensies van het coördinatieraamwerk, door simulatiespelen te ontwikkelen die meerdere dagen duren om verstreken tijd beter te simuleren en door te verkennen hoe creativiteit en lessen uit procesmanagement onze wijze van beheersen kunnen verbeteren. In het onderzoeksveld van CSCW (computer supported cooperative work) zijn de volgende onderwerpen de moeite waard om verder te verkennen: de vele definities van coördinatie proberen te integreren of de verschillen proberen te verklaren; prestatieindicatoren ontwikkelen die onderscheid maken tussen succesvol en niet succesvol balanceren van concurrerende waarden of conflicterende eisen; richtlijnen ontwikkelen voor productief collaboratief ontwerp van werkprocessen; en het verschuiven van onderzoeksvragen van het vergelijken van face-to-face met computerondersteund coördineren naar het vergelijken van goed ontwerp van werkprocessen met slecht ontwerp van werkprocessen.

242

Curriculum Vitae Joeri van Laere was born in Terneuzen on July 24th 1974. After graduating from SSG De Rede in Terneuzen in 1992 he studied Industrial Engineering and Management Science at Eindhoven University of Technology. During his study he was head of the organizing committee of the 1996 Study tour to South Africa. His master thesis “Towards a learning supplier; design of a method for self-diagnosis and organization development” was rewarded with the NIVE year price 1998. After his graduation in 1997 he started as a research assistant at the faculty of Technology, Policy and Management of Delft University of Technology. Besides conducting the research described in this dissertation, he facilitated numerous electronic meetings in the Netherlands, India and South Africa and was involved in teaching courses and supervising students working on their master thesis. He presented and published his work on international conferences and in an international journal. Currently, he is working as an assistant professor in the chairgroup Communication Management within the social sciences group of Wageningen University and Research Centre.

243