www.elsevier.com/locate/eswa

Expert Systems with Applications 32 (2007) 687–702

Expert Systemswith Applications

Intelligent modeling of e-business maturity

George Xirogiannis a,*, Michael Glykas b

a University of Piraeus, Department of Informatics, 80, Karaoli & Dimitriou St., 185 34 Piraeus, Athens, Greeceb University of Aegean, Department of Financial and Management Engineering, 31, Fostini Street, 82 100 Chios, Greece

Abstract

E-business has a significant impact on managers and academics. Despite the rhetoric surrounding e-business strategy formulationmechanisms, which support reasoning of the effect of strategic change activities to the maturity of the e-business models, are still emerg-ing. This paper describes an attempt to build and operate such a reasoning mechanism as a novel supplement to e-business strategy for-mulation exercises. This new approach proposes the utilization of the fuzzy causal characteristics of Fuzzy Cognitive Maps (FCMs) asthe underlying methodology in order to generate a hierarchical and dynamic network of interconnected maturity indicators. By usingFCMs, this research aims at simulating complex strategic models with imprecise relationships while quantifying the impact of strategicchanges to the overall e-business efficiency. This research establishes generic adaptive domains – maps in order to implement the inte-gration of hierarchical FCMs into e-business strategy formulation activities. Finally, this paper discusses experiments with the proposedmechanism and comments on its usability.� 2006 Elsevier Ltd. All rights reserved.

Keywords: Fuzzy cognitive maps; E-business modeling; Strategy planning; Decision support

1. Introduction

Today, there is an increasing demand for a strategic-level assessment of e-business capabilities that can beassembled and analyzed rapidly at low cost and withoutsignificant intrusion into the subject enterprises. The bene-fits from completing such an exercise are quite straightfor-ward, for instance, identification of significant strengthsand weaknesses, establishment of a rationale for action, areference point for measuring future progress, etc.

This paper proposes a novel supplement to strategic-level maturity assessment methodologies based on fuzzycognitive maps (FCMs). This decision aid mechanism pro-poses a new approach to supplement the current statusanalysis and objectives composition phases of typical e-business strategy formulation projects, by supporting‘‘intelligent’’ modeling of e-business maturity and ‘‘intelli-gent’’ reasoning of the anticipated impact of e-business

0957-4174/$ - see front matter � 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.eswa.2006.01.042

* Corresponding author.E-mail addresses: georgex@unipi.gr (G. Xirogiannis), mglikas@

aegean.gr (M. Glykas).

strategic change initiatives. The proposed mechanism uti-lizes the fuzzy causal characteristics of FCMs as a newmodeling technique to develop a causal representation ofdynamic e-business maturity domains. This research pro-poses a holistic set of adaptive domains in order to generatea hierarchical network of interconnected e-business matu-rity indicators. The proposed mechanism aims at simulat-ing the operational efficiency of complex hierarchicalstrategy models with imprecise relationships while quan-tifying the impact of strategic alignment to the overalle-business efficiency. Also, this paper proposes an updatedFCM algorithm to model effectively the hierarchical anddistributed nature of e-business maturity.

This application of FCMs in modeling the maturity ofe-business is considered to be novel. Moreover, it is thebelief of this paper that the fuzzy reasoning capabilitiesenhance considerably the usefulness of the proposed mech-anism while reducing the effort to identify precise maturitymeasurements. The proposed model has both theoreticaland practical benefits. Given the demand for effectivestrategic positioning of e-business initiatives, such a suc-cinct mechanism of conveying the essential dynamics of

688 G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702

e-business fundamental principles is believed to be usefulfor anyone contemplating or undertaking an e-businessstrategy formulation exercise. Primarily, the proposedmodel targets the principle beneficiaries and stakeholdersof strategy formulation projects (enterprise top administra-tion, strategic decision makers, internal auditors, etc)assisting them to reason effectively about the status ofe-business maturity metrics, given the (actual or hypothet-ical) implementation of a set of strategic changes. Never-theless, the explanatory nature of the mechanism canprove to be useful in a wider educational setting.

This paper consists of five sections. Section 2 presents ashort literature overview, Section 3 presents an overview ofthe FCM based system, while Section 4 discusses the newapproach to e-business maturity modeling based on FCMs.Finally, Section 5 concludes this paper and briefly discusesfuture research activities.

2. Literature overview

2.1. E-business drivers

E-business offers promise to apply web and other elec-tronic channel technologies to enable fully the integrationof end-to-end processes. It involves both core and supportbusiness aspects, it focuses on information sharing effi-ciency, not just financial transactions. E-business primaryobjective is business improvement through:

• Deployment of new technologies in the value chain.• Connection of the value chains between enterprises

(B2B) and between enterprises and consumers (B2C)in order to improve service, exploit alternative dis-tribution/communication channels and support costreduction due to the associated value chain optimi-zation.

• Increase of the speed of information processing (mainlyat real-time) and responsiveness by utilizing commoninformation sources (both external and internal).

E-business has a significant impact on every businessfunction. Integrated information technology causes a shiftin the value chain of the enterprise. It causes a considerabledeflation of prices due to radical cost reductions, annihila-tion of profit margins, disintermediation of companies andindustries due to the transparent product/service deliveryto the end customer, increase in cross selling volumes andso forth. On the other hand, no industry is immune tointense competition due to chain reactions that affect allelectronic network partners (Palmer, 2002). This may causea higher level of uncertainty of future business prospects,but it is only fair to say that adaptive risk managementmay reduce such pitfalls. Also, the current enterprise valu-ation can be radical altered by this new business environ-ment therefore enterprises must reconsider their corecompetencies and strategies to maintain their competitiveadvantages.

The new economy associated with e-business has brokendown many of the traditional barriers. The fundamentalshift in focus from optimizing the efficiency of individualenterprises to optimizing the efficiency of a network ofenterprises for competitive advantage is a considerablechallenge (Chung, Yam, Chan, & Potter, 2005). E-businessactivities now operate across an extended network of digi-tally connected partners to enable demand/capacity/priceoptimizations while offering self-service client relationshipsat multiple channels with a significant communicationspeed.

It is the view of this paper that while e-business solutionproviders promise financial prosperity and sales volumes,case studies clearly indicate that awareness, targeted strate-gic planning and holistic organizational alignment are thekey success factors for managing business in the digitalage. Understanding the speed and scope of e-businessimpacts while generating the essential momentum formsthe basis for setting realistic strategic priorities, mappingout a go-forward plan while evaluating the critical factorsfor e-business success. Effective service/product deliverythrough electronic channels requires efficient process con-trol and management of measurable targets, in order tomaintain the necessary range of organizational buy-in, tomanage risk and assure accountability.

2.2. Relevant research in business modeling

2.2.1. Modeling traditional business activities

Enterprises usually employ modeling techniques to drivere-design activities and communicate the impact of internalchange. Such modeling techniques may loosely fit withinthe area of decision support systems (Carlsson & Turban,2002; Shim et al., 2002; Sprague & Watson, 1986). Severalmodeling approaches can be brought to bear on the task ofsupplementing business modeling activities. In particularthe field of knowledge-based systems (Harmon & King,1985; Metaxiotis, Psarras, & Samouilidis, 2003) could ful-fill the desire for more accurate predictive business model-ing tools. The research presented by Lin, Yang, and Pai(2002) proposed generic structures with no formal reason-ing capabilities to model traditional business processes,which could represent a business process in various con-cerns and multiple layers of abstraction.

The research presented by Burgess (1998) modeled busi-ness process models with system dynamics to support thefeasibility stage of business process re-engineering (BPR).Similarly, research (Burgess, 1998) modeled the interactionbetween competitive capabilities of quality and cost duringtotal quality management (TQM) initiatives (Burgess,1996). This model did not decompose hierarchical relation-ships, nor did it allow the connection of the sub-models.Finally, the model required formal definition of causal rela-tionships (e.g. functions), which posed a significant over-head in supplementing the business modeling exercise.The research presented by Crowe, Fong, Bauman, andZayas-Castro (2002) reported the development of a tool

G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702 689

to quantitatively estimate the potential risk level of a pro-cess change effort based on simple arithmetic approxima-tions. The utilization of uncertainty was also suggestedby Jones and Ryan (2002) which proposed a contingencymodel of quality management practices, whereby qualitymanagement orientation, process choice, and environmen-tal uncertainty were the contextualizing variables.

Other research attempts built on theories like non-line-arity. The research presented by Murray, Priesmeyer,Sharp, Jensen, and Jensen (2000) revealed that nonlinearscience offered a practical new frame of reference for busi-ness modeling initiatives (e.g. health care settings).

Research (Kwahk & Kim, 1999) offered a two-phasecognitive modeling (called TCM), to help enterprises iden-tify potential organizational conflicts. It proposed a num-ber of informal/ambiguous techniques to generate andvalidate the organizational cognitive maps, like interviews,observation, group discussions, questionnaires, documentanalysis, and so forth. Causal values were generatedaccording to the pairwise comparison technique with nofuzzy definitions allowed.

Each of the aforementioned tools and techniques offersdistinct advantages in modeling business architectures.However, all of them focus in modeling traditional businessactivities and offer limited functionality in modeling pro-cess – technology integration. They tend to visualize theenterprise as an isolated entity, while e-business practicesbuild on horizontal interconnections between networks ofcoupling value chains. Therefore, it is only fair to say thatsince most contemporary e-business principles depart fromtraditional business practices, contemporary e-businessmodeling tools should also build on contemporary model-ing approaches.

2.2.2. Modeling contemporary e-business activities

While information and communication technology(ICT) in the form of e-business is advocated as an enablerby allowing to be shared by all business stakeholders in thevalue chain, there is little analytical or quantifiable evi-dence that it will actually improve the overall performanceof the enterprise in delivering customer wants. It is usuallyproposed that passing information to all entities in thevalue chain may improve performance, but still no formalreasoning evidence has been provided to support this argu-ment. The impact of the e-business enabled value chain onstrategic decisions, materials/component suppliers, distri-bution channel operations, etc., however, is less well under-stood and exploited. For established enterprises, change isthe key challenge, as argued by Jackson and Harris (2003)and Phan (2003). Such enterprises must rethink fundamen-tal aspects of their strategy, which may lead to a radicaloverhaul of existing ways of doing business, with companystructure and culture becoming much more customer-focused.

Research (Hooft & Stegwee, 2001) discussed a methodfor the development of an e-business strategic framework.However it focused on qualitative analysis based on the

SWOT framework without identifying any causal relation-ships among value chain drivers.

The experimental research presented in Bharati andChaudhury (2003) endeavored to understand factors thataffect decision-making satisfaction in web-based decisionsupport systems. Using a structural equation modelingapproach, the analysis revealed that information qualityand system quality influenced e-business decision-making.While the underlying model built on structured relation-ships, no formal automated reasoning was present.Research (Long & Schoenberg, 2002) presented similarempirical analyses to discuss whether e-business requiresdifferent leadership characteristics.

Research (Disney, Naim, & Potter, 2005) investigatedhow e-business affects the supply chain dynamics of anenterprise in an attempt to establish e-business enabled sup-ply chain models for quantifying the impact of ICTs. It con-cluded that only robust models could enable considerablequantitative insights into the impact of e-business on supplychain dynamic behavior prior to their implementation.

Research (Koh & Kim, 2005) modeled a virtual commu-nity activity framework, integrating community knowledgesharing into business activities in the form of an e-businessmodel. This proposition attempted to model business activ-ities relationships by limiting itself to statistical analysis ofraw electronic interactions, thus presenting limited researchportability to other business cases.

Research (Duffy, 2001) attempted to formalize a blue-print of maturity modeling. This model utilized maturitylevel indicators for each key success driver (KSD) categoryto estimate the overall e-business maturity of the enter-prise. Despite the fact that KSDs were well defined, matu-rity indicators were loosely related to the holistic businessperformance indicators, which could approximate the tra-ditional performance measurement exercise of the enter-prise. Moreover, there was no concrete mechanism thatcould implement the proposed underlying construct, whichquestioned its practical added value.

Software agents (autonomous or semi autonomous)capable of modeling routine, tedious, and recurrent time-consuming e-business activities were proposed by Albrecht,Dean, and Hansen (2003). The implementation of this rea-soning aid used situation calculus as the underlying meth-odology. However, it is fair to say that the proposed toolcould impose significant startup and initialization over-heads. Also, it focused on agents utilizing large amountsof pre-existing concrete knowledge. This prerequisite couldcompromise the precision of the results in the case ofimprecise or incomplete knowledge availability.

Finally, Mahajan and Venkatesh (2000) presented acomprehensive analysis of several contemporary marketingmodeling techniques for e-business. Most of the techniquesdiscussed followed a statistical/stochastic approach to esti-mate the impact of e-business initiatives to the overall busi-ness performance with limited (if any at all) ‘‘intelligent’’reasoning capabilities and limited identification of causalrelationships among e-business performance concepts.

e-customer satisfaction

internal cost

W41

productprice

W13

W12

W23

W52

W35

W34

W63

W46

W56

product defects

company profitability

e-sales volumes

Fig. 1. Simple FCM.

690 G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702

To summarize, it is the belief of this paper that there isno other tool that integrates FCM simulation into e-busi-ness maturity change exercises with the functionality andcharacteristics of the proposed mechanism (presentationwill follow in Section 3). TCM for example may supple-ment business modeling by drawing FCMs with allowednode values of 0 and 1 and no dynamic simulation capabil-ities. Frameworks like MIND, SODA, and COCOMAP(all compared in Kwahk & Kim (1999)) provided method-ologies and guidance that allowed the user to performFCM analysis by identifying node conflicts in multiplemaps, loops, cycles etc. However, nodes in different mapscould not be linked dynamically to create map hierarchies.Also, e-business tools either tend to be case specific (e.g.Albrecht et al., 2003; Disney et al., 2005; Koh & Kim,2005, etc.), or offer limited practical value (e.g. Duffy,2001; Hooft & Stegwee, 2001; Mahajan & Venkatesh,2000).

Accurate predictive models may already exist in e-busi-ness consultancies. Through their experiences they arelikely to have built up databases that could underpin moredetailed approaches such as case based reasoning. Unfortu-nately, the existence and internal features of these modelsare more likely to remain confidential, given their commer-cial sensitivity.

2.3. FCMs as a modeling technique

Fuzzy Cognitive Maps (Kosko, 1986) is a modelingmethodology for complex decision systems, which origi-nated from the combination of Fuzzy Logic (Zadeh,1965) and Neural Networks. An FCM describes the behav-ior of a system in terms of concepts; each concept repre-sents an entity, a state, a variable, or a characteristic ofthe system (Dickerson & Kosko, 1997).

FCM nodes are named by concepts forming the set ofconcepts C = {C1, C2, . . . ,Cn}. Arcs (Cj, Ci) are orientedand represent causal links between concepts; that is howconcept Cj causes concept Ci. Arcs are elements of the setA = {(Cj, Ci)ji} [ C · C. Weights of arcs are associatedwith a weight value matrix wn·n, where each element ofthe matrix wji 2 [�1, . . . , 1] [ R such that if (Cj, Ci) 62 A

then wji = 0 else excitation (respectively inhibition) causallink from concept Cj to concept Ci gives wji > 0 (respec-tively wji < 0). The proposed methodology frameworkassumes that [�1, . . . , 1] is a fuzzy bipolar interval, bipolar-ity being used to represent a positive or negativerelationship.

In practice, the graphical illustration of an FCM is asigned graph with feedback, consisting of nodes andweighted interconnections (e.g. !weight

). Signed and weightedarcs (elements of the set A) connect various nodes (ele-ments of the set C) representing the causal relationshipsthat exist among concepts. This graphical representation(e.g. Fig. 1) illustrates different aspects in the behavior ofthe system, showing its dynamics (Kosko, 1986) and allow-ing systematic causal propagation (e.g. forward and back-

ward chaining). Positive or negative sign and fuzzy weightsmodel the expert knowledge of the causal relationships(Kosko, 1991). Concept Cj causally increases Ci if theweight value wji > 0 and causally decreases Ci if wji < 0.When wji = 0, concept Cj has no causal effect on Ci. Thesign of wji indicates whether the relationship between con-

cepts is positive ðCj!W ji

CiÞ or negative ðCj!W ji � CiÞ, while

the value of wji indicates how strongly concept Cj influencesconcept Ci. The forward or backward direction of causalityindicates whether concept Cj causes concept Ci or viceversa.

Simple variations of FCMs mostly used in business deci-sion-making applications may take trivalent weight values[�1, 0, 1]. This paper allows FMCs to utilize fuzzy wordweights like strong, medium, or weak, each of these wordsbeing a fuzzy set to provide complicated FCMs. In con-trast, Kwahk and Kim (1999) adopted only a simplerelative weight representation in the interval [�1, . . . , 1].To this extend, Kwahk and Kim (1999) offered reducedfunctionality since it does not allow fuzzy weight defi-nitions.

Generally speaking FCM concept activations take theirvalue in an activation value set V = {0, 1} or {�1, 0, 1} ifin crisp mode or [�d, 1] with d = 0 or 1 if in fuzzy mode.The proposed methodology framework assumes fuzzymode with d = 1. At step t 2 N, each concept Cj is associ-ated with an inner activation value at

j 2 V , and an externalactivation value et

aj2 R. FCM is a dynamic system. Initial-

ization is a0j ¼ 0. The dynamic obeys a general recurrent

relation atþ1 ¼ f ðgðeta;w

TatÞÞ; 8t P 0, involving weightmatrix product with inner activation, fuzzy logical opera-tors (g) between this result and external forced activationand finally normalization (f). However, this paper assumesno external activation (hence no fuzzy logical operators),resulting to the following typical formula for calculatingthe values of concepts of FCM:

atþ1i ¼ f

Xn

j¼1;j 6¼i

wjiatj

!ð1Þ

where atþ1i is the value of concept Ci at step t + 1, at

j thevalue of the interconnected concept Cj at step t, wji is theweighted arc from Cj to Ci and f : R! V is a threshold

G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702 691

function, which normalizes activations. Two thresholdfunctions are usually used. The unipolar sigmoid functionwhere k > 0 determines the steepness of the continuousfunction f ðxÞ ¼ 1

1þe�kx. When concepts can be negative(d < 0), function f(x) = tanh(x) is used.

To understand better the analogy between the sign ofthe weight and the positive/negative relationship, it maybe necessary to revisit the characteristics of fuzzy relation(Kaufmann, 1975; Lee, Kim, Chung, & Kwon, 2002). Afuzzy relation from a set A to a set B or (A, B) representsits degree of membership in the unit interval [0, 1]. Gener-ally speaking, sets A and B can be fuzzy sets. The corre-sponding fuzzy membership function is lf : A · B![0, 1]. Therefore, lf(x, y) is interpreted as the ‘‘strength’’of the fuzzy membership of the fuzzy relation (x, y) wherex 2 A and y 2 B. Then this fuzzy relation concept can bedenoted equivalently as x!lf

y and applied to interpret thecausality value of FCM, since wji (the causality value ofthe arc from nodes Cj to Ci) in a certain FCM is interpretedas the degree of fuzzy relationship between two nodes Cj

and Ci. Hence, wji in FCMs is the fuzzy membership valuelf(Cj, Ci) and can be denoted as Cwj;i

j ! Ci.However, we understand that the fuzzy relation (weight)

between concept nodes is more general than the originalfuzzy relation concept. This is because it can include nega-tive (�) fuzzy relations. Fuzzy relations mean fuzzy causal-ity; causality can have a negative sign. In FCMs, thenegative fuzzy relation (or causality) between two conceptsis the degree of a relation with a ‘‘negation’’ of a conceptnode. For example, if the negation of a concept node Ci

is noted as �Ci, then lf(Cj, Ci) = �0.6 means thatlf(Cj, �Ci) = 0.6. Conversely, lf(Cj, Ci) = 0.6 means thatlf(Cj, �Ci) = �0.6.

2.4. Applications of fuzzy cognitive maps

Over the last 10 years, a variety of FCMs have been usedfor capturing – representing knowledge and intelligentinformation in engineering applications, for instance, geo-graphical information systems (Liu & Satur, 1999) andfault detection (Ndouse & Okuda, 1996; Pelaez & Bowles,1995). FCMs have been used in modeling the supervisionof distributed systems (Stylios, Georgopoulos, & Groum-pos, 1997). FCMs have also been used in operationsresearch (Craiger, Goodman, Weiss, & Butler, 1996),web data mining (Hong & Han, 2002; Lee et al., 2002),as a back end to computer-based models and medicaldiagnosis (e.g. Georgopoulos, Malandraki, & Stylios,2002).

Several research reports applying basic concepts ofFCMs have also been presented in the field of business(e.g. Xirogiannis & Glykas, 2004a, 2004b) and other socialsciences. Research in Axelrod (1976) and Perusich, 1996have used FCM for representing tacit knowledge in politi-cal and social analysis. FCMs have been successfullyapplied to various fields such as decision making in com-plex war games (Klein & Cooper, 1982), strategic planning

(Diffenbach, 1982; Ramaprasad & Poon, 1985), informa-tion retrieval (Johnson & Briggs, 1994) and distributeddecision process modeling (Zhang, Wang, & King, 1994).Research like (Lee & Kim, 1997) has successfully appliedFCMs to infer rich implications from stock market analysisresults. Research like (Lee & Kim, 1998) also suggested anew concept of fuzzy causal relations found in FCMsand applied it to analyze and predict stock market trends.The inference power of FCMs has also been adopted toanalyze the competition between two companies, whichare assumed to use differential games mechanisms to setup their own strategic planning (Lee & Kwon, 1998).FCMs have been integrated with case-based reasoningtechnique to build organizational memory in the field ofknowledge management (Noh, Lee Lee, Kim, Lee, &Kim, 2000). Recent research adopted FCMs to supportthe core activities of highly technical functions likeurban design (Xirogiannis, Stefanou, & Glykas, 2004).Summarizing, FCMs contribute to the construction ofmore intelligent systems, since the more intelligent a systembecomes, the more symbolic and fuzzy representations itutilizes.

In addition, a few modifications have been proposed.For example, the research in Silva (1995) proposed newforms of combined matrices for FCMs, the research inHagiwara (1992) extended FCMs by permitting non-linearand time delay on the arcs, the research in Schneider,Schnaider, Kandel, and Chew (1995) presented a methodfor automatically constructing FCMs. More recently,Liu and Satur (1999) has carried extensive research onFCMs investigating inference properties of FCMs, pro-posed contextual FCMs based on the object-oriented para-digm of decision support and applied contextual FCMs togeographical information systems (Liu, 2000).

3. Maturity modeling using FCMs

3.1. Overview of maturity modeling

Despite the rhetoric surrounding technology integrationintelligent mechanisms that support (a) holistic assessmentof e-business maturity (b) proactive identification of theassociated risks and opportunities, (c) reasoning on theimpact of the strategic convergence of processes and tech-nologies to the maturity of the business model are stillemerging. Furthermore, contemporary business modelingtechniques focus on an ex-post performance assessmentof traditional operations. To this extend, the new proposi-tion of this paper is two-fold:

• E-business gap assessment:– The proposed methodology tool aims at providing an

interdisciplinary framework to benchmark currente-business characteristics (if any) by defining maturitymetrics in reference to a wide variety of objectivecharacteristics. The tool proposes a two dimen-sional, though practical, perspective: customer facing

692 G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702

activities and supply chain activities. In contrast,other research proposals as presented in Section2.2.2 (for example Albrecht et al., 2003; Disneyet al., 2005; Koh & Kim, 2005, etc.) follow a singledimensional approach.

– The proposed methodology framework offers a newsource of tangible strategic requirements for e-busi-ness based on a multidimensional analysis and aholistic viewpoint of the enterprise. Also, it aims atencompassing both traditional and contemporaryinfrastructure solution options (essentially a combi-nation of the characteristics presented in Sections2.2.1 and 2.2.2), without being either product or solu-tion centric.

• E-business model evolution and impact assessment:– The proposed tool can be used to quantify the impact

of e-business maturity changes to the efficiency of theenterprise. This tool should be perceived as a decisionaid to support strategy decisions at an executive level,rather than a sophisticated process simulator.

The proposed methodology tool offers a holisticapproach to understanding e-business challenges througha broad coverage of business areas.

3.2. FCM as a supplement to strategic change projects

A typical e-business strategy formulation methodol-ogy (Fig. 2) consists of a series of phases and layers ofanalysis for setting the strategy roadmap of an enter-prise:

Phase 1: Current status analysis/best practices bench-marking.

Phase 2: Vision and strategic positioning.

Vision & strategic positioning

Objective 1 Objective 2

Action plan 1 Action plan 2

…

…

Synthesis & diagnosis

Critical success factors

Best practices

Fig. 2. Overview of busines

Phase 3: Objectives composition, including critical suc-cess factors analysis and selection of strategicalignment indicators (maturity indicators).

Phase 4: Strategic change planning (action planning).

The proposed mechanism focuses on supplementing atypical e-business strategy methodology by providing aholistic strategic alignment evaluation framework basedon e-business maturity indicators. In practice the mecha-nism supplements the recurring feedback loop betweenthe current maturity status, the future strategic objectivesand the action plans for improving e-business maturity(the action plan, in turn, affects the future maturity status).The proposed mechanism actually generates two maturityassessment flows, as explained in Section 3.1:

• ‘‘Current status analysis! Objectives’’ to estimate thegap between existing (‘‘as-is’’) and future (‘‘to-be’’)e-business maturity and establish objectives whichshould bridge this gap.

• ‘‘Action plans! Objectives’’ to estimate the evolutionof e-business, assess its maturity and align objectivesto meet any deviations.

During the third phase of this typical strategy formula-tion exercise, the top management of the enterprise sets theoverall performance targets (strategic maturity). These tar-gets are exemplified further to action plan performancemetrics (tactical maturity) and then to operational perfor-mance indicators (operational maturity). All such metricspresent inherent relationships. In practice, strategic matu-rity metrics must cascade to tactical maturity metrics toallow the middle management to comprehend inherentrelations among the different managerial levels of the enter-prise. Similarly, tactical maturity metrics must propagate

Objective k

Action plan n

Current statusanalysis

Strateg

ic align

men

t perfo

rman

ce evaluatio

n

s strategy formulation.

G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702 693

up the overall strategic maturity metrics. However,relationships between maturity metrics at the samemanagerial level or even relationships between metrics ofdifferent managerial levels with no apparent relation-ships are not always clear and well defined. Thus reaso-ning of the chained impact of maturity metrics to theefficiency of the overall e-business model is not alwaysfeasible.

To resolve this issue, this paper proposes the utilizationof maturity indicators (Fig. 3) to develop the FCMs andreason about the impact of strategic changes to the desired(‘‘to-be’’) e-business models. The proposed mechanism uti-lizes FCMs to interpret:

• e-business maturity metrics as concepts (graphically rep-resented as nodes),

• decision weights as relationship weights (graphicallyrepresented as arrowhead lines),

• decision variables as maturity concept values,• hierarchical decomposition (top-down decomposition)

of maturity metrics to maturity indicators and constitu-ent sub-metrics as a hierarchy of FCMs. This interpreta-tion allows the stakeholders to reason about lower levelFCMs first (constituent indicators) before they reasonabout higher-level e-business maturity metrics (affectedmetrics).

B u s i nes s s t rategy formulation hierarchy

Vision & strategic positioning

Objective 1 Objective 2 Objective k

Action plan 1 Action plan 2 Action plan n

…

…

Fig. 3. Inherent relationships between e-b

Domains

MIs

hie

Strategy

MII MII…

Orgacom

MI MII

Fig. 4. Hierarchical definition of e-business

The proposed mechanism supports reasoning about theoverall or partial e-business strategy implementation usingmaturity indicators from the e-business philosophy. In con-trast to Kwahk and Kim (1999), the proposed mechanismbuilds on hierarchical metrics interrelationships identifiedand utilized by the e-business strategy formulation method-ology. The proposed approach does not perform or guidethe implementation of any stage of the strategy formula-tion methodology. Also, the approach does not performor guide the estimation of the absolute value of any ofthe maturity metrics and/or the overall e-business perfor-mance. It only allows the stakeholders to reason aboutthe qualitative state of e-business maturity metrics usingfuzzy linguistic variables like high-neutral-low cost, high-neutral-low impact of IT infrastructure to cost, etc.

3.3. Maturity domains

The proposed methodology tool (in contrast to othertechniques discussed in Section 2.2.2) builds a hierarchyof domains and indicators to model e-business maturity(Fig. 4). The tool’s strength is its ability to integrate keyareas (domains) of expertise available within the enterpriseessential to e-business operations. The proposed methodol-ogy tool proposes a holistic view to address e-businessmaturity in seven major business domains, namely:

C1

C1.1 C1.2 C1.3

FCMs hierarchies

maturity metrics

usiness strategy and FCM hierarchies.

FCM rarchy

……nizationalpetencies Tax & Legal

MII MII… MII MII…MII

maturity concepts and indicators (M/s).

694 G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702

• Domain 1 – Overall e-business strategy: It addresses crit-ical strategic issues, such as setting strategic direction,analyzing competitors, leveraging information technol-ogy, etc.

• Domain 2 – Organizational competencies: It addressesaspects such as whether an organization requires newskills, new competencies, new ways of working, etc.

• Domain 3 – Channel management: It focuses on the pri-mary e-business processes associated with distributionchannel management, marketing, distribution and logis-tics management, procurement, and customerinteraction.

• Domain 4 – Performance Management: It addresseshow an organization plans, measures, monitors, andcontrols the performance of its e-business capabilitiesand functions.

• Domain 5 – Tactical and Support Operations: It coversissues related to the day-to-day operations, namely con-tent creation, risk management, financial practices, etc.

• Domain 6 – Systems and Technology: It examinese-business enabling technologies for customer and sup-ply chain support and highlights integrated softwaresolutions and trends in technology. Also, it addressese-business security and privacy.

• Domain 7 – Tax and Legal: It addresses an organiza-tion’s e-business tax exposure and liabilities to ensurethat an organization knows its rights, obligations andpotential liabilities.

3.4. New FCM algorithm

As far as the underlying algorithm is concerned, thispaper extends the basic FCM algorithm (as discussed inSection 2.3 and also used by Kwahk & Kim (1999)), byproposing the following updated algorithm:

atþ1i ¼ f k1at

i þ k2 �Xn

j¼1;j 6¼i

wjiatj

!ð2Þ

Profitabi

Head Off

CRM coordination

Internet

e-promotion

Mobile ph

W1 W2 W1 W

CRMcoordination

Fig. 5. Horizontal matu

This paper assumes that coefficients k1 and k2 can be fuzzysets.

Coefficient k1 represents the proportion of the contribu-tion of the value of the concept ai at time t in the compu-tation of the value of ai, at time t + 1. In practice, this isequivalent to assume that wii = k1. The incorporation ofthis coefficient results in smoother variation of conceptvalues during the iterations of the FCM algorithm. Coeffi-cient k2 expresses the ‘‘influence’’ of the interconnectedconcepts in the configuration of the value of the concept ai

at time t + 1. It is the proposal of this paper that such acoefficient should be used to align indirectly causal rela-tionships (essentially, the value of concept Ci) with the cen-tralized/decentralized nature of maturity concept Cj aswell as with the significance of the hierarchical positioningof concept Cj within the strategic framework of theenterprise.

Intuitively, the introduction coefficient k2 imposes twostep of analysis for establishing the ‘‘influence’’ of causalrelationships:

Step 1: estimation of the direct influence of a maturity con-cept Cj to another concept Ci with the weight (wji)of the relationship. Both Ci and Cj should belongto the same maturity domain (or level), that isk2 = 1.

Step 2: approximation of the indirect importance of dupli-cate causal relationships spanning to differentmaturity domains (or levels) using coefficientk2 < 1.

Consider for example a typical enterprise (Fig. 5) withseveral electronic distribution channels (e.g. internet,mobile phones, video conferencing, etc) operating underthe same e-business strategic framework. Let ‘‘CRM coor-dination’’ and ‘‘profitability’’ be interrelated maturity con-cepts. From a theoretical standpoint weight w1 shouldappear to be the same for all duplicate ‘‘CRM coordina-tion–profitability’’ relationships across all channels. From

lity

ice

ones Video conferencing

2 W1 W2

e-promotion CRM coordination

e-promotion

rity decomposition.

G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702 695

a practical standpoint, channels may serve different numberof customers (or even customers with different transactionvolumes). In this example, coefficient k2 models the factthat CRM coordination in a channel which serves manycustomers (or even few customers with large transactionvolumes) is more important to the profitability of the enter-prise in comparison to the CRM coordination in channelswhich serve few customers (or even customers with verysmall transaction volumes), even if the level od CRM coor-dination is the same for all channels.

Similarly, Fig. 6 presents a generic strategy breakdownstructure, accompanied with a sample concept hierarchy.Regardless of weight values w1, w2, w3, coefficients1 = k2,(L1,L1) P k2,(L1,L2) P k2,(L2,L3) model the fact thataffecting maturity concepts at level Li (e.g. concept C5)are more important in determining the value of affectedmaturity concepts at level Li+1 (e.g. concept C3) in compar-ison to other affecting concepts at level Li�1 (e.g. conceptC6). k2,(Li,Lj) stands for the value of coefficient k2 associatedwith levels i and j.

C1

C2 C3

C5

Level 0

C6

W1

W2

W3Level 1

Level 2

Level 3

k2,(L1,L2)

k2 ,(L2,L3)

Fig. 6. Top-down matu

0 00.1

0.5

1

µ

-0.1-0.2-0.35-0.5-0.65-0.8-0.9-1

µµpvvlµnvvlµnvlµnlµnmµnhµnvhµnvvh µz

Fig. 7. Membership functions of

3.5. Assigning linguistic variables to FCM weights and

concepts

3.5.1. Expert linguistic variables

In order to define weight value of the association rela-tionships in an adaptive and dynamic manner, the follow-ing methodology is proposed. Managers are asked todescribe the interconnection influence of concepts using lin-guistic notions. Influence of one concept over another, isinterpreted as a linguistic variable in the interval [�1, 1].Its term set T(influence) is: T(influence) = {negativelyvery-very high, negatively very high, negatively high, nega-tively medium, negatively low, negatively very low, nega-tively very-very low, zero, positively very-very low,positively very low, positively low, positively medium, pos-itively high, positively very high, positively very-very high}.

This paper proposes a semantic rule M to be defined atthis point. The above-mentioned terms are characterized bythe fuzzy sets whose membership functions l are shown inFig. 7.

C4

k2≈ k2,(L1,L2 ) + k2 ,(L 2 ,L3)

rity decomposition.

.2 0.35 1 influence

µpm µps µpvs µpvvs

0.5 0.65 0.8 0.9

µplpvl

linguistic variable influence.

696 G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702

• M (zero) = the fuzzy set for ‘‘an influence close to 0’’with membership function lz.

• M(positively very-very low) = the fuzzy set for ‘‘aninfluence close to 10%’’ with membership functionlpvvl.

• M(positively very low) = the fuzzy set for ‘‘an influenceclose to 20%’’ with membership function lpvl.

• M(positively low) = the fuzzy set for ‘‘an influence closeto 35%’’ with membership function lpl.

• M(positively medium) = the fuzzy set for ‘‘an influenceclose to 50%’’ with membership function lpm.

• M(positively high) = the fuzzy set for ‘‘an influenceclose to 65%’’ with membership function lph.

• M(positively very high) = the fuzzy set for ‘‘an influenceclose to 80%’’ with membership function lpvh.

• M(positively very-very high) = the fuzzy set for ‘‘aninfluence close to 90%’’ with membership function lpvvh.

• Similarly for negative values.

The membership functions are not of the same size sinceit is desirable to have finer distinction between grades in thelower and higher end of the influence scale. The suggestedlinguistics are integrated using a sum combination methodand then the defuzzification method of center of gravity(CoG) is used to produce a weight in the interval [�1, 1].This approach has the advantage that experts do not haveto assign numerical causality weights but to describe thedegree of causality among concepts. The same semanticrule and term set can be used to define the coefficients k1

and k2. A similar methodology can be used to assign valuesto concepts.

Fig. 8. Map categories and a

4. FCM maturity implementation

4.1. FCM hierarchies

This research team uses the Quanta application tool, arobust visual implementation of FCMs. The implementa-tion of Quanta has been funded by the ESPRIT E.U. pro-gramme. The current implementation of the proposedmethodology tool encodes generic maps that can supple-ment the maturity modeling by storing concepts under dif-ferent map categories (Fig. 8a), namely:

• Business category: all concepts relating to core e-busi-ness activities.

• Social category: all tax and legal related concepts.• Technical category: all infrastructure related concepts

with emphasis on technology infrastructure.• Integrated category: essentially all top-most concepts

(e.g. a concept Ci with no backward causality such that"j: wji = 0), or concepts which may fall under more thanone main categories.

The dynamic nature of the approach allows easy recon-figuration. Further maturity concepts may be added, whilematurity concepts may be decomposed further to complywith specialized analysis requirements of enterprises. Thiscategorization is compatible both with the ‘‘process view’’or the ‘‘organizational view’’ (as adopted by Kwahk &Kim (1999)) of the enterprise to allow greater flexibilityin modeling dispersed knowledge flows. The hierarchicaldecomposition of maturity concepts generates a set of

sample FCM hierarchy.

G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702 697

dynamically interconnected hierarchical maturity maps.Each map analyzes further the relationships among con-cepts at the same hierarchical level. Fig. 8b presents sucha sample map hierarchy, which also serves as the FCMoverview.

Currently, the mechanism integrates more than 250 con-cepts, forming a hierarchy of 20 maps. The Quanta inter-face allows the user to utilize a sub-set of these conceptsand maps, on demand. The proposed system can portraythe maturity model following either a holistic or a scalableapproach. This is analogous to seeing the e-business strat-egy of the enterprise either as a single, ‘‘big bang’’ event oras an ongoing activity of setting successive objectives forselected operations. The proposed mechanism can accom-modate both approaches.

Also the current implementation allows easy customiza-tion of the function f and easy reconfiguration of the for-mula Atþ1

i to adapt to the specific characteristics ofindividual enterprises, generation of scenarios for the sameskeleton FCM, and automatic loop simulation until a user-defined equilibrium point has been reached. Alternatively,step-by-step simulation (with graphical output of partialresults) is also available to provide a justification for thepartial results.

The proposed framework exemplifies further e-businessmaturity by decomposing maturity domains into their con-sistent concepts. The following sections exhibit sample(though typical) skeleton maps, which provide relevanceand research interest to this paper.

4.1.1. E-business maturity

The mechanism proposes four (4) basic maturity mapseach consisting of generic e-business maturity metrics asfollows:

• The Strategy top-level map reasons on the maturity ofissues like strategic direction, e-business informationtechnology planning, competition analysis, etc.

Fig. 9. (a) Organizational competencies map

• The Organizational competencies top-level map (Fig. 9a)summarizes concepts like roles and responsibilities,e-business change management, alliance management, etc.

• The Performance management top-level map (Fig. 9b)interconnects concepts like CRM, performance manage-ment, supply chain management, etc.

• The Tactical and support operations top-level map rea-sons on the maturity of issues like content creation,financial practices, e-business project management, riskmanagement, etc.

Concepts denoted as ‘‘#’’ expand further to lower levelmaps. Similarly ‘‘"’’ denotes bottom-up causal propagation.

4.1.2. Social maturity metrics

The mechanism proposes the tax and legal top-level map(Fig. 10a) to support reasoning on the maturity of issueslike tax planning, web site shopping, VAT compliance, res-idence/permanent establishments, tax compliance, transferpricing, contracting with customers, contracting with sup-pliers, regulatory controls, unlawful content, etc.

4.1.3. Infrastructure maturity metrics

The mechanism proposes the Systems and Technology

top-level map (Fig. 10b) to interconnect concepts like infra-structure, capacity planning and management, web quality,encryption, maintenance, network security, technologyselection, database security and control, etc.

4.1.4. Integrated maturity metricsThe mechanism proposes two basic maturity maps con-

sisting of generic e-business maturity metrics as follows:

• The Channel management top-level map (Fig. 11a) sup-ports reasoning on the maturity of issues like channelmanagement, customer integration processes, logisticsmanagement, pricing, direct procurement, productdevelopment, targeted promotion.

, (b) performance management concepts.

Fig. 10. (a) Tax and legal concepts map, (b) systems and technology concepts map.

Fig. 11. (a) Channel management concepts map, (b) high level concepts map.

698 G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702

• The High Level FCM top-level map (Fig. 11b) intercon-nects all maturity domains to reason on the overall e-business maturity of the enterprise.

4.2. Preliminary experiments

Two experiments were conducted by utilizing metricsfrom actual (though random) e-business strategic changeexercises in two major financial sector enterprises. For eachexperiment, a team of experts was engaged to provide lin-guistic variables for the causal weights, the concept valuesand the coefficients values to let the FCM algorithm reasonabout the impact of potential change initiatives, as well asto provide their independent expert estimates (using similarlinguistic variables) of the impact of the strategic changechoices to specific maturity metrics.

Fig. 12 compares the e-business maturity as estimatedby FCMs and the team of experts respectively for the firstexperiment.

The majority of the concepts cascade to several constit-uent metrics, hence the tool traverses complicated conceptinterrelations spreading over different maps and hierar-chies. The FCM mechanism calculated the value of affectedconcepts based on the initial weight and concept value.Similarly, Fig. 13 e-business maturity as estimated by theFCM mechanism and the team of experts respectively forthe second experiment.

4.3. Discussion

Various aspects of the proposed modeling mechanismare now commented on. As far as the theoretical valueis concerned, the proposed mechanism extends previousresearch attempts by (a) introducing a novel supplementto e-business strategy formulation activities which adaptsbetter to the characteristics of e-business initiatives, (b)introducing a holistic framework of e-business maturityassessment, (c) introducing the notion of interconnectedmaturity hierarchies, (d) concentrating on the actual strate-

0.0

0.2

0.4

0.6

0.8

1.0

strat

egic

partn

ersh

ips

inter

nal e

biz ch

ange

s

e-pr

ocur

emen

t man

agem

ent

ebiz

HRM

finan

cial c

ontro

l

cros

s fun

ction

al bu

sines

s cont

rol

cont

ract

man

agem

ent

custo

mer

com

plaint

s man

agem

ent

ITint

erop

erab

ility

ebiz

infra

struc

ture

ITstr

ateg

y & p

lannin

g

prod

uct d

evelo

pmen

t

chan

nel m

anag

emen

t

e-pr

omot

ion

risk na

mag

emen

t

deliv

ery op

erat

ions

orga

nizat

ion &

com

peta

ncies

syste

ms &

tech

nolog

y

FCMs

Experts

Fig. 12. E-business maturity – Scenario A.

0.0

0.2

0.4

0.6

0.8

1.0

strat

egic

partn

ersh

ips

inter

nal e

biz ch

ange

s

e-pr

ocur

emen

t man

agem

ent

ebiz

HRM

finan

cial c

ontro

l

cros

s fun

ction

albu

sines

s con

trol

cont

ract

man

agem

ent

custo

mer

com

plaint

s man

agem

ent

ITint

erop

erab

ility

ebiz

infra

struc

ture

ITstr

ateg

y &pla

nning

prod

uct d

evelo

pmen

t

chan

nel m

anag

emen

t

e-pr

omot

ion

risk na

mag

emen

t

deliv

ery op

erat

ions

orga

nizat

ion &

com

peta

ncies

syste

ms &

tech

nolog

y

FCMs

Experts

Fig. 13. E-business maturity – Scenario B.

G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702 699

gic formulation activity and its impact on the e-businessmodel, (e) allowing fuzzy definitions in the cognitive maps,(f) introducing an interpretation mechanism of fuzzy sets,(g) proposing an updated FCM algorithm to suit betterthe e-business maturity domains, (h) allowing dynamicmap decomposition and reconfiguration.

As far as the practical value of the proposed mechanismis concerned:

• The mechanism does not provide fundamentally differ-ent ‘‘diagnosis’’, compared to the expert estimates. Itprovides reasonably good approximations of e-businessmaturity changes.

• The mechanism tends to under-estimate slightly thematurity of concepts which have several hierarchicaldependencies. This conservatism, however, does notreduce the effectiveness of the proposed mechanism. Itsimply indicates that when complex maturity factors

are involved, it may be safer to assume a conservativestrategic impact scenario.

• The proposed mechanism provides a uniform behaviorregardless of the degree of maturity gap. The first exper-iment (see Fig. 12) involved a financial sector enterprisewith limited electronic presence; while the second exper-iment (see Fig. 13) involved an enterprise with establishelectronic presence seeking to enhance further its alter-native delivery channels.

• The justification of the ‘‘diagnosis’’ (essentially thematurity decomposition) proved helpful in comprehend-ing the sequence of concept interactions (essentially thematurity roadmap).

• The concept-based approach did not restrict theinterpretation of the estimated e-business maturity.The fuzzy interpretation of concept and weight valuesserved as indications rather than precise arithmeticcalculations.

700 G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702

Having established the theoretical and practical value ofthe proposed mechanism, it is useful to discuss also theadded value of incorporating such a mechanism intoe-business strategy formulation exercises. It is the beliefof this paper that the resulting tool provides real value tothe principle beneficiaries and stakeholders of e-businessprojects. For example:

• The mechanism eases the complexity of deriving expertdecisions concerning e-business initiatives. Informalexperiments indicated that the time required by expertsto estimate manually the extensive impact of major stra-tegic changes could pose as a considerable overhead. Onthe other hand the elapsed time for automated estima-tions using FCM decision support can be insignificant,once the map hierarchies have been set up.

• To extend further this syllogism, realistic e-businessstrategy formulation projects should involve continuousargument of change options until an equilibrium solu-tion accepted by all stakeholders has been agreed upon.Informal discussions with the principle beneficiaries andstakeholders of the two e-business projects revealed thatthe proposed FCM decision support can reduce signifi-cantly the maturity estimation overheads, letting thestakeholders focus on the actual strategic planningexercise while exploring in depth alternative objectivesand controlling effectively major strategic changeinitiatives.

• The proposed mechanism can also assist the maturityevaluation of the enterprise on a regular basis. FCMsmay serve as a back end to performance scorecards(Bourne, Mills, Wilcox, Neely, & Platts, 2000; Kaplan& Norton, 1996, 2001) to provide holistic strategic per-formance evaluation and management. However adetailed analysis of this extension falls out of the scopeof this paper.

Senior managers of the two major financial sector enter-prises have evaluated the usability of the proposed tool andhave identified a number of benefits that can be achievedby its utilization as a methodology framework for e-busi-ness maturity measurement. Detailed presentation of theusability evaluation results fall out of the scope of thepaper. However, a summary of major benefits is providedto improve the autonomy of this paper:

• Shared Goals: Concept-driven simulation pulls individu-als together by providing a shared direction and deter-mination of strategic change.

• Shared Culture: All business units feel that their individ-ual contribution is taken under consideration and pro-vide valuable input to the whole strategic changeprocess.

• Shared Learning: The enterprise realizes a high returnfrom its commitment to its human resources, establish-ing a constant stream of improvement within theenterprise.

• Shared Information: All business units and individualshave the necessary information needed to set clearlytheir individual objectives and priorities, while seniormanagement can control effectively all aspects of thestrategic change process.

Summarizing, experimental results showed that FCM-based ex ante reasoning of the impact of e-business matu-rity changes (actual or hypothetical) can be effective andrealistic. This is considered to be a major contribution ofthe proposed methodology to strategic change exercises.

5. Conclusion

This paper presented an intelligent supplement to typicale-business strategy formulation methodologies based onfuzzy cognitive maps (FCM). This decision aid mechanismproposed a new domain-based approach to supplementingthe current status analysis and objectives setting phases oftypical e-business strategy formulation projects, by sup-porting ‘‘intelligent’’ modeling of e-business maturity and‘‘intelligent’’ reasoning of the anticipated impact of strate-gic change initiatives. By using FCM, the proposed mecha-nism drew a causal representation of e-business maturityprinciples; it simulated the operational efficiency of complexstrategy models with imprecise relationships and quantifiedthe impact of strategic change to the e-business model.

Preliminary experimental results indicated that themechanism did not provide fundamentally different esti-mates than expert decisions. It provided reasonably goodestimates of the impact of strategic change initiatives tothe e-business model, while the maintenance effort didnot pose as a prohibitory factor. Moreover, the decompo-sition of maturity metrics supported reasoning of the per-formance roadmap and the complex relationships thataffect the overall e-performance.

The proposed mechanism should not be regarded onlyas an effective e-business modeling support tool. Its mainpurpose is to drive strategic change activities rather thanlimit itself to qualitative simulations. Moreover, the pro-posed mechanism should not be seen as an ‘‘one-off’’ deci-sion aid. It should be a means for setting a course forcontinuous strategic alignment (Langbert & Friedman,2002).

Future research will focus on conducting further real lifeexperiments to test and promote the usability of the tool,but also to identify potential pitfalls. Furthermore, futureresearch will focus on the automatic determination ofappropriate fuzzy sets (e.g. utilizing pattern recognition,mass assignments, empirical data, etc.) for the representa-tion of linguistic variables to suit each particular e-businessproject domain. Finally, further research will focus onimplementing backward map traversal, a form of adbuctivereasoning (Flach & Kakas, 1998). This feature offers thefunctionality of determining the condition(s) Cij thatshould hold in order to infer the desired Cj in the causalrelationship Cij!

wjkCk. Incorporating integrity constraints

G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702 701

reduces the search space and eliminates combinatorysearch explosion. Backward reasoning has been testedextensively in other applications and its integration in theproposed methodology may prove beneficiary.

References

Albrecht, C. C., Dean, D. D., & Hansen, J. V. (2003). Using situationcalculus for e-business agents. Expert Systems with Applications, 24,391–397.

Axelrod, R. (1976). Structure of decision: The cognitive maps of politicalelites. Princeton, NJ: Princeton University Press.

Bharati, P., & Chaudhury, A. (2003). An empirical investigation ofdecision-making satisfaction in web-based decision support systems.Decision Support Systems, 1051, 1–11.

Bourne, M., Mills, J., Wilcox, M., Neely, A., & Platts, K. (2000).Designing, implementing and updating performance measurementsystems. International Journal of Operations and Production Manage-

ment, 20, 754–771.Burgess, T. F. (1998). Modeling the impact reengineering with system

dynamics. International Journal of Operations and Production Man-

agement, 18(9/10), 950–963.Burgess, T. F. (1996). Modelling quality-cost dynamics. International

Journal of Quality and Reliability Management, 13(3), 8–26.Carlsson, C., & Turban, E. (2002). DSS: directions for the next decade.

Journal of Decision Support Systems, 33(1), 105–110.Chung, W., Yam, A. Y. K., Chan, M. F. S., & Potter, A. (2005).

Networked enterprise: A new business model for global sourcing.Production Economics, 45(2), 23–42.

Craiger, J. P., Goodman, D. F., Weiss, R. J., & Butler, A. (1996).Modeling Organizational Behavior with Fuzzy Cognitive Maps.Journal of Computational Intelligence and Organisations, 1, 120–123.

Crowe, T. J., Fong, P. M., Bauman, T. A., & Zayas-Castro, J. L. (2002).Quantitative risk level estimation of business process reengineeringefforts. Business Process Management Journal, 8(5), 490–511.

Dickerson, J. A., & Kosko, B. (1997). Virtual Worlds as Fuzzy CognitiveMaps. In B. Kosko (Ed.), Fuzzy engineering (pp. 125–141). UpperSaddle River, NJ: Prentice-Hall.

Diffenbach, J. (1982). Influense diagrams for complex strategic issues.Strategic Management Journal, 3, 133–146.

Disney, S. M., Naim, M. M., & Potter, A. (2005). Assessing the impact ofe-business on supply chain dynamics. Journal of Production Economics,

89(2), 109–118.Duffy, J. (2001). Maturity models. Blueprints for e-volution. Strategy and

Leadership, 29(6), 19–26.Flach, P. A., & Kakas, A. (1998). Abduction and Induction in Al: Report

of the IJCAI97 workshop. Logic Journal of the Interest Group on Pure

and Applied Logic, 6(4), 651–656.Georgopoulos, V., Malandraki, G., & Stylios, C. (2002). A fuzzy cognitive

map approach to differential diagnosis of specific language impair-ment. Journal of Artificial Intelligence in Medicine, 679, 1–18.

Hagiwara, M. (1992). Extended fuzzy cognitive maps. In Presented at

proceedings of the 1st IEEE international conference on fuzzy systems,New York.

Harmon, P., & King, D. (1985). Expert Systems: Artificial Intelligence inBusiness. New York, NY: John Wiley & Sons.

Hong, T., & Han, I. (2002). Knowledge based data mining of newsinformation of the internet using cognitive maps and neural networks.Journal of Expert Systems with Applications, 23(1), 1–8.

Hooft, F. V., & Stegwee, R. A. (2001). E-business strategy: how to benefitfrom a hype. Logistics Information Management, 14(1/2), 44–53.

Jackson, P., & Harris, L. (2003). E-business and organisational change.Reconciling traditional values with business transformation. Journal of

Organizational Change Management, 16(5), 497–511.Johnson, R. J., & Briggs, R. O. (1994). A model of cognitive information

retrieval for ill-structured managerial problems and its benefits for

knowledge acquisition. In Presented at 27th annual Hawaii interna-

tional conference on system sciences, Hawaii.Jones, R. T., & Ryan, C. (2002). Matching process choice and uncertainty:

Modeling quality management. Journal of Business Process Manage-

ment, 8(2), 161–168.Kaplan, R. S., & Norton, D. P. (1996). Using the balanced scorecard as a

strategic management system. In Harvard Business Review (pp. 75–85),vol. January/February.

Kaplan, R. S., & Norton, D. P. (2001). Leading change with the balancedscorecard. Financial Executive, 17, 64–66.

Kaufmann, A. (1975). Introduction to the theory of of fuzzy subsets.Fundamental theoretical elements. New York: Academic Press.

Klein, J. C., & Cooper, D. F. (1982). Cognitive maps of decision makers ina complex game. Journal of Operation Research Society, 33, 63–71.

Koh, J., & Kim, Y. G. (2005). Knowledge sharing in virtual communities:an e-business perspective. Expert Systems with Applications, 26(2),155–166.

Kosko, B. (1986). Fuzzy Cognitive Maps. Journal of Man-Machine

Studies, 24, 65–75.Kosko, B. (1991). Neural networks and fuzzy systems. Englewood Cliffs:

Prentice Hall.Kwahk, K. Y., & Kim, Y. G. (1999). Supporting business process redesign

using cognitive maps. Decision Support Systems, 25(2), 155–178.Langbert, M., & Friedman, H. (2002). Continuous improvement in the

history of human resource management. Journal of Management

Decision, 40(8), 782–787.Lee, K. C., & Kim, H. S. (1997). A fuzzy cognitive map-based bi-

directional inference mechanism: An application to stock investmentanalysis. Journal of Intelligent Systems in Accounting Finance and

Management, 6(1), 41–57.Lee, K. C., & Kim, H. S. (1998). Fuzzy implications of fuzzy cognitive

map with emphasis on fuzzy causal relationship and fuzzy partiallycausal relationship. Journal of Fuzzy Sets and Systems, 3, 303–313.

Lee, K. C., Kim, J. S., Chung, N. H., & Kwon, S. J. (2002). FuzzyCognitive Map approach to web-mining inference amplification.Journal of Expert Systems with Applications, 22, 197–211.

Lee, K. C., & Kwon, O. B. (1998). A strategic planning simulation basedon cognitive map knowledge and differential game. Journal of

Simulation, 7(5), 316–327.Lin, F. R., Yang, M. C., & Pai, Y. H. (2002). A generic structure for

business process modeling. Business Process Management Journal, 8(1),19–41.

Liu, Z. Q. (2000). Fuzzy cognitive maps: Analysis and extension. Tokyo:Springer.

Liu, Z. Q., & Satur, R. (1999). Contexual fuzzy cognitive map for decisionsupport in geographical information systems. Journal of IEEE

Transactions on Fuzzy Systems, 7, 495–507.Long, P. H., & Schoenberg, R. (2002). Does e-Business Require Different

Leadership Characteristics? An Empirical Investigation. European

Management Journal, 20(6), 611–619.Mahajan, V., & Venkatesh, R. (2000). Marketing modeling for e-business.

Research in Marketing, 17, 215–225.Metaxiotis, K., Psarras, J., & Samouilidis, E. (2003). Integrating fuzzy

logic into decision support systems: current research and futureprospects. Journal of Information Management and Computer Security,

11(2), 53–59.Murray, M. A., Priesmeyer, H. R., Sharp, L. F., Jensen, R., & Jensen, G.

(2000). Nonlinearity as a tool for business process reengineering.Business Process Management Journal, 6(4), 304–313.

Ndouse, T. D., & Okuda, T. (1996). Computational intelligence fordistributed fault management in networks using fuzzy cognitive maps.In Presented at IEEE international conference communication converg-

ing technology tomorrow’s application, New York.Noh, J. B., Lee Lee, K. C., Kim, J. K., Lee, J. K., & Kim, S. H. (2000). A

case-based reasoning approach to cognitive map driven-driven tacitknowledge management. Journal of Expert Systems with Applications,

19, 249–259.

702 G. Xirogiannis, M. Glykas / Expert Systems with Applications 32 (2007) 687–702

Palmer, R. (2002). There’s no business like e-business. Qualitative Market

Research, 5(4), 261–267.Pelaez, C. E., & Bowles, J. B. (1995). Applying fuzzy cognitive maps

knowledge representation to failure modes effect analysis. In Presented

at IEEE annual reliability maintainability symposium, New York.Perusich, K. (1996). Fuzzy cognitive maps for political analysis. In

Presented at international symposium on technology and society.Technical Expertise and Public Decisions, New York.

Phan, D. D. (2003). E-business development for competitive advantages: acase study. Information and Management, 40, 581–590.

Ramaprasad, A., & Poon, E. (1985). A computerised interactive techniquefor mapping influense diagrams (MIND). Strategic Management

Journal, 6(4), 377–392.Schneider, M., Schnaider, E., Kandel, A., & Chew, G. (1995). Construct-

ing fuzzy cognitive maps. In Presented at IEEE conference on fuzzy

systems, New York.Shim, J. P., Warkentin, M., Courtney, J. F., Power, D. J., Sharda, R., &

Carlsson, C. (2002). Past, present and future of decision supporttechnology. Journal of Decision Support Systems, 33(1), 111–126.

Silva, P. C. (1995). New forms of combinated matrices of fuzzy cognitivemaps. In Presented at IEEE international conference on neural

networks, New York.

Sprague, R. H., & Watson, H. J. (1986). Decision Support Systems: Putting

Theory into Practice. Englewood Cliffs, NJ: Prentice-Hall.Stylios, C. D., Georgopoulos, V. C., & Groumpos, P. P. (1997).

Introducing the theory of fuzzy cognitive maps in distributed systems.In Presented at 12th IEEE international symposium on intelligent

control, Istanbul, Turkey.Xirogiannis, G., & Glykas, M. (2004a). Fuzzy Cognitive Maps as a Back

End to Knowledge-Based Systems in Geographically DispersedFinancial Organizations. Knowledge and Process Management, 11,1–18.

Xirogiannis, G., & Glykas, M. (2004b). Fuzzy cognitive maps in businessanalysis and performance-driven change. IEEE Transactions on

Engineering Management, 13(17), 111–136.Xirogiannis, G., Stefanou, J., & Glykas, M. (2004). A fuzzy cognitive map

approach to support urban design. Journal of Expert Systems with

Applications, 26(2), 257–268.Zadeh, L. A. (1965). Fuzzy Sets. Journal of Information and Control, 8,

338–353.Zhang, W. R., Wang, W., & King, R. S. (1994). A-pool: An agent-oriented

open system for distributed decision process modeling. Journal of

Organisational Computing, 4(2), 127–154.