The use of a hybrid fuzzy-Delphi-AHP approach to develop global business intelligence for information service firms

Expert Systems with Applications 37 (2010) 7394–7407

Contents lists available at ScienceDirect

Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa

The use of a hybrid fuzzy-Delphi-AHP approach to develop global business intelligence for information service firms Ming-Kuen Chen a, Shih-Ching Wang b,c,* a

Graduate Institute of Commerce Automation and Management, National Taipei University of Technology, 1, Sec. 3, Chung-Hsiao E. Rd., Taipei 106, Taiwan Graduate Institute of Industrial and Business Management, National Taipei University of Technology, 1, Sec. 3, Chung-Hsiao E. Rd., Taipei 106, Taiwan c Regional R&D Service DEPT., Metal Industries Research and Development Centre, 3, Sec. 8F, 162, Sinyi Rd., Taipei 106, Taiwan b

a r t i c l e

i n f o

Keywords: Information services industry Alliance Modified Delphi approach Fuzzy modified Delphi method Fuzzy analytic hierarchy process (FAHP) Business intelligence (BI)

a b s t r a c t Due to globalization and saturated domestic markets, information service firms, upon growing to a certain size, gradually focus their business efforts on reaching global markets. In order to reduce business risk in developing international markets, using the alliance model is a key strategy for information service firms. On the other hand, firms should handle more accurate business information to support their business intelligence (BI) system to make better business decisions. This research uses a hybrid fuzzy-DelphiAHP approach to propose a more comprehensive framework with specific business elements, and also points out six performance indices for firms to adjust business strategy. Results of this study could have considerable value for the information services industry to develop international markets. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction Despite the financial meltdown’s strong impact on the global economy in 2008, Gartner Dataquest research reported that the information service industry worldwide has exhibited a Compound Annual Growth Rate (CAGR) of 3.43% from 2007 to 2013, and that total production value of the global information service industry market reached US$744 billion in 2007 and was up to US$806 billion in 2008. In Asia, the CAGR is predicted to grow at a rate of 4.45% in the future. Hence, the information service industry is considered a developable sector. However, information service firms responding to saturation in domestic markets and the tendency of internationalization have focused on core competencies and operative strategies to ensure their firm’s survival. The development of Taiwan’s information service industry is one example in which CAGR of production value has 12.11% and export value is up to 10.9% from 2005 to 2009. Apparently, Taiwan’s information service industry in the global market is growing steadily. According to the statistics from the Department of Statistics Ministry of Economic Affairs in Taiwan in 2009, the production value of the information service industry was NT$239 billion (2008, NT$225.8B; 2007, NT$211.2B) and export value was NT$35.0 billion (2008, NT$33.5B; 2007, NT$31.3B). From the standpoint of economic development potential, Taiwan has an important geographic loca-

* Corresponding author. Address: Graduate Institute of Industrial and Business Management, National Taipei University of Technology, 1, Sec. 3, Chung-Hsiao E. Rd., Taipei 106, Taiwan. Tel.: +886 2 27541255x2455; fax: +886 2 27541058/ 27019181. E-mail address: [email protected] (S.C. Wang). 0957-4174/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2010.04.033

tion in Asia’s major economic region, and Taiwan has already developed numerous software systems for the manufacturing industry. Therefore, Taiwan should assess information service sectors having growth potential, and determine the best software service solutions for penetrating the global market. On the other hand, the internationalization of information service firms introduces complex issues different from those faced by firms targeting only domestic markets and these issues affect the types of business models adopted. For instance, internationalized firms should comprehend global market characteristics, product core competencies, and global customer needs. In order to reduce business risk, firms need enough capital to develop international markets – particularly in countries whose domestic market is limited, such as Taiwan and Finland. For this reason, in recent years, Taiwan’s Industrial Development Bureau of the Ministry of Economic Affairs (MOEAIDB) has promoted an ‘‘information services flagship-program” (also called the ‘‘BEST-Program”) (Industrial Development Bureau of Ministry of Economic Affairs (MOEAIDB), 2004) which main purpose of BEST-program is to improve the development of individual firms and push the backbone firms to activate the cooperated firms to form a strategically cooperative way for opening international target markets. Each project (called ‘‘title-flagship-project”) in this program should select a large-scale information service (‘‘Flagship”) having both competitive ability and large growth potential to collaborate with more than three software companies (‘‘Partners”) with complementary core abilities to form an alliance (‘‘Fleet”) to develop target international markets. Apparently, the alliance, or collaborative, model is a safer strategy for SEM firms entering the global market. Additionally, in order to expand effectively and quickly adjust to developing

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407

global market conditions, information service firms should monitor business information using key performance indicators related to critical activities of their business. Briefly, useful information is necessary to business intelligence (BI) upon which information service firms builds better business strategies (Wang, 2005). Therefore, in light of the importance of BI to alliance strategy for developing global markets, this research, based on the framework of Chen and Wang (2010), further integrates BI-based research to propose a more comprehensive framework built on specific business operation factors, and lists performance indicators through a hybrid-fuzzy-Delphi-AHP approach. The results are valuable for firms or governments planning related strategies in entering international markets.

2. Literature review Hoch, Roeding, Purkert, Lindner, and Müller (2000) has divided software firms into package mass-market firms, enterprise solution firms, and professional service firms. Software product programs include working solutions (in pre-sales) and services such as implementation, training, hosting and product upgrades. That is, software implementation is always combined with service contents (Burgel & Murray, 2000; Ganek & Kloeckner, 2007). In Taiwan, the information service industry may be categorized into package, customization, and project industries (MIC, 2009). The business dimensions of information service firms are closely linked with the enterprise or strategic business models (Leidecker & Bruno, 1984). However, compared to physical products, software products are generally expensive to produce, but very cheap to reproduce (Chen, Wang, Chen, & Wang, 2007; Krishnamurthy, 2003; Nambisan, 2001; Westerlund, Rajala, & Svahn 2007). In addition, the information service industry is a knowledge-intensive industry (Rajala & Westerlund, 2007a; Rajala & Westerlund, 2007b), and provides a wide range of products and services to meet customer needs. Compared with other traditional industries, information services industries are ‘‘information-oriented” and provide ‘‘intangible goods”. Therefore, the business models of information service firms are different from other industries (Bonaccorsi, Giannangeli, & Rossi, 2006; Chen, Wang, & Chio, 2009; Hedman & Kalling, 2003; Lee, 2008; Magretta, 2002; Rajala, Rossi, & Tuunainen, 2003). Some researchers have pointed out different aspects of the influence of internationalization, such as pricing (Jain & Kannan, 2002; Sainio & Marjakoski, 2009) and product offerings (Bell, 1997; Ruokonen, 2008), customization (Burgel & Murray, 2000; McNaughton, 1996), intangible assets (Bieberstein, Bose, Walker, & Lynch, 2005). Rajala et al. proposed a reference model with four operative dimensions: product strategy, revenue logic, distribution model, and service and implementation model. Based on Ojala and Tyrväinen (2007), Rajala et al. (2003) proposed to analyze the entry model for software firms to enter the Japanese market. They also pointed out that product strategy is a key for software internationalization. Other researches quote the framework proposed in Rajala et al. (2003). Ojala and Tyrväinen (2006) used individual case studies to summarize and analyze the relationship between a business operation model and a market entry model based on the operation model framework provided by Rajala (2003). They also based on Rajala’s (2003) framework for fulfilling requirements in internationalization, suggested including the strategic technique of marketing regional or global products as a primary factor; most other studies also took Rajala (2003) research as a reference. However, Rajala et al. (2003), Rajala and Westerlund (2007) emphasized the concerns of individual firm businesses, but they fail to consider other key dimensions for entering the global market, e.g. market size and culture (Griffith, 2010; Ojala &

7395

Tyrväinen, 2006), outsourcing (Tiwana, 2004; Willcocks & Choi, 1995; Winkler, Dibbern, & Heinzl, 2008), service and consulting (Ganek & Kloeckner, 2007), domain knowledge (Rajala & Westerlund, 2007a; Rajala & Westerlund, 2007b; Tiwana, 2004), partnership (Coviello & Munro, 1997; Wild, Wild, & Han, 2008), and manpower (Barcus & Montibeller, 2008; Bieberstein et al., 2005). In order to propose a more comprehensive framework, taking careful consideration of other dimensions for information service industry alliances in developing international markets, Chen and Wang (2010) applied a cross case study method (Gable, 1994; Yin, 1994) to offer definitions of each factor, and proposed a business model framework for firms to develop target international markets; Chen and Wang (2010) also used an AHP approach to mark out weights of business elements, including market segment, strategic partners, service/implementation, product competition, distribution/channel model, and revenue efficiency. In addition, they also reported 20 critical factors (see Appendix A). Chen and Wang (2010) have contributed key insights in this framework for allied firms to develop international markets. However, the most important aim of information service firms desiring to expand their market is effective development of new markets. Therefore, firms should allocate capital and resources in a manner optimal for developing various markets. Thus, a BI system is an important tool for alliance performance when developing new markets. Several studies have suggested the benefit of BI for business performance. However, BI systems cover a wide range of tools and have broad scope. The most commonly mentioned BI applications are data warehouses, data mining, OLAP, decision support systems (DSS), and balance scorecards (BSC). The purpose of BI is to provide users with the best possible assistance in the process of decision-making (Eckerson Wayne, 2005 (chap. 3); Wang, 2005). Up to this point, however, there are few studies of performance indicators for alliances of information service firms to expand into the global market. Based on the above review, the chief aim in this article is to contribute the following: (a) to provide a more comprehensive framework with specific business elements, (b) to report performance indicators for alliances of information service firms, and (c) to determine weights of business elements Therefore, based on the literature review and in reference to the former studies, such as Rajala and Westerlund (2007a), Chen and Wang (2010), this study proposes a business strategic architecture in opening global market through a hybrid fuzzy-Delphi-AHP approach. Both Delphi and FAHP (fuzzy analytic hierarchy process) are useful for business strategic planning. The Delphi method is used to effectively create a consensus or conclusion about an issue through insight and knowledge of experts, and it is a qualitative contents; FAHP approach is quantitative approach for the enterprise to use for optimal decision making strategy in business (Basligil, 2005; Chan & Kumar, 2007; Chang, 1996). Due to the current business strategic architecture is still an immature model, there is need to use qualitative research with quantitative research. With this in mind, this study adopts the two methods with fuzzy theory. This article concludes with implications for theory, research, and practice; results of this research could increase success of the information services industry in developing global markets. 3. Research methodology 3.1. Research architecture The research architecture of this article is structured as shown in Fig. 1. Firstly, the author, based on the model proposed by Chen and Wang (2010), determines the final dimensions (and subcriteria) through fuzzy-modified-Delphi method. Secondly, this study reports the performance indicators through modified Delphi

7396

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407

Object: Propose business elements/sub-criteria Method: Fuzzy Modified Delphi Method

Object: Propose performance indicators Method: Modified Delphi Approach

Business Intelligence

Object: Point out weights of elements Method: Fuzzy AHP (FAHP)

Business strategy in developing global market : Supporting

Fig. 1. Research architecture by hybrid fuzzy-Delphi-AHP.

approach. Finally, the author uses available data in an AHPquestionnaire to propose overall weights of business elements/ sub-criteria through Fuzzy-AHP.

respondents, and the rate of returns (or effectiveness) for each research method. Due to concern for research cost and time, the study focuses on several Taiwanese larger-scale alliances for opening global markets.

3.2. Data collection The data collection method used in these approaches selected several typical information service firms participating in the ‘‘BEST-Program” (see Section 1). Because the gross scale of all participating firms, including lead companies (‘‘Flagships”) and their partners, has surpassed more than half of the information service industry in Taiwan, and has attained representative status, this study selected several decision-makers or experts of flagship-projects (of the BEST-Program), including 9 CEOs, 5 CIOs, 4 Project leaders, and 4 consultants. Each respondent (i.e., CEO or CIO) is a leader of some ‘‘Flagship” team, who has over 15 years experience in the information service industry domain. In addition, each ‘‘Flagship” represents a large strategy alliance with 3–5 cooperators (information service firms). Therefore, the sample size and composition can be strongly representative in Taiwan (such as ACER, MiTAC, and DSC.). Table 1 lists sample sizes of questionnaires,

Table 1 Sample collecting lists. Sample

Questionnaires Respondents Invalid samples Effectiveness Document(projects of firms)

Method A

B

C

25 23 2 84.0% 8

18 15 1 83.3% 8

25 24 4 80.0% 8

A: fuzzy modified Delphi method. B: modified Delphi approach. C: fuzzy AHP.

3.3. Business elements: a fuzzy modified approach Although Chen and Wang (2010) used a cross case study to propose a framework with six elements and 20 criteria, more studies need to be conducted to ascertain the quantity of elements. Therefore, this study uses a mathematical method to quantify these elements. Nevertheless, the business operative dimension involves much strategic policy and higher order decision-making. Thus, the author used fuzzy modified approach to find the degree of consensus from experts at these elements. Modified Delphi method (Murry & Hommons, 1995) is an applicable research instrument when there is incomplete knowledge about an issue. It is an iterative process for collecting and refining expert opinion. In order to obtain more complete knowledge, reduce time spent, and increase rate of return, in general, researchers reference highly relevant issues from literature prior to setting the Delphi-questionnaires in the first round. On the other hand, to prevent extreme values of expert opinions from obstructing convergence, the research has used fuzzy theory as a substitute for the traditional geometric average of the Delphi method to extract more precise weights of the dimensions. A fuzzy set is a class of objects with a continuum of grades of membership (Saaty, 1980, 1994). In fuzzy set theory, the membership function assigns to each object a grade of membership ranging between zero and one. Researchers always consider different issues with different membership functions; the most commonly used membership functions include triangles membership functions, trapezoid membership functions and Gauss membership functions. This research adopts triangles membership functions for the primary membership functions. The equation of triangles

v~A

LRA

LMA

LUA

Fig. 2. Triangle membership function.

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407 Table 2 Specialists-reply lists. Elements

Experts

Market segment Strategic alliance Service model Product strategy Distribution/channel strategy Revenue strategy New technology application Investment strategy Intellectual property Manpower training Others *

Mean

LRA

LMA

LUA

4.591 3.909 4.182 4.227 4.455 3.864 2.318 2.455 2.182 2.227 –

4 3 3 3 3 3 1 2 1 1 –

4.564* 3.829* 4.093* 4.147* 4.375* 3.779* 2.217 2.405 2.043 2.081 –

5 5 5 5 5 5 3 3 3 3 –

Selected elements (i.e. over the threshold ‘‘S” value = 3.0).

membership functions includes three parameters, i.e. LRA, LMA, and LUA, as shown as Fig. 2. The parameters LRA, LMA, and LUA, respectively, denote the smallest possible value of factor ‘‘A”, the most promising value, and the largest possible value that describes a fuzzy event. Each number in the pair-wise comparison matrix represents the subjective opinion of decision makers and is an ambiguous concept; fuzzy numbers work best to consolidate fragmented expert opinions. The triangular fuzzy numbers are established as follows:

m~ ¼ ðLRA ; LMA ; LU A Þ; LRA  LMA  LU A

ð1Þ

LRA ¼ MinðX Ai Þ;

ð2Þ

i ¼ 1; 2; 3; . . . ; n

LMA ¼ ðX Al  X A2  . . . X An Þ1=n

ð3Þ

LU A ¼ MaxðX Ai Þ;

ð4Þ

i ¼ 1; 2; 3; . . . ; n

~A is triangles membership functions where A is problem (factor); m for ‘‘A”; i is expert; XAi is the evaluated number of ‘‘A” from the ith expert; LRA is the smallest number of ‘‘A” from experts; LMA is the mean number of ‘‘A” from experts; and LUA is the largest number of ‘‘A” from experts. Base on the definition of elements and sub-criteria proposed by Chen and Wang (2010), this study utilizes FMDM to establish a series of Delphi-questionnaires with five point types (1: strongly

7397

disagree, 2: disagree, 3: slightly agree, 4: agree, 5: strongly agree). The authors implemented the questionnaire in three rounds (since June 2008). According to the data obtained from the questionnaire, the author calculated the weight of fuzzy numbers – LMA values of dimensions – through software (Microsoft Excel 2003). The final results of six dimensions total are extracted from over 13 elements, such as new technology application, investment strategy, intellectual property, and manpower training, as listed in Table. 2. Based on the Delphi method, the study set a suitable threshold of ‘‘S = 3.0” by expert consensus, and eliminate LMA’s having ‘‘S” less than three (Table 2). Obviously, these selected elements are the most influential factors for the information service industry to enter global markets (Fig. 3). The authors obtained results similar to the results found in formal research (Chen & Wang, 2010), and some elements are slightly adjusted, such as ‘‘service/implementation model” becomes ‘‘service model”, ‘‘competition of product” becomes ‘‘product strategy”, and ‘‘revenue efficiency” becomes ‘‘revenue strategy”. After obtaining the main elements described above, the authors once again ran the fuzzy-Delphi approach to refine these sub-criteria. Finally, this study takes two rounds to list 13 sub-criteria (i.e. LMA value greater than ‘‘S = 3.0”), such as cognition of regional culture, cognition of market scale, and market geographical position (Appendix A). The authors find the results of sub-criteria are more appropriate and accurate items than those proposed by formal research (Chen & Wang, 2010) (Appendix A). For example, the five sub-criteria of service model are adjusted into two sub-criteria; and the three sub-criteria of distribution/channel strategy are adjusted into two sub-criteria. Obviously, the 13 sub-criteria are more comprehensive than the formal 20 sub-criteria proposed by Chen and Wang (2010). Thus, this study proposes a more comprehensive framework (Fig. 4) for the information service firms to plan entry strategies for global markets than the formal model proposed. Moreover, the results of this study are adequately integrated with the consensus of experts and business decisionmakers; higher expert and content validity are achieved. 3.4. Performance indicators: a modified Delphi approach Regarding BI concerns, as described in Section 2, firms need to watch appropriate indicators to monitor business performance

Fig. 3. Radar map of expert opinions by FMDM.

7398

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407

Distribution/channel strategy Product strategy -Product function and quality -Competition of product

-Distribution channel Model -Distribution channel efficiency

Revenue strategy -Revenue model -Pricing strategy

Market segment -Cognition of regional culture -Cognition of market scale -Market’s geographical position

Strategic alliance -Positioning the cooperative role -The degree of cooperation

Service strategy -Integrity of service mechanism -The depth of knowledge of industrial domain

Fig. 4. Framework of information service industry in developing global markets by fuzzy-Delphi approach.

Table 3 The performance indicators of six business elements. Dimensions

Indicator

Relevant study

Market segment Strategic alliance Service model

Number of markets/subsidiaries (NMS) Partner scale (PC) Integrity of service mechanism (or SOP) (SSOP)

Product strategy Distribution/channel strategy Revenue strategy

Product/service complement (PSC) Number of distribution/channel (NDC) Total revenue of alliance (TRA)

Chen and Wang (2010), Chen and Wang (2010), Chen and Wang (2010), Rajala et al. (2003) Chen and Wang (2010), Chen and Wang (2010), Chen and Wang (2010),

for evaluating their strategy in developing global markets. Furthermore, based on the framework (Fig. 4) obtained by fuzzy-Delphi approach, the authors use modified Delphi approach again to extract evaluation indicators of the six elements for alliances of firms in global markets. The steps of the modified method are summarized below:

Winkler et al. (2008) Shapiro and Varian (1999), Willcocks and Choi (1995) Krishnan, Kriebel, Kekre, and Mukhopadhyay (2000), Bell (1997), Rajala et al. (2003) Rajala et al. (2003) Rajala et al. (2003)

firms to develop international markets in the six business elements?” (3) Perform a questionnaire survey in the 2nd round. (4) Perform a questionnaire survey in the 3rd round. (5) Synthesize expert opinions to form a consensus. In total, the authors operated the questionnaire over four rounds as listed in Appendix B. The evaluation performance indicator terms for alliances of information service firms are listed in Table 3. After introducing the new framework and listing the six performance indicators above, the research shows a conceptual model (Fig. 5) that shows a two-way relationship between the BI system

(1) Select the experts in the field: Based on literature review of performance indices, the authors selected 12 experts from the ‘‘flagship-projects” group (Table 1). (2) Perform a questionnaire survey in the 1st round: The openended questionnaire could include such questions as ‘‘What are the evaluation indicators for alliances of information service

Distribution/ channel strategy -Distribution channel Model -Distribution channel efficiency Product strategy -Product function and quality -Competition of Product

Service model -The depth of knowledge of industrial domain -Integrity of service mechanism

NDC PSC

SSOP BI System

PC

Strategic alliance -Positioning the cooperative role -The degree of cooperation

TRA NMS

Revenue strategy -Revenue model -Pricing strategy

Market segment - Cognition of regional culture -Cognition of market scale -Market’s geographical position

Fig. 5. The relationship between the BI and framework (conceptual model).

7399

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407

and the business framework; thus, firms may effectively adjust their market entry policy while obtaining real information from these performance indicators through their BI system. Hence, this model is more convenient for firms to build an entry strategy for markets spanning the globe.

practical cases, the human preference model is uncertain and decision-makers might be reluctant or unable to assign exact numerical values to the comparison judgments (Kahraman, Cebeci, & Ulukan, 2003). Since some of the evaluation criteria are subjective and qualitative in nature, it is very difficult for the decision-maker to express his/her preferences using exact numerical values and provide exact pairwise comparison judgments. The traditional AHP cannot be applied in a straightforward manner to solving uncertain decision-making problems (Deng, 1999). This is because the researcher is typically unable to explicitly relate his/her preferences due to the fuzzy nature of the comparison process (Kahraman et al., 2003). In order to overcome all these shortcomings, FAHP was developed for solving the hierarchical problems. Buckley (1985) initiated trapezoidal fuzzy numbers to express the decision maker’s evaluation on alternatives with respect to each criterion while Laarhoven and Pedrycz (1983) were using triangular fuzzy numbers. Chang (1996) introduced a new approach for handling FAHP, with the use of triangular fuzzy numbers for pair-wise comparison scale of FAHP, and the use of the extent analysis method for the synthetic extent values of the pair-wise comparisons. Based on the new framework (Fig. 4) by fuzzy-Delphi method, as expressed in Section 3.1, this study uses FAHP to assign more accurate weight values for each item (i.e. elements and subcriteria). Obviously, based on the reference architecture using fuzzy-Delphi method, the results (i.e., weights) applying FAHP approach is useful for information service firm to plan or choose a better strategy in developing global market. According to FAHP approach, there are eight steps:

3.5. Overall weights of elements: a FAHP approach Based on the framework (Figs. 4 and 5), firms need to realize the significance (such as weight or priority) of each element/sub-criteria to allocate business resources or make better decisions in developing global markets. This issue becomes a multi-criteria decisionmaking (MCDM) problem. However, analytic hierarchy process (AHP) is a widely used MCDM tool first proposed by Saaty (1980), Saaty (1994); AHP is a tool accessible to decision makers and researchers; and, it is one of the most widely used MCDM tools (Vaidya & Kumar, 2006). In traditional AHP, pair-wise comparison is made using a nine-point scale, which converts human preferences (i.e., numbers 3, 5, 7, and 9 meaning ‘generally important’, ‘strongly important’, ‘very important’, and ‘absolutely important’; and 2, 4, 6, and 8 for compromises between 3, 5, 7, and 9). Thus the AHP uses only absolute scale numbers for judgments and for their resulting priorities. Even though the discrete scale of AHP has the advantages of simplicity and ease of use, it is not sufficient to take into account the uncertainty associated with the mapping of one’s perception to a number. In spite of its popularity and simplicity in concept, this method is often criticized for its inability to adequately handle the inherent uncertainty and imprecision associated with the mapping of the decision-maker’s perception to exact numbers. In the traditional formulation of the AHP, human judgments are represented as exact numbers. However, in many

The business elements/ criteria of information service firms in entering global market

Level 1 -Goal

(1) Determine problems: Determine the problems to be solved, so as to ensure future analyses are correct (this study focuses

Alternative

Level 3 -Sub-Criteria •Cognition of regional culture(C1) Market Segment (E1) •Cognition of market scale(C2) •Market’s geographical position(C3) Level 2 -Elements

Alternative 1

Strategic Alliance (E2)

•Positioning the cooperative role(C4) •The degree of cooperation(C5)

Service Model (E3)

•Integrity of service mechanism(C6) •The depth of knowledge of industrial domain(C7)

Product Strategy (E4)

•Product function and quality(C8) •Competition of product (C9)

Distribution/ Channel Strategy (E5)

•Distribution channel Model (C10) •Distribution channel efficiency(C11)

Revenue Strategy (E6)

•Revenue model (C12) •Pricingstrategy(C13)

Alternative 2

Alternative n

Fig. 6. FAHP hierarchy structure.

~ 1

µ ~(x)

~ 3

~ 2

~ 4

~ 6

~ 5

~ 8

~ 7

~ 9

A

0

1

2

3

4

5

6

Fig. 7. Scale of fuzzy numbers.

7

8

9

x

7400

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407

Table 4 Definition of fuzzy numbers.

kmax ¼

n X

aij ðwi =wj Þ

ð6Þ

j

Fuzzy number

Definition

~ ¼ ð1; 1; 1Þ 1 ~ ¼ ð1; 2; 3Þ 2

Equally important

~ ¼ ð2; 3; 4Þ 3 ~ ¼ ð3; 4; 5Þ 4

Moderately more important

~ ¼ ð4; 5; 6Þ 5 ~ ¼ ð5; 6; 7Þ 6

Strongly more important

~ ¼ ð6; 7; 8Þ 7 ~ ¼ ð7; 8; 9Þ 8

Very strongly more important

~ ¼ ð8; 9; 10Þ 9

Extremely more important

CI ¼ ðkmax  mÞ=ðm  1Þ

Judgment values between equally and moderately Judgment values between moderately and strongly Judgment values between strongly and very strongly Judgment values between very strongly and extremely

on ‘‘Business Intelligence for Alliances of Information Service Industry Firms in Developing Global Markets”). (2) Set up the hierarchy architecture: Determine the key elements and sub-criteria as the hierarchic layer of FAHP through FMDM as shown in Fig. 4. This study set up the hierarchy architecture as shown in Fig. 6. The first layer is the ultimate goal, the second layer is the element, and the third layer is the sub-criteria. Under this framework, the authors summarize questionnaires and calculate weights by FAHP. With this framework, firms select a better alternative (e.g. joint venture, subsidiary or dealer) depending on their business operation needs for entering the market. ~ij ): Compare the (3) Set up fuzzy paired comparison matrices (a relative importance between items (elements and sub-criteria) given by decision-makers in pairs; set up paired comparison matrices. Simply stated, the fuzzy AHP substitutes the ~ij ), specific figure for aij with triangular fuzzy numbers (a implying that triangular fuzzy numbers are substituted into the pair-wise comparison matrix as Eq. (5).

~ij ; a ¼ ½a

~ij ¼ ðLij ; M ij ; Rij Þ; a

~ij ¼ 1=a ~ji ; a

Finally, the authors found 18 questionnaires to be valid, meaning that if aij of a questionnaire was consistent, then ~ij were also consistent. its fuzzy paired comparison matrices a Hence, the author could convert these values to fuzzy numbers according to the definitions in Fig. 7 and Table 4. And fuzzy numbers of matrices were obtained by calculating the geometric mean. (4) Calculate fuzzy weight value: In this study, the three positive and negative value matrices were calculated using the ‘‘Column Vector Geometric Mean Method” proposed by Buckley (1985), using the following equations: 1 e i ¼ ða ~i1  a ~i2  . . .  a ~in Þn ; 8i ¼ 1; 2; . . . ; n Z ei  ðZ e1  Z e2  . . . Z e n Þ1 ~i ¼ Z w

ð8Þ ð9Þ

where, aij: Column i row j of matrix, i; j = 1, 2, . . . , n. e i : Column vector mean value of fuzzy number, Z i = 1, 2, . . . , n. ~ i : weight of ith factor. w : multiply fuzzy numbers, for example, assuming two trie ¼ ða1; b1; c1Þ; B e ¼ ða2; b2; c2Þ. angular fuzzy numbers: A

eB e ¼ ða1; b1; c1Þ  ða2; b2; c2Þ A ¼ ða1  a2; b1  b2; c1  c2Þ: /: divide fuzzy numbers, e. g. two triangular fuzzy numbers e and B, e A

8i; j ¼ 1; 2; . . . ; n

e B e ¼ ða1; b1; c1Þ/ða2; b2; c2Þ ¼ ða1=c2; b1=b2; c1=a2Þ: A/

ð5Þ The authors removed four invalid questionnaires and analyzed 20 available questionnaires (Table 1). Then, following the AHP approach, a consistency index (CI) of 0.1 or less was considered acceptable, reflecting an informed judgment that could be attributed to the knowledge of the analyst. Hence the decision-maker’s pair-wise comparison matrices were acceptable. Therefore, the authors selected the valid sample (i.e. CI < 0.1, by Eq. (6) and (7)) among the questionnaires through AHP software (Choice-Maker 2002 and Microsoft Excel 2007).

ð7Þ

(5) Defuzzification: Various defuzzication methods are available, and methods used in defuzzified fuzzy ranking generally include the mean of maximal (MOM), center of area (COA), and a-cut. Utilizing the COA method to find out the Best Nonfuzzy Performance value (BNP) is simple and practical without the need to bring in the preferences of any evaluators (Hsieh, Lu, & Tzeng, 2004; Opricovic & Tzeng, 2003). Hence, the COA method is used in this research. The BNP value of the fuzzy number can be found as following.

Table 5 Fuzzy weights of elements with respect to the overall objective.

E1 E2 E3 E4 E5 E6

Average fuzzy value

Fuzzy weights

BNP

STD_BNP

Rank

(1.3901,1.6903,1.9130) (0.6751,0.7486,0.9026) (0.8119,0.9416,1.1146) (1.1279,1.1675,1.6297) (0.5910,0.6733,0.9676) (0.5854,0.6948,0.8332)

(0.1889,0.2857,0.3692) (0.0917,0.1265,0.1742) (0.1103,0.1592,0.2151) (0.1532,0.1973,0.3145) (0.0803,0.1138,0.1867) (0.0795,0.1174,0.1608)

0.2813 0.1308 0.1615 0.2217 0.1269 0.1193

0.2700 0.1256 0.1551 0.2129 0.1219 0.1145

1 4 3 2 5 6

Table 6 Fuzzy weights of sub-criteria with respect to market segment.

C1 C2 C3

Average fuzzy value

Fuzzy weights

BNP

STD

Rank

(1.0029, 1.1842, 1.3782) (1.2681, 1.5120, 1.7636) (0.4830, 0.5585, 0.6698)

(0.2631, 0.3638, 0.5005) (0.3327, 0.4646, 0.6404) (0.1267, 0.1716, 0.2432)

0.3758 0.4792 0.1805

0.3629 0.4628 0.1743

2 1 3

7401

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407

BNP ¼ ½ðRij  Lij Þ þ ðM ij  Lij Þ=3 þ Lij

w Lw ~ i of the market segment is Ew ing the w 1 ; E1 ¼ ðE1 ; Mw Rw E1 ; E1 Þ,

ð10Þ

(6) Normalization: The STD_BNP value of the fuzzy number can be found as following.

e ij = STD BNP ¼ NW i ¼ DF A

X

e ij DF A

L1 R1 R2 R3 R4 R5 R6 1 ELw 1 ¼ E1  ðE1 þ E2 þ E3 þ E4 þ E5 þ E6 Þ

¼ 1:3900  ð1:9130 þ 0:9026 þ 1:1146 þ 1:6297 þ 0:9676 þ 0:8332Þ1 ¼ 0:188852  1 M1 M2 M3 M4 M5 M6 ¼ EM1 EMw 1 1  E1 þ E2 þ E3 þ E4 þ E5 þ E6

ð11Þ

¼ 1:6903  ð1:6903 þ 0:7486 þ 0:9416 þ 1:1675 þ 0:6733 þ 0:6948Þ1 ¼ 0:285719  1 Rw L1 L2 L3 L4 9m5 þ EL6 E1 ¼ ER1 1  E1 þ E2 þ E3 þ E4 þ E5 6

(7) Compute the overall hierarchy weight: After the weights for various hierarchies and elements are computed, computation results for the overall hierarchy weight are compiled. According to the FAHP procedure above, the authors calculated weight values of elements and sub-criteria as listed in Appendix D, and Table 5–11. The results are illustrated by the following statements: (1) Appendix D shows the fuzzy comparison matrix of elements with respect to the overall objective, in which triangular ~ij ) are obtained by Eq. (8). fuzzy numbers (a (2) Table 5 presents the results of the fuzzy weights of elements with respect to the overall objective, in which the summary triangular fuzzy numbers are obtained by Eq. (9). e.g. assum-

¼ 1:9130  ð1:3901 þ 0:6751 þ 0:8119 þ 1:1279 þ 0:5910 þ 0:5854Þ1 ¼ 0:369208   Mw Rw Ew ¼ ELw 1 1 ; E1 ; E1 ¼ ð0:18885; 0:28572; 0:36921Þ

Next, the fuzzy weight value of elements are defuzzified by Eq. (10), e.g. assuming the weight values of ‘‘market segment (E1)” is (E1LDw, E1MDw, E1RDw):

E1 ¼

    Lw þ EMw =3 þ ELw ERw  ELw 1  E1 1 1 1 Þ

¼ ðð0:36921  0:18885Þ þ ð0:28572  0:18885ÞÞ=3 þ 0:18885Þ ¼ 0:28126 0:2813

Table 7 Fuzzy weights of sub-criteria with respect to strategic alliance.

C4 C5

Average fuzzy value

Fuzzy weights

BNP

STD

Rank

(1.3430, 1.4799, 1.5909) (0.6286, 0.6757, 0.7446)

(0.5751, 0.6865, 0.8069) (0.2692, 0.3135, 0.3777)

0.6895 0.3201

0.6829 0.3171

1 2

Average fuzzy value

Fuzzy weights

BNP

STD

Rank

(0.9838, 1.1307, 1.2648) (0.7907, 0.8844, 1.0165)

(0.4313, 0.5611, 0.7128) (0.3466, 0.4389, 0.5728)

0.5684 0.4528

0.5566 0.4434

1 2

Average fuzzy value

Fuzzy weights

BNP

STD

Rank

(0.6870, 0.7484, 0.8366) (1.1952, 1.3362, 1.4556)

(0.2997, 0.3590, 0.4445) (0.5214, 0.6410, 0.7733)

0.3677 0.6453

0.3630 0.6370

2 1

Table 8 Fuzzy weights of sub-criteria with respect to service model.

C6 C7

Table 9 Fuzzy weights of sub-criteria with respect to product strategy.

C8 C9

Table 10 Fuzzy weights of sub-criteria with respect to channel strategy.

C10 C11

Average fuzzy value

Fuzzy weights

BNP

STD

Rank

(1.0743, 1.3312, 1.6088) (0.6255, 0.7496, 0.9189)

(0.4250, 0.6398, 0.9465) (0.2475, 0.3602, 0.5406)

0.6704 0.3828

0.6366 0.3634

1 3

Table 11 Fuzzy weights of sub-criteria with respect to revenue strategy.

C12 C13

Average fuzzy value

Fuzzy weights

BNP

STD

Rank

(1.0098, 1.1232, 1.2493) (0.8004, 0.8903, 0.9903)

(0.4509, 0.5579, 0.6902) (0.3574, 0.4421, 0.5470)

0.5663 0.4489

0.5578 0.4422

1 2

7402

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407

Table 12 Fuzzy weights of element/sub-criteria by FAHP. Item

Local weights

Overall weights

BNP

STD

Rank

E1 C01 C02 C03

(0.1889, (0.2631, (0.3327, (0.1267,

0.2857, 0.3692) 0.3638, 0.5005) 0.4646, 0.6404) 0.1716, 0.2432)

(0.0497, 0.1040, 0.1848) (0.0628, 0.1327, 0.2364) (0.0239, 0.0490, 0.0898)

0.2813 0.1128 0.1440 0.0543

0.2700 0.0996 0.1272 0.0479

(1) 3 2 11

E2 C04 C05

(0.0917, 0.1265, 0.1742) (0.5751, 0.6865, 0.8069) (0.2692, 0.3135, 0.3777)

(0.0527, 0.0869, 0.1406) (0.0247, 0.0397, 0.0658)

0.1308 0.0934 0.0434

0.1256 0.0825 0.0383

(4) 6 13

E3 C06 C07

(0.1103, 0.1592, 0.2151) (0.4313, 0.5611, 0.7128) (0.3466, 0.4389, 0.5728)

(0.0476, 0.0893, 0.1533) (0.0382, 0.0699, 0.1232)

0.1615 0.0967 0.0771

0.1551 0.0854 0.0681

(3) 4 8

E4 C08 C09

(0.1532, 0.1973, 0.3145) (0.2997, 0.3590, 0.4445) (0.5214, 0.6410, 0.7733)

(0.0459, 0.0708, 0.1398) (0.0799, 0.1265, 0.2432)

0.2217 0.0855 0.1499

0.2129 0.0755 0.1323

(2) 7 1

E5 C10 C11

(0.0803, 0.1138, 0.1867) (0.4250, 0.6398, 0.9465) (0.2475, 0.3602, 0.5406)

(0.0341, 0.0728, 0.1767) (0.0199, 0.0410, 0.1010)

0.1269 0.0642 0.0366

0.1219 0.0835 0.0476

(5) 5 12

E6 C12 C13

(0.0795, 0.1174, 0.1608) (0.4509, 0.5579, 0.6902) (0.3574, 0.4421, 0.5470)

(0.0359, 0.0655, 0.1110) (0.0284, 0.0519, 0.0880)

0.1193 0.0708 0.0561

0.1145 0.0625 0.0495

(6) 9 10

(3) Tables 6–11 are the fuzzy comparison matrices of the sub-criteria with respect to sub-criterion (C1 to C13), and weight values of sub-criteria obtained by Eq. (8)– (11). Finally, this study obtained the overall weights as summarized in Table 12. The results show that the ranking of the weights of the elements are: market segment (0.2700), product strategy (0.2129), service model (0.1551), strategy alliance (0.1256), distribution/ channel strategy (0.1219), and revenue strategy (0.1145). These results manifest the two most influential elements for information service firm to develop international market are market segment and product strategy. This means the company should strengthen these two areas, including cognition of market scale, cognition of regional culture, and competition of product. The least influential element is revenue strategy. That is, the optimal outcome depends on the other five elements. Among these elements, there are three (i.e. product competition, market segment, and service model) whose weights are over 50% in all elements. On the other hand, the results show that the ranking of the weights of sub-criteria are: competition of product (C9), cognition of market scale (C2), cognition of regional culture (C1), integrity of service mechanism (C6), distribution channel Model (C10), positioning the cooperative role (C4), product function and quality (C8), depth of knowledge of industrial domain (C7), revenue model (C12), pricing strategy (C13), market’s geographical position (C3), distribution channel efficiency (C11), and the degree of cooperation (C5). The result shows the top five sub-criteria (sub-criteria having weights over 50%) and their weights of global ranking are: competition of product (0.1323), cognition of market scale (0.1272), cognition of regional culture (0.0996), integrity of service mechanism (0.0854), and distribution channel model (0.0835). These results indicate that the sub-criteria are fully distributed over product strategy, market segment and distribution/channel strategy, and this distribution is more in accordance with the elements ranking results.

BEST-Program as a business benchmark to analyze their strategic advantages in these six elements. In order to evaluate performance for Taiwan’s information service industry in developing global markets, the authors collected data on eight flagship-projects (Appendix C) from 2006 to 2008. They collected each case’s internal documents (i.e. projects) and downloaded others’ data from public information websites; the authors also conducted in-depth interviews with over 11 CEOs and several consultants. Finally, this study found performance of each of the eight cases as listed in Table 13. According to the performance list by indicators (Table 13), cases 1 and 7 (denoted by *) executed better on the six indicators. Overall, as they expanded their international markets, cases 1 and 7 employed the following priorities: (1) building up strategic alliances and identifying the complementarity of core competencies, (2) deciding product line and product package, (3) analyzing the factors of geography and culture in different target markets, then adopting a proper alliance model, (4) developing a mode of viable product and service output, (5) establishing a customer-oriented service model, (6) setting up different channels of distribution, and (7) defining the allocation of benefits for each partner. Because of space limitations, the following discussion focuses on findings of case 1 (benchmark) that relate specifically to the strategy in

Table 13 Performance of cases (flagship-projects). Case

1* 2 3 4 5 6 7* 8

4. Case survey: performance and strategy This study has proposed the operative framework (Fig. 4), six performance indicators (Table 2), and weights of six elements and 13 sub-criteria (Fig. 6). Furthermore, the author selected the

KPI NMS (%)

PC (%)

SSOP

PSC

NDC (%)

TRA (%)

300 110 140 170 190 130 160 150

62 55 41 52 38 60 68 48

A B B C A A A B

A A B A B B A B

360 33 120 250 100 80 300 230

136 –42 20 43 14 5 63 16

Note: NMS, NDC, and TRA – the growth rate from 2007 to 2008 (two years). PC – the proportion of partner scale (revenue) in alliance. SSOP – A (complete SOP), B (incomplete SOP), C (no service SOP). PSC – A (better complementary), B (some complementary), C (noncomplementary).

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407

the six business elements, and suggestions for case 1’s future development. (1) Market segmentation (E1): Mainland China and Southeast Asia were the main target markets. Case 1 crossed retail and channel sector markets based on the medium-sized ERP market in the Chinese manufacturing domain. They also expanded the scope of business and sales. Branches (including commercial offices) formed 25 bases of operation, covering 80% of the mainland Chinese ERP market. The dealer model was employed for the SME (small and medium enterprise) market. Next actions will include EC, BI, and CRM-based solutions, while also supporting the ERP II market and related demands. Better business models should be drawn for balancing the cost between ‘‘time” and ‘‘distance” in broader Asian markets. In other words, two issues should be enhanced, namely ‘‘market scale (C1)” and ‘‘market’s geographical position (C3)”. (2) Strategic alliances (E2): The principle criterion for partner selection depends on complementarity for Enterprise Resource Planning (ERP) solution (such as product maturity). A total of five partners were selected by the major firm. In addition, strategic alliances were made on three levels (the higher the level, the closer the cooperative relationship). Although they have succeeded in extending to mainland China or Vietnam, they need to plan different alliance models more focused on ‘‘the degree of cooperation (C5)” to fulfill various worldwide markets (customers’) needs. (3) Service model (E3): Case 1 used ‘‘structuralized solutions” and ‘‘uniformized service mechanisms” to promote the global market, and the case’s quantification duplication mechanism counseling partner caused uniform service quality. As far as the superior product and service of partners, the major firm also took advantage of this opportunity to introduce internal use, e.g. KM and eMarketing system. Thus, for the new tendency in applications or service model, such as SaaS (i.e., software as a service), they have to consider the impact on their traditional service. (4) Product strategy (E4): The early stage focused on expanding ERP to ERPII of major firms; the middle stage focused on portfolio architecture to enhance strategic alliances in opening overseas market share; and the final stage, shedding home culture to adapt business operations to the local culture, focusing on collaboration with local firms. On the other hand, based on a new technology development, such as ‘‘cloudy technology”, there should be more research on Case 1 about new product for this tendency to hold their competition. (5) Distribution/channel model (E5): Case 1 focused on the manufacturing market to strengthen the maturity of products, and they achieved market share of 40% of manufacturing software applications through a direct sales model. Case 1 used direct and dealer models to enter the mainland Chinese market, and sold more than 700 software packages. Case 1’s project cases include several ERP-family products to satisfy customer needs (one-stop-shopping). Especially, since Case 1 has closely cooperated with a bigger channel brain to extent China market. This successful strategy is an essential key decision for developing global market.

7403

(6) Revenue logic (E6): The main source of revenue for case 1 came from licensing fees, as well as other income sources, such as consultancy and services. Overall, in order to grasp global customer needs and raise ERP market penetration, the firms of ‘‘case 1” should enhance their strategic alliances to fulfill a broader market and build a highly value-added service model through an integrated CRM/BI system. The authors also suggest case 1 could clarify the various pricing models in its revenue strategy for different markets to achieve their business objectives.

5. Conclusion and future research This research has combined qualitative research and quantitative analysis to propose a majority of points for information service firms planning strategy for expanding international market share by a hybrid fuzzy-Delphi-analytic hierarchy process approach (FDAHP), in which the new framework consists of the six business elements and 13 sub-criteria, including: market segment, including three criteria (cognition of regional culture, cognition of market scale, market’s geographical position), strategic alliance, including two criteria (positioning the cooperative role, the degree of cooperation), service model (integrity of service mechanism, the depth of knowledge of industrial domain), product strategy (product function and quality, competition of product), distribution/channel strategy (distribution channel model, distribution channel efficiency), revenue strategy (revenue model, pricing strategy). This research also suggests six indictors for BI systems for alliances developing global markets, including: growth rate of number in foreign markets (GNM), proportion of partners scale (PS), degree of service standard operation process (DSOP), proportion of alliance product complementarity (PAPC), number of channels (NCC), and growth rate of alliance revenue (GAR). Clearly, the proposed model, focusing on internationalization, business strategy, and alliance aspect, is a new framework in which it is more meticulous and workable structure than former proposed model, combing key elements mentioned previously (such as Chen & Wang, 2010; Rajala et al., 2003). All in all, the findings of this research can make a tremendous contribution to the information service industry in planning operative strategies for developing target markets all over the world. In addition, this study verifies the practicability for the proposed model through case study; it appears this model can fit the analytic needs in developing overseas markets for information service firms. However, according to the analysis above, the authors foresee the new technology or service type (such as SaaS, ‘‘cloudy” technology) will impact business strategy. Therefore, the researcher needs to take in others related business issue. For the future, the author hope more sound research will be performed to explore this issue, and suggest a focus on correlative research, including: enhanced evaluation of performance of firms, study on various domains of information services, various benchmark cases-study, and proposals for suitable entry models for different markets.

Acknowledgements The authors thank referees for their useful comments. The authors are so grateful to get valuable aspects from CEOs, experts, and project leaders who work for the famous information service corporations and the ministries of government in Taiwan.

7404

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407

Appendix A. Summary of the elements/sub-criteria of information service firms in developing global markets (comparison table)

Chen and Wang (2010) proposed (by cross case study)

This study proposed (by FMDM)

Elements

Elements

Market segment

Strategic alliance

Service/ implementation model*

Sub-Criteria













Competition of product

Distribution/channel model











Revenue efficiency



Cognition of regional culture Cognition of market scale Market’s geographical position Positioning the cooperative role Effects of partner supplement Degree of cooperation Integrity of service mechanism Talented person for internationalized service Image and appraisal in the market The depth of knowledge of industrial domain Service pricing model Function and R&D of product Quality and image of product, product’s capacity in market Product pricing model Efficiency of distribution/channel model Distribution/channel cost controls Distribution/channel benefit assignment Strategy of revenue Rate of revenue and expenditure (or profit and loss)

Total: 20 *

Is renamed term by fuzzy-Delphi approach.

Market segment (E1)

Strategic alliance (E2)

Sub-criteria







Cognition of regional culture (C1) Cognition of market scale (C2) Market’s geographical position (C3) Positioning the cooperative role (C4) Degree of cooperation (C5)

Service model* (E3)

Integrity of service mechanism (C6)

Depth of knowledge of industrial domain* (C7)

Product strategy (E4)

Product function and quality (C8)

Competition of product* (C9)

Distribution/channel strategy* (E5)

Distribution channel model* (C10)

Distribution channel efficiency (C11)

Revenue strategy* (E6)

Revenue model* (C12)

Pricing strategy* (C13) Total: 13

Appendix B. The key performance indicator of BI by modified Delphi method (MDM) (4 rounds; (): Number of votes by experts) Round

Elements Strategic alliance

1st Round

3rd Round

4th Round

Market segment

Service model

Distribution/channel strategy

Revenue strategy

– Partner scale (4) – Market share of partner (3) – Market competition (2) – Number of branches (1) – Degree of willingness to cooperate (1) – Positioning the cooperative role(1) – Expertise (1) – Partner scale (8) – Market share of partners (2) – Positioning the cooperative role (1) – Market competition (1)

– Product/service complement (7) – Product Integrity (3) – Product Expansion(2)

– Number of markets/ subsidiaries (3) – Overall market share (3) – Market profitability (2) – Market positioning (2) – The extent of the Market (2)

– Integrity of service mechanism (or service SOP) (6) – Customer satisfaction (1) – Growth rate contract (2) – Business integration (2) – Value-added of service (1)

– Number of distribution/ channel (5) – Function/role of channel (3) – Growth rate of revenue (2) – Costs (1) – Integrity management process (1)

– Total revenue of alliance (6) – Gross margin (3) – Sources of revenue (1) – KPI (1) – Number of customers (1)

– Product/service complement (8) – Product Integrity (4)

– Number of markets/ subsidiaries (7) – Overall market share (3) – Market profitability (2)

– Integrity of service mechanism (or service SOP) (8) – Customer satisfaction (1) – Growth rate contract (3)

– Total revenue of alliance (8) – Gross margin (2) – Number of customers (2)

– Partner scale (10) – Market share of partners (1) – Market competition (1)

– Product/service complement (9) – Product Integrity (3)

– The number of markets/subsidiaries (9) – Overall market share (2) profitability – Market (1) The number of market/ subsidiary: NMS (12)

– Integrity of service mechanism (or service SOP) (10) – Customer satisfaction (1) – Growth rate contract (1)

– Number of distribution/ channel (6) – Function/role of channel (2) – Growth rate of revenue in channel (2) – Integrity management process (2) – Number of distribution/ channel (10) – Growth rate of revenue in channel (1) – Function/role of channel (1) The number of distribution/channel: NDC (12)

Partner scale: PC (12)

Product/service complement: PSC (12)

Integrity of service mechanism(or service SOP): SSOP (12)

– Total revenue of alliance (9) – Gross margin (1) – Number of customers (2)

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407

2nd Round

Product strategy

Total revenue of alliance: TRA(12)

(): Number of consensus from experts.

7405

7406

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407

Appendix C. Summary of cases (‘‘Flagship-project participating in Taiwan’s BEST-Program) Item Project

Employees/customer type/target market

Number of partner (service contents)

Case 1 Case 2

1044/manufacturing (SME)/mainland China, Southeast Asia 515/bank and securities/mainland China, Southeast Asia

Case 3 Case 4 Case 5

1017/ICT,SI, Government, Traffic/Asia, Middle East, Europe 414/TFT-LCD, traditional/manufacture, Government 290/remote backup, security monitoring, hosting, network security solutions 1700/bank and securities, financial, manufacturing and logistics/ mainland China, Southeast Asia 295/bank and clearing, 7-11 retailer, expense finance/Thai, Hong Kong, mainland China 226/petroleum and chemical, steel, semiconductor/mainland China

5 (GSCM, logistics, e-invoice, KM, e-Learning) 4 (data mining, domain know-how, consultant, elearning) 3 (PKI,KM, search engine) 5 (HR, Web report, ERP, PLM, e-Learning) 4 (security, network, disaster, SLM, data mining)

Case 6 Case 7 Case 8

4 (SI, KM,CRM,BI) 3 (SI, reader card machine, card) 4 (network communication solution, KM, expert system platform, MES)

Appendix D. Fuzzy comparison matrix of elements with respect to overall objective by FAHP E1

E2

E3

E4

E5

E1

(1, 1, 1)

E2

(0.2979, 0.4387) (0.4087, 0.6700) (0.5564, 0.9585) (0.3823, 0.5848) (0.4248, 0.6466)

(2.2797, 2.8262, 3.3567) (1, 1, 1)

(1.2711, 1.6528, 2.1313) (0.6187, 0.7793, 0.9962) (1, 1, 1)

(0.8669, 1.1676, 1.4878) (0.4825, 0.6039, 0.8002) (0.5480, 0.7403, 0.9962) (1, 1, 1)

(1.3791, 2.1098) (1.0059, 1.7242) (1.0911, 1.8610) (2.2384, 3.4798) (1, 1, 1)

E3 E4 E5 E6

0.3538, 0.5123, 0.7090, 0.4723, 0.5081,

(0.8414 1.0730, 1.3547) (1.0079, 1.6145, 1.6715) (0.6027, 0.7511, 1.0102) (0.6257, 0.7867, 0.9868)

(0.8975, 0.7682, 1.0393) (0.5584, 0.7203, 0.9313) (0.5757, 0.7398, 0.9701)

References Basligil, H. (2005). The fuzzy analytic hierarchy process for software selection problems. Journal of Engineering and Natural Sciences, 2, 24–33. Buckley, J. J. (1985). Fuzzy hierarchical analysis. Fuzzy Sets and Systems, 17, 233–247. Bell, J. (1997). A comparative study of the export problems of small computer software exporters in Finland, Ireland and Norway. International Business Review, 6(6), 585–604. Bieberstein, N., Bose, S., Walker, L., & Lynch, A. (2005). Impact of service-oriented architecture on enterprise systems, organizational structures and individuals. IBM Systems Journal, 44(4), 691–709. Bonaccorsi, A., Giannangeli, S., & Rossi, C. (2006). Entry strategies under competing standards: Hybrid business models in the open source software industry. Management Science, 52(7), 1085–1098. Barcus, A., & Montibeller, G. (2008). Supporting the allocation of software development work in distributed teams with multi-criteria decision analysis. Omega, 36(3), 464–475. Burgel, O., & Murray, G. C. (2000). The international market entry choice of start-up companies in high-technology industries. Journal of International Marketing, 8(2), 33–62. Chang, D. Y. (1996). Applications of the extent analysis method on fuzzy AHP. European Journal of Operational Research, 95, 649–655. Chan, F. T. S., & Kumar, N. (2007). Global supplier development considering risk factors using fuzzy extended AHP-based approach. Omega, 35(4), 417–431. Chen, M. K., Wang, S. C., Chen, K. H., & Wang, C. H. (2007). A study on the current status and trend of e-business IT application in Taiwan manufacturing industry. International Journal of Electronic Business Management, 5(4), 310–318. Chen, M. K., Wang, S. C., & Chio, C. H. (2009). The e-business policy of global logistics management for manufacturing. International Journal of Electronic Business Management, 7(2), 86–97. Chen, M. K., & Wang, S. C. (2010). The critical factors of success for information service industry in developing international market: Using analytic hierarchy process (AHP) approach. Expert Systems with Applications, 37(1), 694–704.

(0.2874, 0.3520, 0.4467) (0.3242, 0.4054, 0.5474)

E6 1.7057, 1.3617, 1.4200, 2.8408,

(0.6488, 0.8249, 1.0629)

(1.5465, 2.3543) (0.8451, 1.3633) (1.0309, 1.7371) (1.8270, 3.0841) (0.9408, 1.5413) (1, 1, 1)

1.9679, 1.0653, 1.3517, 2.4667, 1.2123,

Coviello, N., & Munro, H. (1997). Network relationships and the internationalization process of small software firms. International Business Review, 6(4), 361–386. Deng, H. (1999). Multicriteria analysis with fuzzy pair-wise comparison. International Journal of Approximate Reasoning, 21, 215–231. Eckerson Wayne, W. (2005). Performance dashboards: Measuring, monitoring, and managing your business. Wiley. Gable, G. G. (1994). Integrating case study and survey research methods: An example in information systems. European Journal of Information Systems, 3(2), 112–126. Ganek, A., & Kloeckner, K. (2007). An overview of IBM service management. IBM Systems Journal, 46(3), 375–386. Griffith, D. A. (2010). Understanding multi-level institutional convergence effects on international market segments and global marketing strategy. Journal of World Business, 45(1), 59–67. Hedman, J., & Kalling, T. (2003). The business model concept: Theoretical underpinnings and empirical illustrations. European Journal of Information Systems, 12, 49–59. Hoch, D., Roeding, C., Purkert, G., Lindner, S., & Müller, R. (2000). Secrets of software success: Management insights from 100 software firms around the world. Boston, MA: Harvard Business School Press. Hsieh, T. Y., Lu, S. T., & Tzeng, G. H. (2004). Fuzzy MCDM approach for planning and design tenders selection in public office buildings. International Journal of Project Management, 22(7), 573–584. Jain, S., & Kannan, P. K. (2002). Pricing of information products on online servers: Issues, models, and analysis. Management Science., 48(9), 1123–1142. Krishnan, M. S., Kriebel, C. H., Kekre, S., & Mukhopadhyay, T. (2000). An empirical analysis of productivity and quality in software products. Management Science, 46(6), 745–759. Kahraman, C., Cebeci, U., & Ulukan, Z. (2003). Multi-criteria supplier selection using fuzzy AHP. Logistics Information Management, 16(6), 382–394. Krishnamurthy, S. (2003). E-commerce management, text and cases, Thompson learning. New Yark: South-Western College Pub.

M.-K. Chen, S.C. Wang / Expert Systems with Applications 37 (2010) 7394–7407 Laarhoven, P. J. M., & Pedrycz, W. (1983). A fuzzy extension of Saaty’s priority theory. Fuzzy Sets and Systems, 11(3), 229–241. Lee, Y. S. (2008). Using modified Delphi method to explore the competition strategy for software companies of Taiwan. Journal of Informatics and Electronics, 3(1), 39–50. Leidecker, J. K., & Bruno, A. V. (1984). Identifying and using critical success factors. Long Range Planning, 17(1), 23–24. Magretta, J. (2002). Why business models matter. Harvard Business Review, 80(5), 86–92. Murry, J. W., & Hommons, J. O. (1995). Delphi: A versatile methodology for conducting qualitative research. The Review of Higher Education, 18(4), 423–436. McNaughton, R. B. (1996). Foreign market channel integration decisions of Canadian computer software firms. International Business Review, 5(1), 23–52. Nambisan, S. (2001). Why service business are not product businesses. MIT Sloan Management Review, 42(4), 72–80. Ojala, A., & Tyrväinen, P. (2006). Business models and market entry mode choice of small software firms. Journal of International Entrepreneurship, 4(2), 69–81. Ojala, A., & Tyrväinen, P. (2007). Market entry and priority of small and mediumsized enterprises in the software industry: An empirical analysis of cultural distance, geographic distance, and market size. Journal of International Marketing, 15(3), 23–149. Opricovic, S., & Tzeng, G. H. (2003). Defuzzification within a fuzzy multicriteria decision model. International Journal of Uncertainty. Fuzziness and Knowledgebased Systems, 11(5), 635–652. Rajala, R., Rossi, M., & Tuunainen, V. K. (2003). Software vendor’s business model dynamics case: TradeSys. Annals of Cases on Information Technology, 5(1), 538–548. Rajala, R., & Westerlund, M. (2007a). A business model perspective on knowledgeintensive services in the software industry. International Journal of Technoentrepreneurship, 1(1), 1–19. Rajala, R., & Westerlund, M. (2007b). Business models – a new perspective on firms’ assets and capabilities, observations from the Finnish software industry. Entrepreneurship and Innovation, 8(2), 115–125. Ruokonen, M. (2008). Market orientation and product strategies in small internationalising software companies. Journal of High Technology Management Research, 18(2), 143–156.

7407

Shapiro, C., & Varian, H. R. (1999). Information rules: A strategic guide to the network economy. Boston, MA: Harvard Business School Press. Saaty, T. L. (1980). The analytic hierarchy process. New York: McGraw-Hill. Saaty, T. L. (1994). How to make a decision: The analytic decision processes. Interfaces, 24(6), 19–43. Sainio, L. M., & Marjakoski, E. (2009). The logic of revenue logic? Strategic and operational levels of pricing in the context of software business. Technovation, 29(5), 368–378. Tiwana, A. (2004). An empirical study of the effect of knowledge integration on software development performance. Information and Software Technology, 46(13), 899–906. Vaidya, O. S., & Kumar, S. (2006). Analytic hierarchy process: An overview of applications. European Journal of Operational Research, 169, 1–29. Wang, Zhuo. (2005). Business intelligence. Taiwan: DrMaster Culture Limited Company. Westerlund, M., Rajala, R., & Svahn, S. (2007). Managing networked business models in the software industry. The Business Review. Cambridge, 7(1), 53–57. Wild, J. J., Wild, K. L., & Han, J. C. Y. (2008). International business – the challenges of globalization (4th ed.). Upper Saddle River, New Jersey: Person Education. Willcocks, L., & Choi, C. J. (1995). Co-operative partnership and ‘total’ IT outsourcing: From contractual obligation to strategic alliance? European Management Journal, 13(1), 67–78. Winkler, J. K., Dibbern, J., & Heinzl, A. (2008). The impact of cultural differences in offshore outsourcing: Case study results from German–Indian application development projects. Information Systems Frontiers, 10(2), 243–259. Yin, R. K. (1994). Case study research: Design and methods. California: SAGE Publications.

Web references Industrial Development Bureau of Ministry of Economic Affairs (MOEAIDB), Taiwan. September (2004), the BEST (Best E-Services Taiwan) export plan. Available from http://www.timglobe.com.tw/. Market Intelligence & Consulting Institute (MIC), III, Taiwan. (2009). Information service industry news. Available from http://mic.iii.org.tw/intelligence/.