Hyperknowledge and Expert Systems: A Case Study of ... - CiteSeerX

Proceedingsof the 28th Annual Hawaii InternationalConferenceon SystemSciences- 1995

Hyperknowledge and Expert Systems: A Case Study of Knowledge Formation Processes Pirkko Walden and Christer Carlsson Institute for Advanced Management Systems Research(IAMSR) Abe Akademi University, DataCity A 3208,20520 Bibo, Finland by the unfortunate anomaly that numerical precision and relevancefor strategic decisions tend to become conflicting considerations (51. There is a more general problem behind this observation: imprecision is the basis for richer, more meaningful and better information [6], [7]. The second issue is the ongoing controversy between various schools of thought in strategic management [3]; one of the dominating schools seemsto argue against the use of computer-supportedsystems (cf. Mintzberg [3]) which are describedto produce information which lack in richness,is too aggregated,is late and unreliable.

Abstract Strategic management is defined as the system of action programs which form sustainable competitive advantages for a corporation, its divisions and its business units in a strategic planning period. We have developed a system called Woodstrat to serve as a support system for these action program activities on both the corporate, the divisional and the business unit levels. The system is modular and is built around the logic of strategic management. The innovation in Woodstrat is that these modules are linked together so that essential competitive elements interact to form the strategic management context. The intermodular links are based on expert knowledge on strategic management; and user-developed expert knowledge is also used to guide the users trough the process of working out sustainable competitive advantages. The process is made user-supportive with a hyperknowledge user interface.

l

l

l

l

1. Introduction Among managers strategic management is normally understood to cover both the strategic planning process and the implementation of its results. It is an integrated program of means by which a firm securesand sustains competitive advantages[I]. This involves understanding and transforming some proper and consistentselection of strategic managementconceptsinto specific strategic action programs. In our specific context, the Finnish forestry industry, this translates in to working out how to position strategic businessunits of a corporation such that the corporation can build and sustain competitive advantagesin a very tough Europeanmarket. Strategic managementhas been the focus of much controversy [2], [3], [4] during the last few years. The first issue is that much of the information and knowledge used by senior managersin strategic management,but also in toplevel planning and decision making, is characterised

hard information is often limited in scope, lacking richness and often failing to encompass important noneconomicand nonquantitativefactors; much hard information is too aggregatedfor effective use in strategy making; much hard information arrives too late to be of use in strategy making; a surprising amount of hard information is unreliable;

In his criticism of the Mintzberg approach to strategic management Ansoff [2] describes this as aimed at an implicit strategy formation, which amounts to strategy and capabilities being allowed to evolve organicalty, through trial and error, and thus without any explicit strategy formulation. This emergent strategy approach relies on the interaction of people in an organisation to eventually come up with a joint strategy. Nevertheless,strategic managementis done in practice, quite successfully.In our experience the process is often complex and time consuming, and there are a number of moments which could be made more effective and with better quality if some support systemtechnology would be applied. This belief is stated despite the Mintzberg observation [3] that nothing much could be gained with computer support. As a starting point, we need to construct a conceptual skeleton: strategic management is the process through which a company for a chosenplanning period (i) defines its operational context, (ii) outlines and decides upon its

73 1060-3426/W

$4.00 Q 1996 IEEE

Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95) 1060-3425/95 $10.00 © 1995 IEEE

Proceedingsof the 28th Annual Hawaii International Conferenceon SystemSciences- l!J95

shtegic goals and long-term objectives,(iii) exploresand decidesupon ita strengths, weaknesses,opportunities and threats, (iv) formulates ita sustainablecompetitive advantagesand (v) developsa program of actions,which exploit its competitive advantages and ensure profitability, financial balance, adaptability to sudden changes and a sound developmentof its capital structure. In order to show that it could be done, we have constructed a hybrid object-oriented expert and hyperknowledge systemto serve as a support systemfor the strategic managementprocess.We will show that the hybrid system allows both systematicmodelling and adaptive, interactive learning by combining Mint&erg’s criticism of strategic planning with the ideas of cognitive mapping and hyperknowledge (section 2); then we will demonstrate that the system is useful as a support system for strategic management(section 3), and finally summa&e the knowledge formation processeswe have seen taking place among tbe managers who have worked with the system in a series of studies in 1993-94. This paper is a continuation of previously published research work (cf. PI, [91, WI, N-1, WI). - -

The assumption of quantifiation: the strategy making processcan be driven by hard data, and qmmtitative aggregates of detailed facts about the organisation and its environment. The assumption ofpredetermination: this emphasisesthe (wrong) belief that managementneedspreplanned actions in order to act wisely; this is possible only if the context for strategy making is stable, or at least predictable, the processitself as well as its consequences(strategies)can he predetermined. Having outlined his path of attack Mintzberg then commits himself to develop arguments contrary to each assumption,and thus to show that the rationality assumed in strategic planning is a “grand fallacy”: because anulysis is not synthesis, strategic planning is not strategy formation.

The Mintzberg description of strategic planning is - with some simplification - a processin which the four asaumptions are used to formulate some basic scenarios which then is linked with a standard, numerical planning framework, i.e. something like the following construct (cf. fig. 1):

2. The M intzberg fallacy, strategy formation and cognitive mapping The Ansoff-Mint&erg controversy may seem academic and without interest for the construction of support systems in strategic management;unfortunately enough this is not the case - the doubt the controversy has created about the usefulness of computer support reappears as resistanceto the use of any kind of knowledge basedsuppoti systems.The only remedy seems to be to take issue with the key arguments and show that they are wrong; there is also a bonus to be gained in the process- we can verify that the support system we have constructed will avoid most of the common pitfalls in the discussion. In a recent book Mintzberg [3] expands his criticism of the most commonly used approachesto strategic planning. He finds a number a limitations and problems with strategic planning per se, which he summarises as the following key assumptions(cf. [3], pp 222-225): The assumption of formalisation which states that the strategymaking processcan be programmedby the use of systems. This amounts to a general assumption that analysis can provide synthesis, which Mintzberg finds to be a major leap of faith. The assumption of detachment: if a strategic planning system does the thinking then thought is assumedto be detachedfrom action, strategy from operations,ostensible thinkers from real doers, and “strategists”from the objects of their strategies.

Fig. 1: Strategic planning fmmework. An alternative strategy formation processwould be one in which a strategy emerges as a result of working through and articulating strategic visions, which are derived from qualitative belief systems. The notion of an emerging strategy could be describedin the terms of fig. 2; we have tried this notion in a series of studies with the Woodstrat in 1993-94: (i) in interactive mode1building seminars with the managers, in which the basic logic of the system was formed; (ii) in studies of the knowledge basesused for actual strategic planning, and (iii) in workshops with the managers,in which they built their strategic plans.

Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95) 1060-3425/95 $10.00 © 1995 IEEE

Proceedings of the 28th Annual Hawaii International Conference on System Sciences -

Fig. 2: Strategy

formation

1995

members of an organisation participating in a strategy formation process. Every strategy is constrained and enhanced by a network of other strategies, which makes for a complex and ill-structured context. Cognitive maps when formalised with computer-supported tools - can be used to show and simulate the interaction and the interdependencies of multiple belief systems. There is a technology available to accomplish this - it has been called hyperknowledge. The idea of creating a hyperknowledge environment is a fairly recent one (cf. [13], [14], [15]). Chang et al [13] introduced it with a cognitive metaphor, ” . . . a decision maker cognitively possesses many diverse and interrelated pieces of knowledge (i.e. concepts)... “which appears close to our metaphor of the strategy formation process (cf. fig.2). There are some useful characteristics of a hyperknowledge environment (cf. [ll]): (i) the user can navigate through and work with diverse concepts; (ii) concepts can be different epistemologically, (iii) concepts can be organised in cognitive maps, (iv) the concepts can be made interrelated and interdependent, (v) relations can be structured or dynamic, and (vi) relations can change with or adapt to the context. There are also a couple of problems with hyperknowledge (cf. [12]): (vii) it has turned out to be too informal and unstructured for handling complex problems, and (viii) users get lost in a conceptually overrich environment, i.e. they lose touch with the task they try to accomplish. Some ways to handle these problems have been shown in [ll], and have been incorporated in the knowledge-based support system discussed here.

process.

The managers have visions of how to create and articulate sustainable competitive advantages, which converges as a synthesis of visions of a market position (MP) competitive position (CP), productivity (PRO), investments (INV) and cost effectiveness (CES); the visions are enhanced or restricted by facts about the context, which identify the market potential (EXT), key competitive factors (by SWOT), the competitors (COMP) and the set of possible strategies (STRAT); the visions and the facts are combined to generate options and form strategic action programs which are defined both for the strategic business unit (SBU) and corporate (CORP) levels; finally, the visions and the strategic action programs are evaluated in terms of profitability and capital structure (specified as an income estimate (INC), a balance estimate (BAL) and financing (FIN)) which represent stakeholder interests. This description is quite close to Eden’s [16] formulation of the process, but there is also a major difference between Eden and Mintzberg: the latter “proves” the fallacies of strategic planning by quoting various sources, and stating that they have obvious errors, in order to “show” that strategic planning must fail; the former shows a practical way to cope with complexity in an orderly and systematic fashion, and proceed to demonstrate that qualitative belief systems can both be analysed and synthesised; he, furthermore, demonstrates that this can be done with computer support. In our work with the Woodstrut also we have seen that modern information technology makes it possible for line managers to use formal procedures to fonn novel strategies for their key products in their most important market segments; this in real time and in anticipation of strategic moves by their key competitors. Colin Eden [16] describes cognitive maps as “actionoriented representations of the world”. They are built on the premise that assertions about strategic issues should have consequences and implications; these are built on qualitative belief systems, which may differ between the

3. Woodstrat The Woodstrat was developed as a series of prototypes in 1992-94; after the first versions with a Lisp-based expert systems shell proved to be too inflexible, the next versions were built with Toolbook, which introduced the notion of Izyperknowfedge. The present full-scale system was built as a hybrid system in Visual Basic in which the features of the Lisp and Toolbook versions were rebuilt as objects. With Visual Basic it was possible to fully exploit the graphical user interface (GUI) technology; we have used multiple-document interface (MDI), object linking and embedding (OLE), dynamic data exchange (DDE), effective graphics and the possibilities to add custom controls by calling procedures in dynamic-link libraries (DLLs). The finished application is an .exe file that uses a runtime DLL. Let us work through the Woodstrut system step by step: Business Unit I is one of the SBUs (the name and all the figures are fictitious); it is operating in several countries, with well-defined product groups and specified customer segments. Markets and segments differ for different prod75

Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95) 1060-3425/95 $10.00 © 1995 IEEE

Proceedings of the 28th Annual Hawaii International Conference on System Sciences -

1995

uct groups, and their importance varies over the planning period; the 1993 volumes and prices are used as starting points (the two small input boxes in the lower left hand corner, cf. fig. 3) in order to calibrate the estimates to be given. The strategic Market Position (MP) is detennined hierarchically: segments are defined for each product group and product groups are selected for each country; for each segment demand and price development estimates are made and consolidated to product group and country levels. The weighted average of the growth and price development estimates update the Net sales line in the income statement (cf. fig 7) through functional links when the CEO-button is activated. In this way the market visions are formulated and immediately evaluated in terms of the income statement’ . Fig. 4 : Competitive

Fig. 3: Market

Position

position. Fii. 5: Competitors

The Woo&rat is supported with a fairly extensive database of country specific economic indicators and related forecasts, and market and segment specific forecasts on the development of price and demand levels. We found out that this helps the SBU managers to calibrate their assessments of growth and price developments. The Competitive Position (CP) is activated (with logical links) from the same level as the MP, and the MP and CP are worked out in parallel. The CP is determined in terms of critical success factors (these are SBU-specific and were determined successively in a series of seminars with the managers) by assessing the relative changes to the previous year, and as changes to the CP worked out for the previous year. This process is one of reassessing visions on MP when evaluated against critical success factors and the relative strengths of the competitors.

Three selected competitors are evaluated on the same critical success factors; as this is a benchmarking approach, quite a lot of time was used to select “good” competitors; CSF and CP averages were determined for the competitors; the relative difference in competitive positions is calculated and transferred with logical links to the CP, where it is used as a basis for assessing relative strategic CPs. There are functional links to MP, which are used to calculate an estimated development in volumes and prices; this function is only voluntary as the SPU managers can always make their own estimates and thus override the suggested development. This process anchors visions of relative CPs to facts (as they are known) about competitors. A summary of relative strategic CPs and the expected development of the studied markets is shown in a summary graph (cf. fig. 6), which now is the first summarised, visual consequence of the MP and CP visions.

cf the principles shown in fig. 2. 76

BEST COPY AVAILABLE Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95) 1060-3425/95 $10.00 © 1995 IEEE

Proceedings of the 28th Annual Hawaii International Conference on System Sciences - 1995

I50

Total Sales

IQ0

I*W-B

IQ

l!!fi!z I994 199s

Q

1996 1997 1996

Year

Market Development

/+ghfZT 1994

I995

1996

1: 199)

199s

Fig. 7: Income Fig. 6: Market

Statement

Graph

The main benefit with the automatic linking of the modules of the CEO Report is that the managers do not have to interrupt their work on strategic assessments in order to check on the profitability and the financing - a task which is hard on non-specialists. The links have also turned out to be major time-savers. As Return on Net Assets (RONA) is a forestry industry standard, a graphical simulation module for RONA allows the managers to quickly find critical sales or operating cost levels for reaching target levels of the RONA (cf. fig. 8):

The third element, the Production Position (PR0)2, estimates productivity as a consequence of the MPs and the CPs. Production soM is detemrined and transferred from the growth and price development specified in the MP; productivity is determined as a weighted average of several factors - labour, raw material, electricity, steam and technology. The productivity factors are updated from corresponding cost lines of the income statement (cf. fig 7) through functional, knowledge based links. The productivity and profitability measures are numerical consequences of key success factors of the MP and CP visions. The CEO Report is activated from the summary level of MP. Functional, knowledge based links from the MP and a module for raw material costs update the revenue and cost lines of the projected income statement (cf. fig. 7), which is linked with the balance sheet, the statement of funds and the key ratios; all these modules update each others with logical, knowledge based links in a way which follows proper accounting principles. Most of the lines are further specified in more details, but the summary reports are mostly sufficient for strategic planning purposes.

Fig. 8: RONA

Simulation

A feature which won much approval among the SBUmanagers were the integrated modules for simulating investment alternatives and corresponding alternative financial models, which showed their impacts on cash flows and key ratios (cf. fig. 9, 10).

’ cf the principles shown in fig. 2. 77

BEST COPY AVAILABLE Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95) 1060-3425/95 $10.00 © 1995 IEEE

Proceedings of the 28th Annual Hawaii International Conference on System Sciences -

iv.

V.

Fig. 9: Investment

Plan

vi.

vii.

1995

competitors; the results were used to establish relative competitive positions; these positions were then used to establish a new set of reference points for the estimates of price and volume developments; the price and volume estimates were reiterated with the proposed product mix (PM) combinations, which are essential for estimating the use of raw material (a major cost factor) and production capacity; the MP, CP and PM modules are then used to determine the net sales and variable costs of the income statement in the CEO module; when worked out with a complete set of variable and fixed costs the resulting profit is either satisfactory or nonsatisfactory; in the latter case it triggers reiterations of the MP, CP and PM modules; if satisfactory it is further evaluated in terms of the capital structure, the use of funds and key ratios; reiterations can be triggered on a number of reference points; the CP detennines the need for investments; the corresponding module triggers the financing of these investments; reiterations are triggered by reference points in the income statement, the capital structure, the use of funds and key ratios; the MP and the CEO modules update and trigger the determination of the productivity index; reference points induce reiterations of investments and of the MP and CP modules;

There are a number of minor processes also included in the system, but the major processes shown in (i) - (vii) represent the strategy fonnation process. Fig. 10: Financing

Plan

4. Experiences

of strategy

formation

with

Woodstrut is used for strategy formation, which was seen in the use of the various modules:

Woodstrat.

i.

The Woodstrat was run in 9 SBUs in a series of seminars this spring3 in sessions which involved about 40 managers in a 2-month period and culminated in presentations given by the COOS of the SBUs to the corporate board. In these presentations they used Woodstrut to demonstrate their strategic solutions and to show the internal logic of their plans. As we had included an on-line Memo function in the system and had charged the mangers with continuously commenting on their assessments and conclusions in all modules, we could follow their reasoning. The most effective management team collected 15 pages of information in their Memo.

ii.

.. . III.

the database of exxternal data served as an instrument to establish reference points for growth and price estimates; the MP module served both analysis and synthesis; it was first used to build estimates of price and volume developments for specific product groups in specific market segments; then it was used to get a feeling for the expected development in a product group or in a country; then the estimates were reiterated until some acceptable levels were reached; the CP module was used for similar analysis / synthesis iterations in the same countries, product groups and market segments, but the iterations were now in terms of critical success factors; reference points for these were established with analogous iterations of the critical success factors for three key

3 The system has now been used in an additional 8 SBUs with more than 40 managers participating. 78

Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95) 1060-3425/95 $10.00 © 1995 IEEE

BEST COPY AVAILABLE

Proceedingsof the 28th Annual Hawaii International Conferenceon SystemSciences- 1995 In the following we have briefly summarked some of the

Competitive Position

experienceswe collected from the SBUs; a more detailed report on the findings will be forthcoming.

.

Market Position

.

.

.

.

.

compared with SBUs from other divisions our assessmentsof market and competitive positions are made more in depth; market information will from now on be updated by the marketing managers with assistancefrom the international salesoffices; the Woods&at systemwill be in a continuous use; working through the MP took about 70% of the total planning time; we have never discussedthe markets as thoroughly before;

the problem is to get the information about our competitors; it required a lot of work, and it was difIicult because our competitor information is not sufficient; we would like to have a possibility to change the critical success factors, as we intend to change our product mix during the planning period;

The hyperknowledgeenvironment made the evaluations of relative CPs fast and efficient. A summary of the cornments showed the basic problem to be the lack of sufficient knowledge of the competitors (which in our mind was surprisingly good). The gradual assessmentof relative competitive positions worked very well despite the fact that the assessmentswere basically analytical; the supporting discussion in the Memo showed systematic and consistent reasoning about (i) the abilities of the competitors and (ii) the possibilities to develop critical successfactors through investments and increased productivity.

The MP module helped in determining the expected volume and price development in key market segments. The managerswere able to determine their expectedmarket developmentin more detail than ever before, and also found out that they now have a basis for quick and systematic evaluations of unexpected changes in both volumes and prices. The rather abstract visions building described by Mintzberg turned out to be very practical and precise,and to serve as a basis for strategy formation. The CP module was one of the key modules from the very beginning. In most SBUs it was found out that: (i) the managementteam does not know their key competitors as well as they woulid like to know them; (ii) they are not too well aware of what their critical successfactors are, and what impact various levels of these might have on their successin various market segments,and (iii) they wert not too sure about bow they could improve on these factors or how this improvementwould work out in a chosen planning interval. The CP module offered a possibility to work out all implications of the various factors systematically, and it offered a good basis for explaining their competitive solutions and defending their requirements for investmentsin a manner much more persuasive than before. The CP module required quite a lot of work, which felt tedious from time to time, but it was probably the most rewarding of all the modules.

CEO Report .

. .

working through it took about 10% of the total planning time, very straight-forward; we think the CEO was (in a technical sense)the most complex part of the system; a few more links could be added; interest costs should be activated for reservations,for long-term investments and for corporateoverbeadcosts paid in advance;

The userswere quite familiar with building the elements of the CEO Report as this part of the strategy formation processhas been routine for a number of years. The contribution of Woodstiat was that the knowledgebasedlinks made operations more precise and much faster than before; all required operations were fully automated and a good selection of graphics objects helped to produce effective and useful summary reports. The investments module, with an integrated financing module, was built to help the SBU managers work out their investmentproposals.It was found to save quite a lot of time as complex and detailed investment plans, with a number of alternatives scheduled for a number of years, and their impact on financing, key ratios, capital structure, etc. could be evaluated.

79

Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95) 1060-3425/95 $10.00 © 1995 IEEE

Proceedingsof the 28th Annual Hawaii International Conferenceon SystemSciences- 1995 expected development of demand and prices. It worked basically as expected, but there was a further demand for better depth of knowledge and more industry-specific information. Tbe byperknowledge environment with quick access to data, and the possibility to browse back and forth between various databases pmved to be timesaving and to promote the benchmarking pwe wrnted to get started. It may be a point of interest to note that the Mintzberg description of the limitationa of ‘bati information” in strategy making simply is incorrect for this type of environment.

Investments . . .

maintenance investments should he taken as default, as they have to be made anyway; there should be a possibility to run simulations on strategic investments; there should be a possibility to see a summary of ail the investments per year;

The investments module, and its interlinked financing module, was technically more advanced than the other modules; it seems that investment analysis was intuitively understood by only a select few, but that Woouktrat made the logic quite clean for all SBU managers.

Memo l

Production Position .

l

.

advantages: no additional work is needed in order to get the productivity; ail cost factors are linked from existing data in the CEO Report, generally very difficult to measure; labour and energy are important factors, raw material is standard; there should be a possibility for benchmarking; the figures are more or less abstract;

l

l

The Memo module turned out to be an exciting tool for us, the system developers, as it displayed ail the insights the users gained in their discussions of the various aiternatives. It also showed their conclusions on various items and their evaluation of their competitors. A number of factors were not known too well and the Memo was used to register those points were follow-up studies of the competitors were required; a number of questions and ideas were forwarded to sales offhs in Europe for verification and follow up studies. The Memo revealed that we were following Q strategy formation process (as described in section 3), which was carried out by the management teams of the nine SBUs. Now with these insights as a basis it is interesting to foilow-up on Mintzberg’s discussion of strategic planning:

At some point we almost believed that the determination ofproductivity is a form of black magic, as nobody seems to be able to come up with a reasonable method. Finally, we built and tried out our own version of the productivity measures, and they seemed to work. Productivity is an abstract and complex measure, and the idea was to gain some insight into the effects of the MP and CP visions of changes in productivity. There are knowledge based links to simplify the calculations, and (as we found in the Memos) the users turned out to gain an intuitive understanding of the relations between market developments and the productivity. External Data l

.

.

I used the memo as a report-generator, I would like to have two memos; one for my internal continuous use and the other as a formal presentation and as a check -list; I really used the memo; 15 written pages, it took a lot of time; I used it always when there was a change in either market or competitive positions; if you do not use the memo, later you do not have a chance to remember the reasoning behind your decisions;

I reaiised that the SBUs have not used the external data as tnuch as tbey should have; some of the SBUs are very optimistic about their competitive position, which does not seem merited with the market information available; we did not use the external data very much, there is still not too much specific information available about the pulp business; we would like to have the possibility to do our own forecasts in the external database;

(i)

(ii)

(iii)

The External data module was built to be easily acc..ssible and to give a benchmarking basis for estimates of the

the assumption offormalisution appears to be carrect, i.e. the strategy making process can he programmed by the use of a system; the assumptions of detachment appears to be incorrect; the strategic planning was done by the SBU management teams; the assumption of quuntifrcation seems to be correct, i.e. the management teams favoured bard data and systematic aggregates as far as possible; knowiedge was not generaiised to the extent described by Mintzherg; this was mainly due to the technology we had implemented in the system, which allowed

80

Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95) 1060-3425/95 $10.00 © 1995 IEEE

Proceedingsof the 28th Annual Hawaii International Conferenceon SystemSciences- 1995

(iv)

the users to work with knowledge on many levels of aggregationsimwltaneousiy; the assumption of predetermination is not necessary, nor is it relevant; an interactive support system, with routines for continuous updates of all central knowledge elements, allows the user to quickly adapt to and find out useful and relevant responses to a quickly changing and unpredictable

l

l

environment.

If we follow the criteria establishedby Popper [17], i.e. one false premise nulEifiesthe conclusion, we have shown that Mint&erg is wrong about the “grand fallacy”.

5. Conclusions

l

l

time as I now have a fairly complete basis as a starting point; COO of a pulp mill although I bave for several years done strategic plen-

ning “my way”, after having used the Wood&at system I would not change back for anything; COO of a building materials unit the final version which we presentedto the corporate executive board took us 1,5 day to finish (pmparatory work done) ; our group - six people - worked as a team and found the system very useful for teamwork; COO of a pulp mill the sydtemguides tbe user to focus on important issues which eliminates unnecessarywork; COO of a saw mill the support system “imprisoned us; the drawback was that we concentratedtoo much on details in the MP and CP; COO of a saw mill we bad our basic visions and missions in mind, and found out through Woodstrut that our visions of cbiorine free pulp changed; there exists - after all - a demand big enough for the new product we bad discussed;COO of a pulp mill the planning processwas real team work; COO of a building materials unit

Woodstrut is a hybrid of an object-orientedexpert system and a byperknowledge support system, and is thus constructed to provide both advanced level knowledge support in strategic management and an environment for linking assessmentsof qualitative factors with systematic quantitative evaluationsof their consequences. It turned out that the technology used in Woodstrut is instrumental for continuous analysis and synthesis processes,which can be made to interact both in a logical / causaland a temporal setting. As a by-product this actually shows that Mintzberg’s main argument for estabiisbing “the grand fallacy” of strategic planning - because

Thus strategy formation seemsto be a processin which analysis becomes synthesis through the intervention of Woodstrat - a knowledge-basedsupport system.

analysis is not synthesis, strategic planning is not strategy formation - cannot be correct. Woudsfrut is a support system for strategy formation.

References

Tbe links between the logical elements of the system foilow an intuitive, internal logic which has gradually emerged tbrougb interactive work with the SBUmanagers. This has created the foundations for a quick and effective user acceptance. Woodmat appearsto have a number of useful featuresas reportedby the users: I am very pleasedwith the system, it really is a working system; although I also have (paper) documentsof the SBU plans, I work only with the Woodstrat system; I have got very positive feedback from the CEO on the computer-basedpresentation of the division’s and the SBUs’strategic plans; perhapsI ought to allow the SBUs to carry out group and division consoiidations - this I believe would generateinternal competition to attain divisional goals and objectives; senior VP Woodstrat compared to “my old way”? - I worked more thoroughly and I used more time than normally; next year the work will be accomplishedin much less

l

l

1. Day, George, Weik, Bartonand Wensley, Robin, (ed.) The Interface of Marketing and Strategy, Jay Press Inc., Greenwich (1990) 2. Ansoff, H. Igor, Critique of Henry Mint&erg’s The Design School: Reconsideringthe Basic Premisesof Strategic Management’,Strategic Management Journal, Vol 12 (1991), 449-461 3. Minkberg, Henry, The Rise and Fall of Strategic Planning, Prentice Hall (1994) 4. Mintzberg, Henry, Patterns in Strategy Formation, ManagementScience,Vol 24, No. 9, (1978) 934-948 5. Carisson, Christer, Approximate Reasoning Through Fuzzy MCDM-Methods, Operational Research ‘87, Elsevier, Amsterdam (1988), 817-828 6. Gougen, M., On Fuzzy Robot Planning, Memo No 1 on AI, UCLA, Los Angeles (1974) 7. Rockart, John F. and DeLong, David W., Executive Supprl Systems,Dow Jones-Irwin, Homewood (1988) 8. Walden, Pirkko and Carlsson, Christer, Enhancing Strategic Market Management with Knowledge-Based Systems, HICSS-26 Proceedings, IEEE Computer Society Press, Los Alamitos, 240-248

Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95) 1060-3425/95 $10.00 © 1995 IEEE

Proceedingsof the 28th Annual Hawaii International Conferenceon SystemSciences- 1995 13. Chang, Ai-Mei, Holsapple, Clyde W. and Whinston, Andrew B., Model Management lssuea and Directiona, lki-

9. Carlssoa, Christer, Expert Systems as Conceptual Frame-

works and Management Support Systems for Strategic Management, Internutional Journal of Information Resource Management, Vol2, No. 4 (1991), 14-24 10. Carlsson, Christer, New Instruments for Management Research,Hwnan Systems Management, Vol 10, No. 3 (1991),

sion Sumrt

Systems, Vol9 (1993), 19-37

14. Lange, Danny B., Object-Grieated Hypennodeling of Hypertext Supported Information Sptemq HICZZ&M Proceedings, IEEE Computer Society Press,Loa AJamitoa, 380-389 15. Cershman, Anatole and Got&man, Bdward, Use of Hypermedia for Corporate ffiowledge Dissemination, HICSS-26 Proceedings, IEEE Computer Society Press, Lee Alemitos, 41 l-420 16. Eden, Colin, Strategy Development and ImplementationCognitive Mapping for Group Support, in Hendry and Johnson (eds): Strategic Thinking: Leadership aad the hknagement of Chauge, (1993) 17. Popper, Karl, The Iogic of Scientific Discovery, Harper & Row, New York (1959).

203-220

11. Carlsson, Christer Knowledge Formation in Strategic Management, IfXCSS-27 Proceedings, IEEE Computer Society Press,Los Alamitos, 221-230 12. Walden, Pirkko and Catlsson, Christer, Strategic Management with a Hyperknowledge Support System, HiCSY-27 Proceedings, IEEE Computer Society Press, Los Alamitos, 241-250

82

Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95) 1060-3425/95 $10.00 © 1995 IEEE