Knowledge manipulation activities: results of a Delphi ...

Information & Management 39 (2002) 477–490

Knowledge manipulation activities: results of a Delphi study C.W. Holsapplea, K.D. Joshib,* a

School of Management, Carol M. Gatton College of Business and Economics, University of Kentucky, Lexington, KY 40506-0034, USA b School of Accounting, Information Systems and Business Law, College of Business and Economics, Washington State University, P.O. Box 644750, Pullman, WA 991164-4750, USA Received 20 January 2000; accepted 26 April 2001

Abstract Knowledge-based organizations are hosts for multitudes of knowledge management (KM) episodes. Each episode is triggered by a knowledge need and culminates with the satisfaction of that need (or its abandonment). Within an episode, one or more of the organization’ processors (human and/or computer-based) manipulate knowledge resources in various ways in an effort to meet the need. This paper identifies and characterizes a generic set of elemental knowledge manipulation activities that can be arranged in a variety of patterns within KM episodes. It also indicates possible knowledge flows that can occur among the activities. This descriptive framework was developed using conceptual synthesis and a Delphi methodology involving an international panel of researchers and practitioners in the KM field. The framework can serve as a common language for discourse about knowledge manipulation. For researchers, it suggests issues that deserve investigation and concepts that must be considered in explorations of KM episodes. For practitioners, the framework provides a perspective on activities that need to be considered in the design, measurement, control, coordination, and support of an organization’ KM episodes. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Delphi study; Framework; Knowledge management episode; Knowledge; Manipulation; Knowledge-based organization

1. Introduction A hallmark of the emerging knowledge economy is the rise of knowledge-based organizations [14]. In these, knowledge is regarded as a crucial resource processed by a joint human–computer system in changing the organization’ state of knowledge and of producing outputs. Individually, each human or automated processor is a knowledge worker that has a particular set of skills for manipulating knowledge. Collectively, an organization’ knowledge processors are arranged into a system that amplifies the knowledge work to be accomplished. *

Corresponding author. Tel.: þ1-509-335-5722. E-mail addresses: [email protected] (C.W. Holsapple), [email protected] (K.D. Joshi).

Knowledge management (KM) involves attempts to get the right knowledge to the right processor at the right time in the right representation and at the right cost. The task of recognizing and satisfying the needs of a modern organization is both important and challenging. These can be modest or voluminous, simple or complex, routine or novel, well specified or vague, stable or volatile, of low priority or urgent. We shall term what occurs from the time of recognizing a knowledge need through its satisfaction (or abandonment) as a KM episode which may be independent or interdependent with other episodes and active at any given time in an organization. Each involves one or more knowledge processors operating on some knowledge resources and constrained or guided by various influences. Fig. 1 illustrates a KM episode. But what knowledge manipulation activities are allowed in a KM

0378-7206/02/$ – see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 7 2 0 6 ( 0 1 ) 0 0 1 0 9 - 4

478

C.W. Holsapple, K.D. Joshi / Information & Management 39 (2002) 477–490

Fig. 1. Architecture of a KM episode during the conduct of knowledge management.

episode? The answer to this question is important. Indeed, a recent survey found that a majority of respondents preferred an activity-oriented KM [20]. However, there has been little agreement among researchers or practitioners on what they are [29]. This paper presents a descriptive framework of basic knowledge manipulation activities that can occur in an episode. The framework was developed through a Delphi process involving an international panel of over 30 KM researchers and practitioners. The result is a relatively comprehensive, unified perspective on the kinds of knowledge manipulation activities that can occur in a KM episode. This offers several benefits. It can serve as a common language for discussion about an organization’s KM episodes. It gives a foundation for suggesting how each of the knowledge manipulation activities should be accomplished and how they should be configured within episodes. Its characterization of each manipulation activity is suggestive of functionalities that would be helpful to include in the design of computer-based processors for performing or supporting the activity. It could be applied to highlight and investigate KM issues, such as a means for measuring, controlling, and coordinating manipulation activities, etc.

level, while others deal with higher-level knowledge activities. For instance, the activities identified by Arthur and APQC [2], Wiig [31], van der Spek and Spijkervet [30], Alavi [1], and Szulanski [28] appear to be more elemental than those identified by LeonardBarton [17], and Choo [6]. The higher-level activities seem to be groupings of more elemental activities. For example, decision making is an activity that may involve a composite of the more elemental activities identified by Arthur and APQC [2]. An examination of the activities reveals considerable variation. No one view subsumes the others. This suggests there is a need for a generic framework of knowledge manipulation activities that not only describes each activity clearly and completely but also identifies their possible inter-relationships. Here, the focus is on elemental activities (and their sub-activities) rather than higher-level, composite ones. The focus is on activities that directly manipulate knowledge and produce knowledge flows within a KM episode, rather than activities that start or control the episode. The latter managerial influences on KM episodes have been addressed elsewhere [13]. This framework can be applied to multiple concepts of knowledge.

2. Background

3. Methodology

A comparative analysis of KM frameworks in the literature indicates that they identify various KM activities [11]. These are summarized in Table 1. Some frameworks treat these activities at an elemental

Through a synthesis of concepts, best practices, and issues in the literature, an initial descriptive framework of knowledge manipulation activities was developed. This evolved through a Delphi process, involving two

C.W. Holsapple, K.D. Joshi / Information & Management 39 (2002) 477–490 Table 1 Summary of knowledge management activities identified in the literature Author

Knowledge management activities

Alavi [1]

Acquisition (knowledge creation and content development) Indexing Filtering Linking involves screening, classification, cataloging, integrating, and interconnecting internal and external sources) Distributing (packaging and delivery of knowledge in form of Web pages) Application (using knowledge)

Arthur and APOC [2]

Share Create Identify Collect Adapt Organize Apply

Choo [6]

Sensemaking (includes ‘‘information interpretation’’) Knowledge creation (includes ‘‘information transformation’’) Decision making (includes ‘‘information processing’’)

Holsapple and Whinston [14]

Procure Organize Store Maintain Analyze Create Present Distribute Apply

Leonard-Barton [17]

Shared and creative problem solving Importing and absorbing technological knowledge from the outside of the firm Experimenting prototyping Implementing and integrating new methodologies and tools

Nonaka [21]

Socialize (convert tacit knowledge to tacit knowledge) Internalize (convert explicit knowledge to tacit knowledge) Combine (convert explicit knowledge to explicit knowledge) Externalize (convert tacit knowledge to explicit knowledge)

Szulanski [28]

Initiation (recognize knowledge need and satisfy that need) Implementation (knowledge transfer takes place) Ramp-up (use the transferred knowledge) Integration (internalize the knowledge)

van der Spek and Spijkervet [30]

In the act process Develop Distribute Combine Hold

Wiig [29]

Creation Manifestation Use Transfer

479

480


rounds of critique and evaluation by a panel of KM practitioners and academicians. 3.1. Preliminary phase Boundary conditions, framework evaluation criteria, and evaluation standards were determined prior to starting development of the framework. Three boundaries were defined: business, descriptive, and detail. The business boundary confined the development to a consideration of KM within business organizations. The descriptive boundary confined it to describing knowledge manipulation activities, rather than prescribing their configurations in the conduct of KM. The framework was developed in a top–down fashion. Because this process could not continue indefinitely, a detail boundary was set as two levels, i.e. to identify basic knowledge manipulation activities at one level and their sub-activities at the second. Four criteria were selected for evaluating the framework: completeness, correctness, conciseness, and clarity. These criteria are similar to those used for theory evaluation [3,15]. Each criterion was applied in guiding development of the initial framework. Each Delphi iteration involved answering questions assessing the framework with respect to the four criteria, while incorporating revisions to address panelists’ comments. An iterative approach was used in developing an initial framework to account for notions of elemental knowledge manipulation found in a survey of the literature (e.g. [8–10,21,22,24–27]). This approach is similar to the process of matching where concepts, ideas, language along with their inter-relationships are synthesized, organized, and unified in an inductive fashion over multiple iterations [16]. A Delphi study was conducted to gather KM researchers’ and practitioners’ perspectives and to revise the framework. This also provided an independent expert assessment of the framework. The Delphi methodology used here is similar to that used by Bacon and Fitzgerald [4] in developing an information technology framework. Our framework development involved two rounds, at which point all suggestions were either outside the development boundaries or not deemed sufficiently important to warrant a third round. In the fall of 1996, researchers and practitioners who had contributed to the KM literature and presented at KM conferences were identified for partici-

Table 2 Primary perspectives for viewing KM Primary perspectives

Frequency (%)

Information systems Management Strategic management Computer science Public administration Philosophy Cognitive sciences/artificial intelligence Finance Human centered design Communication Economics Management science Organizational behavior Sociology Innovation strategy Value creation

22 13 13 9 9 9 3 3 3 3 3 3 3 3 3 3

pation on the KM panel. Those for whom a mailing address could be readily identified became the 122 candidates for the Delphi panel. This number seemed to be reasonable for the purpose of capturing a variety of viewpoints (Table 2). Of the 122 candidates, 31 (25.4%) agreed to read the framework description and respond. Their principal work activities covered five continents, with the majority in North America. Their primary business interests were manufacturing, service, consulting, and education. Panelists approached the field from diverse perspectives, with even balance of researchers and practitioners (43% for each, with remainder in both). At the time of the study, panelist experience in the KM field ranged from 1 to 15 years, with a majority having at least 5 years of experience. All panelists were active contributors to the field (e.g. via articles, books, and presentations); 40% of the panelists had at least 10 KM-related publications and 50% had done KM presentations at conferences at least 10 times. A questionnaire was designed to elicit comments on the framework (its comprehensiveness, completeness, clarity, and conciseness). The questionnaire was pilot tested and the feedback was used to refine its content. The final instrument consisted of both open-ended questions for structured elicitation and Likert-scale items; a seven-point scale (from ‘‘not at all successful’’ to ‘‘extremely successful’’) was used. Open-ended questions allowed panelists to express dissatisfaction,


reasons for it, and suggestions on ways to make improvements. The questionnaire, with a letter of invitation to participate, a self-addressed return postcard, a paper describing the initial framework, and a self-addressed postpaid return envelope was mailed to each candidate. All were given 12 weeks to reply. The responses were captured and organized in three database tables—one for demographics, one for numeric responses, and one for open-ended answers. Responses to open-ended questions were organized by questionnaire item and carefully reviewed and evaluated. The first round of data analysis divided the responses into two groups: (1) to be considered in framework revision and (2) beyond the research boundaries. Comments in the first group were further classified into (1) concerns that were repeated and/or seemed to be of major importance, (2) concerns that were less frequent and/or less important and (3) concerns that occurred infrequently and/or seemed relatively unimportant. A summary response analysis document was prepared organizing critiques into these categories for each of the questionnaire items. This response analysis document was used to guide revision of the initial framework. Three types of revisions were made: (1) fundamental; (2) additive and (3) clarifications. The fundamental modifications were extensive and entailed developing new concepts stimulated by participants’ comments, detailing and further characterizing the concepts existing in the initial framework, and further justifying the elements. Additive changes introduced new elements, describing the nature of each and its relationships with other elements. Clarification was needed when an element was already present but panelist comments indicated a need to explain it more clearly or give it more emphasis. Panelists from the first round were invited to participate in the second round. Each received a mailing comprised of an invitation letter, a paper describing the revised framework, the response analysis document, a questionnaire, and a self-addressed postagepaid envelope. Panelists were given 8 weeks to respond. Out of the 31 panelists from the first round, 17 (55%) responded in the second. Data analysis was similar to the first round. The resultant response analysis document showed substantial agreement with the round-two framework in terms of both qualitative and quantitative assessments.

481

4. A descriptive framework of knowledge manipulation activities The conduct of KM in an organization results in learning and projection, thereby adding value. The conduct of KM is guided and shaped by managerial influences; enabled and constrained by environmental influences and organizational resources. In a KM episode, processors use their knowledge handling skills to perform manipulation on knowledge resources. That is, the mechanisms and results of knowledge manipulation activities are expressions of processors’ knowledge manipulation skills. Skill is the ability to apply one’s knowledge effectively and readily to execution and performance [19]. Participants’ knowledge manipulation skills differ widely, but we do not attempt to classify the diversity that differentiates them. Instead, we focus on the activities that are executed and performed using those skills. The framework identifies a set of inter-related knowledge manipulation activities that appear to be common across diverse organizations. Identification of these activities forms a starting point for understanding how knowledge is processed in organizations and how this changes over time. They highlight major activities with which a chief knowledge officer (CKO) needs to be concerned. Participants’ knowledge manipulation skills need to be cultivated, harnessed, and organized in the performance of these activities. Fig. 2 identifies the major activities of acquiring, selecting, internalizing, and using knowledge. The latter refers to the activities of externalizing and generating knowledge. Arrows indicate major knowledge flows. For instance, execution and performance of the acquisition activity entails a knowledge flow into that activity from the environment and a consequent flow to the internalizing or using activity. The internalizing activity produces a knowledge flow that impacts the state of the organization’s knowledge resources; externalizing impacts the environment: these are learning and projection, respectively. Aside from the main knowledge flows, activities can send and receive ancillary messages. (For simplicity, these messages are not represented in the figure.) An example is a request from one activity to another (e.g. the generating activity requests a knowledge flow from the selection activity). Requests can range from procedural (specifying how the activity should be

482


Fig. 2. Major knowledge manipulation activities.

carried out) to nonprocedural (merely indicating what is needed). They can range from explicit (e.g. a command) to implicit (e.g. involving recognition of a need). Some may require fast responses; others tolerate performance of an activity in the background. They can range from one-time requests to standing requests that require continual monitoring. Another example is a feedback message, where an activity comments on the value of a knowledge flow that is received. Each instance of knowledge manipulation can be performed by one participant or by some configuration of multiple participants. Conversely, a participant may exercise knowledge manipulation skills to perform multiple activities within a KM episode. For instance, a person (or a computer system) may participate in

both acquiring and generating knowledge within a single KM episode. Managerial, resource, and environmental factors affect how the knowledge manipulation skills of an organization’s participants are deployed to accomplish the activities. These factors also influence the pattern of activities that occur within a specific KM episode: which are performed, how many instances are performed, which one is first, which is last, etc. In all cases, the framework furnishes the following characterization of the activities. 4.1. Acquiring knowledge This refers to the activity of identifying knowledge in the environment and transforming it into a


representation that can be internalized, and/or used. Sub-activities include the following. Identifying appropriate knowledge from external sources. This includes locating, accessing, valuing, and/or filtering knowledge from outside sources. Capturing identified knowledge from outside. This involves extracting, collecting, and/or gathering knowledge deemed to be sufficiently valid and useful. Organizing captured knowledge. This involves distilling, refining, orienting, interpreting, packaging, assembling, and/or transforming captured knowledge into representations that can be understood and processed by another knowledge manipulation activity. Transferring organized knowledge. This involves communication channel identification and selection, scheduling, and sending. This can be to an activity that immediately uses the knowledge or internalizes it for subsequent use. There are many issues related to the acquisition activity. These include whether to acquire knowledge (e.g. outsource) or generate it in-house; preventing knowledge overload (e.g. avoid acquiring knowledge that does not add significant value to the organization; developing procedures and guidelines for effective performance of each of the sub-activities (e.g. reduce costs associated with execution of each sub-activity). 4.2. Selecting knowledge This is the activity of identifying needed knowledge within an organization’s existing knowledge resources and providing it in an appropriate representation to an activity that needs it (i.e. to an acquiring, using, or internalizing activity). This activity is analogous to acquisition, except that it manipulates resources already in the organization. However, the two activities can require different skills, levels of effort, and costs. Sub-activities include the following. Identifying appropriate knowledge within the organization’s existing resources. This includes locating, accessing, valuing, and/or filtering knowledge. Capturing identified knowledge, which involves retrieving, collecting and/or gathering it from the organization’s knowledge resources.

483

Organizing captured knowledge. This involves distilling, refining, orienting, interpreting, packaging, assembling, and/or transforming captured knowledge into representations appropriate for subsequent manipulation. Transferring organized knowledge to an appropriate activity (e.g. the one that requested it). This involves determining the medium through which knowledge is transferred, identification and choice, scheduling, and sending it. This flow can support the acquiring, internalizing, or using activities. Numerous issues concerned with knowledge selection include how to adapt skills at acquiring to the activity of selecting and vice versa; how to perform the selection activity on knowledge resources that embody implicit/tacit knowledge; what techniques/ tools to employ to make the selection activity effective and efficient for each of the main types of knowledge resources; ensuring that the needed knowledge reaches participants that need it and does so when it is needed. 4.3. Internalizing knowledge This involves incorporating or making the knowledge a part of the organization. The framework views internalizing as an activity that alters an organization’s knowledge resources by acquiring, selecting, or generating the knowledge. It receives knowledge flows from these activities and produces flows that impact the organization’ state of knowledge. Internalizing knowledge is a culminating activity in organizational learning. Sub-activities include the following. Assessing and valuing knowledge to be internalized. This involves determining the suitability of the knowledge, which depends on its degree of assessed utility and validity. Targeting knowledge resources. This identifies knowledge resources that are to be impacted by the knowledge flow produced by internalization. Structuring knowledge. This represents knowledge to be conveyed in the appropriate forms for the targets. Delivering the knowledge representations as targeted. This involves modifying existing knowledge resources (adding to them, deleting from them, increasing the organizational density, or perhaps

484


fundamentally altering a knowledge resource). It involves depositing, storing, updating, disseminating, distributing, and sharing knowledge with respect to targeted knowledge resources. This also involves channel identification and choice, scheduling, and sending. 4.4. Using knowledge This is the activity of applying existing knowledge to generate new knowledge and/or produce an externalization of knowledge. Where the activity of using knowledge exists, there is the possibility that organizational learning is innovative and adds value. 4.4.1. Generating knowledge This is an activity that produces knowledge by processing existing knowledge, which has resulted from selection, acquisition, and/or prior generation. The knowledge generated may be new to the organization. This is crucial because ‘‘new knowledge provides the basis for organization renewal and sustainable competitive advantage’’ [23]. Alternatively, it may currently exist or have previously existed in the organization. Generation of knowledge that is not ‘‘new’’ can occur for economic reasons (e.g. it is cheaper to generate than to select), for training reasons, for validity checking, due to a lack of awareness about its existence, or due to not having been internalized when previously acquired or generated. Subactivities include the following. Monitoring the organization’s knowledge resources and the external environment by invoking selection and/or acquisition activities as needed. Evaluating selected or acquired knowledge in terms of its utility and validity for the production of knowledge. Producing knowledge from a previously existing base. This can involve creating, synthesizing, analyzing, and constructing knowledge. Transferring the produced knowledge for externalization and/or internalization. This involves channel identification and choice, scheduling, and sending. Broadly, there are two types of generation: derivation and discovery. Derivation involves the use of

process knowledge (e.g. procedures, rules) and descriptive knowledge (e.g. data, information) to generate process and/or descriptive knowledge. It employs KM skills that are of an analytical, logical, and constructive nature. In contrast, discovery generates knowledge in less structured ways, via skills involving creativity, imagination, and synthesis. The exact path from the initial to the discovered knowledge cannot always be preconceived or traced. Herman Helmholtz, a German physiologist and physicist, described his scientific discoveries as progressing through three stages: saturation (finding out everything he could learn on a subject), incubation (reflecting on what has been absorbed, by thinking about and mulling over what he has learned through the research), and illumination (arriving at a sudden solution). French mathematician Henry Poincare added a fourth stage to this—verification [5]. Issues related to generation include what methods or techniques facilitate production of knowledge; how to effectively interface with selection and acquisition activities for efficient monitoring and to minimize evaluation required during generation; how to coordinate-related instances of generation; means of learning not only in the sense of generating knowledge but also in the introspective sense of how to better generate knowledge in the future, learning about generation itself and internalizing the result (e.g. as process knowledge) for subsequent selection; guidelines for governing whether needed knowledge should be acquired, selected, or generated. 4.4.2. Externalizing knowledge The activity of making something available outside the organization is termed externalization. Applying this to an organization’s KMC, the framework views externalizing knowledge as the activity that uses existing knowledge to produce organizational outputs for release into the environment. It yields projections (i.e. embodiment of knowledge in outward forms) for external consumption, in contrast to internalization which may also yield projections, but which are retained as knowledge resources. Externalization is only partially a knowledge manipulation activity because it can involve physical activities, such as the act of producing a product through transformation of raw materials. However, the flow of material can be seen as secondary to the flow of knowledge that


enables, facilitates, and guides it [7]. Sub-activities include the following. Targeting the output. This involves determining what needs to be produced for targeted elements of the environment. Producing the output. This involves applying, embodying, controlling, and leveraging existing knowledge to produce output for the target. The output is a representation of the knowledge used to produce it. Transferring the output. This is concerned with packaging and delivering the projections that have been produced for targets in the environment. This involves channel (medium through which knowledge is transferred) identification and choice, scheduling, and sending. Externalization results in projections. When an organization transfers an output (e.g. in the form of products, services, knowledge artifacts), it is projecting. The process of effective projection adds value to an organization in forms such as profits, image, customer loyalty, and visibility. Once externalization occurs, its impact (e.g. in the forms of sales, etc.) can be captured through the knowledge acquiring activity. This environment interaction results in a feedback loop. Externalization issues include how to innovatively apply generated knowledge for the purpose of external projections; how to mesh or intersperse instances of generation into an instance of externalization; means of identifying appropriate targets and the nature of knowledge acquisition effort with respect to them; understanding the differences and commonalties between externalization and internalization.

5. Analysis of Delphi responses In examining panelist reactions to the framework, no major or crippling reservations were detected. However, several concerns related primarily to pushing beyond framework boundaries were expressed. 1. Three of the 17 respondents had reservations about referring to generation (or some aspect of generation) as a kind of use activity. Some were unclear about the interplay of generation and externalization.

485

Comment: This is largely a presentational issue. The use activity can be further explained to support inclusion of generation and externalization within it. Also, if framework users choose to separate generation and externalization from the ‘‘use’’ umbrella, there is essentially no effect other than eliminating the ‘‘use’’ term and losing some conciseness. The trade-off between clarity and conciseness should be considered in preparing any alternative presentation of the framework. 2. Each of the following activities was perceived by one or another panelist as missing from the activities. The ‘‘conversion from knowledge to knowhow’’. Organizing, categorizing, storing, and sharing knowledge. Unlearning. ‘‘Location of stakeholder value’’. Sharing/socializing, creating, capture/storage, and learning. Comment: As these points were very infrequently made, we do not interpret them as being major. Most of these elements are present in the framework. However, they may need to be presented more clearly or highlighted. Those that are absent (location of stakeholder’ value and knowledge conversions) do not appear to be elemental knowledge manipulation activities, but may perhaps be explained in terms of the existing activities. 3. The following points were identified by one or another panelist as needing clarification or more explanation. Tacit KM should be explored further. The figure depicting knowledge manipulation activities and their relationships need to be further explained. Especially the knowledge flows. Elaborate on the sequencing of knowledge flows. Need to explain the difference between knowledge acquisition and selection. Clarify the distinction between using and internalizing knowledge. More discussion of value capture. Comment: None of these points was raised by more than one panelist. Considering each of the activities with respect to tacit (or explicit) knowledge is an interesting future research direction, as are similar considerations with respect to other

486

4.

5.

6.

7.


attribute dimensions of knowledge. The framework focuses on activity characterizations and their relationships (in the form of knowledge flows). The generic framework does not advocate any particular sequencing among the knowledge manipulation activities. One respondent suggests that more attention needs to be given to the issues of knowledge replenishment (e.g. adding new knowledge, discarding irrelevant knowledge). Comment: This is an aspect of the framework’s internalization activity. However, types of alteration, criteria for alteration, and mechanisms to execute alteration are not detailed within the description of the internalization activity. For instance, there are several types of knowledge alteration (e.g. addition, deletion, updates, modification, shifts). One respondent states: ‘‘This is yet another model of KM activity when what is needed is a synthesis of the best of many existing models . . .. Also, Nonaka’ model is elegant and simple, and works at both individual and organizational learning level with four activities’’. Comment: The initial framework was constructed by synthesizing existing frameworks and other ideas from the literature. The framework reported here built upon this by further synthesis of panelists’ suggestions and responses to their critiques. Future study may investigate how the generic framework’s activities relate to those of more specialized frameworks such as those of Nonaka. The intent here is to be generic rather than focus on a particular attribute dimension of knowledge (e.g. the tacit versus explicit dimension). One respondent asks: How does the ‘‘data– information–knowledge–skill (wisdom) information–food–chain’’ fit into this activity model? What is the relation between ‘‘information’’ and knowledge? Comment: Little consensus exists on distinctions between ‘‘information’’ and ‘‘knowledge’’ and it is not the purpose of this framework to take a position, but rather to be open and able to accommodate multiple positions. One respondent says: ‘‘It is unclear why these six activities are selected and why they are treated as separate’’.

Comment: In the interest of conciseness and a bottom-line focus, such explanation is omitted. In addition, panelists quantitatively evaluated the framework in terms of the four criteria. Graphical displays of participants’ responses to Likert-scale items in the second round are presented in the figures. All respondents rate the framework as at least somewhat successful on all criteria. For every criterion, a preponderance of respondents rate the framework as at least moderately successful. Fig. 3 shows relative frequency distributions of responses for each criterion. Over 80% of respondents indicate a moderate or higher degree of success with respect to every criterion. Moderate satisfaction is the mode for each of the criteria. A ‘‘majority of respondents gauge the framework’ completeness and conciseness as being in the successful to extremely successful range’’. A ‘‘majority of respondents gauge the framework’s accuracy and clarity as being in the moderately successful to extremely successful range’’. Fig. 4 presents the relative frequency distributions for the four criteria. This is an alternative way of looking at results, showing that a preponderance of responses are at least at the moderately successful level across all four criteria. 6. Implications This paper presents a generic framework of basic knowledge manipulation activities that operate on an organization’s knowledge resources with KM episodes. Identification and explanation of knowledge manipulation activities and their inter-relationships allows for better understanding of the nature and the dynamics of activities that manipulate an organization’ knowledge resources. Each of the four knowledge manipulation activities and any of their sub-activities can be further characterized an analyzed in greater detail. For instance, such an analysis has been performed for the knowledge selection activity [12]. It fleshes out concepts of selection functionalities in greater detail. Based on these, it identifies issues related to knowledge selection, uses the framework’ concepts to organize a characterization of current technological offerings for knowledge selection, and describes the interaction of all the other activities with the selection activity.


Fig. 3. Responses for each of the framework’s criteria.

Fig. 4. Responses for each of the evaluation measures.

487

488


Similar analysis can be carried out for each of the other activities. Alteration of knowledge resources through knowledge valuing (i.e. what knowledge should be attained, retained and replenished) is a very serious issue faced by organizations [18]. This is an aspect of the framework’s internalization activity. However, types of alteration, criteria for alteration, and mechanisms to execute alteration are not detailed within the internalization activity. The framework’s activities are not confined to manipulation of explicit knowledge, but can also be performed on tacit/implicit knowledge. The manipulation skills and mechanisms needed to carry out a particular activity on one type of resource can substantially differ from those useful for another type of knowledge resource. This framework can be used to systematically generate, study, and discuss, KM-related issues. The framework provides a platform for communication and sharing of ideas related to knowledge manipulation activities among practitioners. The common

Debra Amidon Sulin Ba Thomas J. Beckman Kesper Deboer Marc Demarest Alain Godbout Valerie Cliff Ming Ivory Linda Johnson Mark A. Jones Sam Khoury Kai Larsen Dirk Mahling Eunika Mercier-Laurent Philip C. Murray Brian Newman David Paradice Gordon Petrash Dave Pollard Larry Prusak David Skyrme Charles Snyder Kathy Stewart Karl Sveiby

language can aid practitioners in administering KM initiatives by addressing issues in a systematic fashion (e.g. the types of activities that need more attention, types of skills and competencies, tools and techniques that need to be cultivated and developed to execute these activities). The framework’s characterization of knowledge manipulation activities is descriptive, aiming to identify relevant activities and their knowledge–flow relationships. It does not advocate any particular methodology or process for coordinating these activities. They can be combined in various configurations in order to define a process or methodology.

Acknowledgements Funding for this research was provided by Kentucky Initiative for Knowledge Management. We are indebted to the following persons for their participation as Delphi panelists. Those who participated in every round are indicated by an asterisk ().

ENTOVATION International, Ltd., USA University of Southern California, USA George Washington University & IRS, USA Andersen Consulting, USA The Sales Consultancy, USA Godbout Martin Godbout & Associates, Canada ICL Enterprise Consultancy, UK James Madison University, USA Western Kentucky University, USA Andersen Consulting, USA The Dow Chemical Company, USA Center for Technology in Government, USA University of Pittsburgh, USA EML Conseil—Knowledge Management, France Knowledge Management Associates, USA The Newman Group & The KM Forum, USA Texas A&M University, USA The Dow Chemical Company, USA Ernst & Young, Canada IBM Corporation, USA David Skyme Associates Limited, England Auburn University, USA Georgia State University, USA Sveiby Knowledge Management, Australia


Robert Taylor Karl Wiig Andrew Whinston Fons Wijnhoven Dennis Yablonsky Michael Zack David Paradice One participant prefers to remain anonymous.

489

KPMG Management Consulting, UK Knowledge Research Institute, Inc., USA University of Texas, Austin, USA University of Twente, The Netherlands Carnegie Group, Inc., USA Northeastern University, USA Texas A&M University, USA

References [1] M. Alavi, KPMG Peat Marwick U.S.: One Giant Brain. Harvard Business School (Case), 9-397-108, Review on 11 July 1997. [2] A. Arthur, The American Productivity and Quality Center (APQC), The Knowledge Management Tool: External Benchmarking Version, 1996. [3] S. Bacharach, Organizational theories: some criteria for evaluation, Academy of Management Review 14 (6), 1989, pp. 496–515. [4] J. Bacon, B. Fitzgerald, The Field of IST: A Name, A Framework and A Central Focus, Executive Systems Research Center Working Paper Series, Reference: 96/5, 1–35, 1996. [5] B. Carson, Think Big—Unleashing Your Potential for Excellence, Harper-Paperbacks, New York, 1992. [6] C. Choo, An Integrated Information Model of the Organization: The Knowing Organization, http://www.fis.utoronto.ca/ people/faculty/choo/FIS/KO/KO.html1#contents, 7 December 1996. [7] D. Cook, C. Chung, C. Holsapple, Information flow first, Material flow next! APICS—The Performance Advantage (1995) 38–39. [8] Ernst, Young, Innovation in Action—Selling Knowledge on the Net. Perspectives on Business and Innovation—Managing Organizational Knowledge, http://www.businessinnovation. ey.com/journal/features/whykno/body.html, 2 October 1997a. [9] Ernst, Young, Twenty Questions on Knowledge in the Organization. A Joint Survey by Business Intelligence and the Ernst & Young Center for Business Innovation, Top-Line Finding, http://www.businessinnovation.ey.com/research/ knowle/survey/topline.html, 2 October 1997b. [10] G. Hedlund, A model of knowledge management and the Nform corporation, Strategic Management Journal 15, 1994, pp. 73–90. [11] C. Holsapple, K. Joshi, Description and analysis of existing knowledge management frameworks, in: Proceedings of the Hawaiian International Conference on System Sciences, Maui, January 1999a. [12] C. Holsapple, K. Joshi, Knowledge selection: concepts, issues, and technology, in: J. Liebowitz (Ed.), Handbook of Knowledge Management, CRC Press, Boca Raton, FL, 1999b. [13] C. Holsapple, K. Joshi, An investigation of influences on the conduct of knowledge management, Kentucky Initiative for Knowledge Management, Research Paper No. 124, College of Business and Economics, University of Kentucky, 1999c.

[14] C. Holsapple, A. Whinston, Knowledge-based organizations, The Information Society 5 (2), 1987, pp. 77–90. [15] F. Kerlinger, Foundations of Behavioral Research, Harcourt Brace Jovanovich College Publishers, New York, 1986. [16] G. Krogh, J. Roos, K. Slocum, An essay on corporate epistemology, Strategic Management Journal 15, 1994, pp. 53–71. [17] D. Leonard-Barton, Wellsprings of Knowledge, Harvard Business School Press, Boston, 1995. [18] J. Liebowitz, K. Wright, Valuation of human capital as a component of knowledge assets, in: Proceedings of the Association of Information Systems Conference, Baltimore, MD, 1998. [19] Merriam Webster’s Collegiate Dictionary, 10th Edition, Merriam-Webster Incorporated, Springfield, 1995. [20] P. Murray, A. Myers, The facts about knowledge, Information Strategy Online, http://www.info-strategy.com/knowsur1, 26 October 1998. [21] I. Nonaka, The knowledge creating company, Harvard Business Review (1991) 96–104. [22] G. Petrash, Dow’s journey to a knowledge value management culture, European Management Journal 14 (4), 1996, pp. 365– 373. [23] J.B. Quinn, The Intelligent Enterprise, Free Press, New York, 1992. [24] G. Rifkin, Fast Company, Buckman Labs Is Nothing but Net, 9 April 1997, http://www.fastcompany.com/03/buckman. html, 25 April 1997. [25] R. Ruggels, The big idea—why knowledge? why now? Perspectives on Business Innovation—Managing Organizational Knowledge, http://www.businessinnovation.ey.com/ journal/features/whykno/body.html, 2 October 1997. [26] T. Stewart, Intellectual Capital, Doubleday/Currency, New York, 1997. [27] K. Sveiby, What Is Knowledge Management? 4 March 1997, http://www.svieby.com.au/KnowledgeManagement.html, 4 April 1997. [28] G. Szulanski, Exploring internal stickiness: impediments to the transfer of best practice within the firm, Strategic Management Journal 17 (Winter Special Issue), 1996, pp. 27–43. [29] S. Tan, H. Teo, B. Tan, K. Wei, Developing a preliminary framework for KM in organizations, in: Proceedings of the Association for Information Systems Conference, Baltimore, MD, 1998. [30] R. van der Spek, A. Spijkervet, Knowledge management: dealing intelligently with knowledge, in: J. Liebowitz, L.

490


Wilcox (Eds.), Knowledge Management and Its Integrative Elements, CRC Press, New York, 1997. [31] K. Wiig, Knowledge Management Foundations, Schema Press, Arlington, 1993. Clyde W. Holsapple has held the Rosenthal Endowed Chair in MIS and directed the Kentucky Initiative for Knowledge Management for more than a decade. He has served as editor of the Journal of Organizational Computing and Electronic Commerce, area editor of Decision Support Systems and the ORSA Journal on Computing, and associate editor of Management Science. He is editor of Springer’s new Handbook on Knowledge Management. His earlier books include Decision Support Systems: A KnowledgeBased Approach, The Information Jungle: A Quasi-Novel Approach to Managing Corporate Knowledge, Business Expert Systems, and Foundations of Decision Support Systems. Dr. Holsapple has published over 100 scholarly articles.

Kshiti D. Joshi is an Assistant Professor in the School of Accounting, Information Systems, and Business Law at Washington State University. She holds a BA in Mathematical Statistics and an MA in Operational Research from the University of Delhi. She also earned an MS degree in Industrial and Operations Engineering from the University of Michigan. Dr. Joshi holds a PhD in Decision Science and Information Systems from University of Kentucky. Her research articles have been accepted for publication in Decision Support Systems, Information Systems Journal, The Information Society, Knowledge Management Handbook, and Handbook of Electronic Commerce, Journal of Strategic Information Systems. She has been awarded an NSF grant to study gender differences in information systems career choice.