An analytical network process-based framework for successful total quality management (TQM): An assessment of Turkish manufacturing industry readiness

ARTICLE IN PRESS Int. J. Production Economics 105 (2007) 79 96 www.elsevier.com/locate/ijpe An analytical network process-based framework for successful total quality management (TQM): An assessment of Turkish manufacturing industry readiness Ozden Bayazit a,, Birsen Karpak b a Department of Business Administration, College of Business, Central Washington University, 20000 68th Avenue W, Lynnwood, WA 98036, USA b Department of Management, Youngstown State University, Youngstown, OH 44555, USA Received 3 February 2004; accepted 19 December 2005 Available online 9 May 2006 Abstract In this study, we have developed an analytic network process (ANP)-based framework to identify the level of impact of different factors on total quality management (TQM) implementation and to assess the readiness of the Turkish manufacturing industry to adopt TQM practices. ANP is a methodology recently introduced by Saaty for multiple criteria problems where there is feedback and interdependence among decision attributes and alternatives. We determined the factors that affect the level of implementation of TQM by doing literature searches and further refined those factors through a survey conducted among 250 large manufacturing companies in Turkey. We ended up with 32 factors. When we applied the model into large manufacturing companies zero defect and costly and long-term study turned out to be the most influential factors contrary to those of survey respondents quality improvement and higher revenue. The results of our decision model show that the Turkish manufacturing industry has a readiness level of 59.2% for implementing TQM. Model identifies a number of factors for successful application; therefore, an understanding of the critical factors would help managers to advance TQM implementation. Since there is feedback and interdependence among these factors, ANP proves to be an effective framework for assessing readiness to adopt TQM and facilitating TQM implementation. r 2006 Elsevier B.V. All rights reserved. Keywords: Factors affecting total quality management implementation; Multiple criteria analysis; Quality management; Analytic network process; Decision analysis 1. Introduction In the early 1980s, consumers became more powerful and started to demand high-quality goods and services at reasonable prices. The globalization Corresponding author. Tel.: +1 425 640 1574; fax: +1 425 640 1488. E-mail addresses: bayazito@cwu.edu (O. Bayazit), bkarpak@ysu.edu (B. Karpak). of trade has made high-quality low-cost products available throughout the world. These factors increased the pressure on companies around the world to improve their goods and services. Technologies and methodologies such as total quality management (TQM) have helped them do this (Wadsworth et al., 2002). In Turkey, manufacturing organizations represent a dynamic and important sector of the economy and they are aware of the importance to their survival of assuring quality in 0925-5273/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ijpe.2005.12.009

80 ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 their products. A considerable number of organizations have tried to implement these practices and have failed to achieve much, while many others have implemented TQM with great success. The overwhelming volume of literature on TQM is primarily focused on the elements of TQM and the approaches taken to assure a successful implementation; however, less attention has been devoted to identify the critical success factors for the implementation of TQM program (Dayton, 2001). Black and Porter (1996) developed a model that identifies a set of critical factors of TQM, their relative importance and the interrelationships between each. Recently Conca et al. (2004) conducted a similar study to identify critical success factors of TQM and empirically tested with the answers of 108 ISO certified firms in Spain. Since these critical factors are interdependent and there is feedback among them we contend that our analytic network process (ANP)-based framework is an enhancement to earlier studies. Since the critical success factors of TQM have not been studied extensively throughout the world, it is the intention of this study to investigate these factors and identify the relative importance of each of them in a successful TQM implementation and measure the readiness of the Turkish manufacturing industry to adopt it. The approach in this paper is to use the ANP to investigate the degree to which TQM practices were adopted in the Turkish manufacturing industry and to identify the impact of different factors on successful TQM implementation. This industry is particularly appropriate for the study of the effectiveness of TQM program implementation since the majority of organizations that have implemented TQM consist of manufacturing companies in Turkey. ANP requires expert judgments to assess the relative importance of different factors with respect to each other. In our study these expert judgments were obtained via survey of 250 manufacturing companies in Turkey. ANP is a new methodology introduced by Saaty (2001b) that extends the analytic hierarchy process (AHP) for decision making to cases of dependence and feedback. As Wang et al. (2004) pointed out, more and more researchers are realizing that AHP is an important generic method and are applying it. Whereas, ANP is relatively new and there are few applications as of yet. Some examples of ANP applications include re-engineering, supply chain performance, logistics, quality function deployment, energy policy planning, project selection decisions, and performance measurement systems (Hamalainen and Seppalainen, 1986; Partovi and Corredoira, 2002; Sarkis and Talluri, 2002; Agarwal and Shankar, 2002; Partovi, 2001; Lee and Kim, 2000; Ashayeri et al., 1998; Meade and Sarkis, 1998; Sarkis, 1998, 1999, 2003; Karpak and Bayazit, 2001; Saaty, 2001a, c). In this paper we developed a framework that facilitates finding the importance of different factors on TQM implementation. In addition, we applied ANP for the first time to assess the readiness of manufacturing industry in Turkey to adopt TQM based on the survey of 62 companies. Since constructing such a framework can best be approached by studying organizations that have implemented TQM, we have excluded the ones which stated that they did not implement TQM. The paper is organized into five sections and begins with a literature search for the factors affecting TQM implementation. The methodology of the study is explained in Section 3. Section 4 introduces an ANP-based framework which identifies the importance of different factors on TQM implementation and Turkish manufacturing industry readiness to implement TQM. The overall conclusion is given in Section 5. 2. Background There is a huge amount of published literature on TQM. A dominant theme in these writings is that TQM is an approach to management that is characterized by the principles of customer focus, continuous improvement, and teamwork (Ugboro and Obeng, 2000; Wadsworth et al., 2002; Chan and Quazi, 2002; Hellsten and Klefsjo, 2000; Scharitzer and Korunka, 2000; Young et al., 2001; Woon, 2000; Fok et al., 2001). It is broadly agreed that TQM is an integrated management philosophy aimed at continuously improving the performance of products, processes, and services to achieve and surpass customer expectations. A number of research studies have been carried out to examine the implementation process of TQM and investigate the critical success factors for implementing TQM. A common conclusion of these studies is that the way TQM is implemented is central to its long-term success within an organization (Ghobadian and Gallear, 2001). Flynn et al. (1995) surveyed 42 US manufacturing firms and measured the degree of use of quality management practices. Constructs they used were top

ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 81 management support, customer relationship, workforce management, work attitudes, process flow management, statistical process control/feedback and product design. They found out that different core quality management practices led to success in different dimensions of quality. Allen and Kilmann (2001) reported that using a cross-functional planning approach when developing strategic plans, forming quality councils and teams, and customer focus are the important TQM practices. Black and Porter (1996) conducted a study to determine the TQM critical success factors using members of the European Foundation for Quality Management. They determined that TQM critical success factors were: people and customer management, supplier partnerships, communication of improvement information, customer satisfaction orientation, external interface management, strategic quality management, teamwork structures for improvement, operational quality planning, quality improvement measurement systems, and corporate quality culture. Dayton (2001) determined that all of the TQM critical success factors identified in the Black and Porter (1996) study were important to US quality assurance professionals as well. Research conducted by Tsang and Antony (2001) showed the 11 critical success factors for the successful implementation of TQM in the UK service sector are: customer focus, continuous improvement, teamwork and involvement, top management commitment and recognition, training and development, quality systems and policies, supervisory leadership, communication within the company, supplier relationship and supplier management, measurement and feedback, and cultural change in employees behaviors and attitudes. Motwani (2001) found seven critical success factors for TQM implementation after examining six empirical studies. He recommended that attention should be given mostly to these five constructs: top management commitment, employee training and empowerment, quality measurement and benchmarking, process management, and customer involvement and satisfaction. Ugboro and Obeng (2000) conducted research among 800 members of the Association for Quality and Participation. According to their study, top management leadership and commitment, teamwork, flow of information within the organization, employee involvement and empowerment are the critical strategies for successful TQM programs. Ghobadian and Gallear (2001) examined 31 TQM implementation plans and identified the most common initiatives: training, TQM education course, teamwork, creating quality council/steering group, quality assurance processes, and mission/ vision development. Laszlo (1999) concluded that a successful implementation of a quality management approach within any organization requires commitment, corporate culture, and investment. The critical factors in TQM found in the literature vary from one author to another, although there is a common core formed by the following requirements (Conca et al., 2004): customer-based approach, management commitment and leadership, quality planning, management based on facts, continuous improvement, human resource management (involvement of all members in the firm, training work teams, and communication systems that eliminate communication barriers), learning, process management, and cooperation with suppliers. Several studies that have been devoted to examining the implementation process of TQM have emphasized goals and results. Reed et al. (2000) concluded that TQM has the potential to generate competitive advantage. They claimed that generating competitive advantage depends on not only on TQM but also on the fit between the strategy, firm orientation, and the environment. Today a growing number of organizations implement TQM to generate a competitive advantage (Nilsson et al., 2001; Chan and Quazi, 2002). A study by Chong (1998) argued that TQM might provide a fundamental way of conducting business, making the organization more competitive and viable, with TQM driving change and improvement. Many TQM studies we examined claim that the successful implementation of TQM could also generate improved products and services, lower costs, more satisfied customers, and empowered employees (Agus and Abdullah, 2000; Wadsworth et al., 2002; Chin et al., 2003). Chin et al. (2003) contend that successful TQM results increased return on investment, and market share as well. The survey respondents perceived customer focus and leadership to be the most important elements to implement TQM in Hong Kong manufacturing industries (Chin et al., 2003, 2002a). TQM has the potential not only to increase competitiveness, but also to improve organizational effectiveness (Ugboro and Obeng, 2000; Fok et al., 2001). Radovilski et al. (1996), surveyed 235 companies, showed that increases in profit, market share, and productivity, reductions in defects and costs of achieving quality are among the benefits of

82 ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 implementing TQM. Huang and Lin (2002) reported that cost reduction, sales increase, product image, service quality, and overall reputation are some of the improvements companies achieved by implementing TQM. Sousa and Voss (2002) recently wrote a reflective review article on TQM. They have synthesized and organized the literature in the field and offered suggestions for further research. They found out that one important area of research in TQM has been the examination of the extent to which TQM practices have an impact on firm performance the quality performance model. Some TQM practices did not have a significant impact on organizational performance. They stated that further research is needed to clarify the relative importance between mechanistic/process/technical ( core ) and nonmechanistic/sociobehavioral ( infrastructure ) TQM aspects. They also found out that some rigorous academic studies have started to question the universal validity of TQM practices, and to investigate the influence of the organizational context on TQM practice. We concur with Conca et al. (2004) that the critical factors of TQM differ from one author to another. The current literature emphasizes various critical factors of TQM implementation based on different implementation practices. However, as demonstrated in Sousa and Voss (2002) there are certain constructs such as top management commitment, customer focus, supplier quality management, employee involvement, employee empowerment, employee training, and statistical process control usage which are all present in the previous frameworks proposed. The framework we propose in this research encompasses common constructs expanded by all relevant variables and relationships including factors comprising expected benefits, risks, and costs of TQM. It also includes Turkish manufacturing industry-specific constructs such as dynamic structure of Turkey, unavailability of country-specific TQM models, and conflict of Turkish management structure and TQM. 3. Methodology background In this study ANP serves as the decision analysis tool and we implemented it using Super Decisions r, a sophisticated and user friendly software that implements ANP (Saaty, 2001a). ANP makes it possible to deal systematically with the interactions and dependencies among the factors in a decision system. Our factors, as mentioned before, were based on the results of our literature search and on the survey conducted among 250 large manufacturing companies in Turkey. In this section, we describe the ANP methodology and the survey carried out to identify relative importance of different critical success factors of TQM. 3.1. Analytic network process The ANP is a generalization of AHP and can be used to treat more sophisticated decision problems. ANP is a coupling of two parts. The first consists of a control hierarchy (or network) of criteria and subcriteria that control the feedback networks. The second part consists of the networks of influence that contain the factors of the problem and the logical groupings of these factors into clusters. Each control criterion (or sub-criterion) has a feedback network. A supermatrix of limiting influence that gives the priorities of the factors in the network is computed for each network (Saaty, 2001c). We used Saaty s benefits, opportunities, costs, risks (BOCR) approach. The BOCR are called the merits of the decision (Saaty, 2001a). Often, though not in this work, Saaty further decomposes the merits into control criteria such as economic, political, and social, each of which will have a decision network (sometimes referred to as a subnet) associated with it. Here we directly set up the subnets for each merit. Each decision network is composed of clusters, their elements, and links between the elements. A link between an element (the parent ) and the elements it connects to in a given cluster (its children ) makes up the usual AHP pairwise comparison set. The interactions, feedback, influences, and dependencies in the system are expressed through these links. Links between elements within the same cluster are called inner dependencies, whereas links between a parent element in one cluster and its children in another cluster are called outer dependencies (Saaty, 1999). Inner and outer dependencies are the best way decision-makers can capture and represent the concepts of influencing or being influenced, between clusters and between elements with respect to a specific element. Pairwise comparisons are made systematically for all combinations using the fundamental comparison scale (1 9) of AHP that is used to indicate how many times an element dominates another. The decision-maker can express his

ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 83 preference between each pair of elements verbally as equally important, moderately more important, strongly more important, very strongly more important, and extremely more important. These descriptive preferences would then be translated into numerical values 1, 3, 5, 7, 9, respectively with 2, 4, 6, and 8 as intermediate values for comparisons between two successive judgments. Reciprocals of these values are used for the corresponding transpose judgments. In making judgments, the decisionmaker can incorporate experience, knowledge and hard data (Harker and Vargas, 1990). Tangibles can be included in the model alongside intangibles. After the pairwise comparisons are completed, the results are synthesized. Recognizing that BOCR often are not equally important in coming to a decision, they are rated separately, one at a time, with respect to the strategic criteria to establish their priorities for the decision. Finally, the BOCR priorities are used to weight and combine the results of the decision networks beneath them (Saaty, 2001a). For Costs and Risks pairwise questions, one asks which is the more costly or risky, so the results from those subnets must be inverted as shown in the formula below (for each alternative s priorities): If we have the priorities p i ; i ¼ 1;...; 4, for each B, O, C, R, the formula is represented as p 1 B þ p 2 O þ p 3 1=C þ p 4 1=R. The results are normalized to yield the overall priorities of the alternatives. The ANP methodology is explained in Saaty s (2001b) book, so we are not going to explain all the intricacies of the methodology due to space limitations. Below we give enough of the general approach to enable the reader to follow the paper with ease. Step 1 (BOCR priority development): There are two ways to combine the BOCR subnet priorities depending upon whether they are equally weighted or not: (a) Multiplicative analysis: if benefits, costs, opportunities, and risks are all equally important, obtain for each alternative a single overall weight using the ratio of the four in the form: BO/CR or (benefits times opportunities) over (costs times risks) and then choose the alternative with the largest value. (b) Additive analysis: as we explained above, if benefits, costs, opportunities, and risks are not equally important, rate the BOCR one at a time with respect to high-level personal or corporate strategic criteria that are used to assess the merits of the different decisions we make (Saaty, 2001b). Step 2 (Model construction): Determine one decision network for each merit: benefits, opportunities, costs, and risks. If in some cases one or more of the merits are unimportant for the decision, leave them out. Determine all the elements that affect the decision and group them into clusters for each network. Each decision network must contain a cluster of alternatives. The other clusters contain elements that are related to the issue such as criteria or stakeholders. Step 3 (Formulating the links and performing paired comparisons between the clusters/elements): In each network formulate the links between the elements and perform the following paired comparisons to derive eigenvectors and to form a supermatrix. (a) Cluster comparisons: Perform paired comparisons on the clusters that influence a given cluster with respect to the control criterion for that network. Weights derived from this process will be used to weigh the elements in the corresponding column blocks of the supermatrix for that network. (b) Comparisons of elements: Perform paired comparisons on the elements within the clusters. Compare the elements in a cluster according to their influence on an element in another cluster to which they are connected (or on elements in their own cluster). (c) Comparisons for alternatives: Compare the alternatives with respect to all the elements from which they are connected. Step 4 (Constructing the supermatrix): The outcome of the process above is the unweighted supermatrix. Its columns contain the priorities derived from the pairwise comparisons of the elements. In an unweighted supermatrix, its columns may not be column stochastic. To obtain a stochastic matrix, i.e., each column sums to one, multiply the blocks of the unweighted supermatrix by the corresponding cluster priority. Raise the supermatrix to a large power to capture first, second, and third degree influences. When the differences between corresponding elements of a column are less than a very small number, for successive powers of the supermatrix, the process has converged. Step 5 (Obtaining overall outcome): Use the resulting ratings, respectively, to multiply the benefit priorities of the alternatives, the opportunities

84 ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 priorities of the alternatives, the normalized reciprocals of the costs priorities of the alternatives, and the normalized reciprocals of the risks priorities of the alternatives ðp 1 B þ p 2 O þ p 3 1=C þ p 4 1=RÞ. Since the alternatives that have the highest priority under Costs and Risks are more costly or risky, convert the priorities so that less preferred alternatives have lower values than more preferred ones, by taking the reciprocal of each alternative s priority. Add these four resulting numbers for each alternative to obtain its overall priority. Repeat for all the alternatives and normalize. Finally, select the alternative with the highest priority. 3.2. The survey A sample of 250 large companies in Turkey was randomly drawn from the list of 500 biggest manufacturing companies in Turkey as determined by the Istanbul Chamber of Industry. Companies in the service industry were not included. Questionnaires were mailed directly to quality managers of the selected companies. In this research, quality managers and directors were the dominant respondents. A cover letter was included in the mailing, explaining the purpose of the study, along with the questionnaire and a stamped return envelope. Of the 250 surveys mailed, 25 were returned as undeliverable and 100 usable responses were received for a net response rate of 40.0%. Some of the questionnaires that were returned were incomplete. Sixtytwo percent of the companies who responded had stated that they fully implemented TQM. Fourteen percent of the respondents reported that they had not implemented TQM. Twenty-four percent of them reported that they were preparing to implement TQM (Bayazit, 2003). The major purpose of our research was to identify relative importance of factors affecting the adoption of TQM implementation and finding out how favorable the conditions were for adopting TQM implementation in the Turkish manufacturing industry. We needed the experts who had TQM implementation experience, since ANP requires expert judgment. Therefore, we selected the 62 companies that had already implemented TQM for our analysis. Those companies quality managers and directors had the required experience to judge the relative importance of different factors on TQM implementation. In actuality, those were the firms who answered all the questions. The questionnaire consisted of 25 questions. The first part of the questionnaire addressed issues related to demographic information and the implementation process of TQM. The second part of the questionnaire addressed the issues related to the factors affecting TQM implementation. The factors we used were a combination of those determined based on the literature search and those added from the survey. Initially we developed a total of 28 factors based on literature research. Some of the open questions in the survey addressed issues related to improvements achieved, implementation goals, difficulties and success factors for implementing TQM and these helped us to determine the rest. We compiled a total of 45 factors through the survey analysis. Most of them matched those we found in the literature. Yet some of the factors originated from Turkish management structure and were completely different. We have gone through some initial evaluation of the factors to eliminate those that were irrelevant to the decision problem. Of the factors that were refined from the survey, 13 were eliminated and 32 were perceived to be significant to the decision problem and hence considered as the main factors. These factors were then categorized under advantages, opportunities, risks, and disadvantages. We used the term advantages and disadvantages for Saaty s terms benefits and costs. The main categories also were divided into sub-categories (Karpak and Bayazit, 2001). The judgments used in the analysis used the fundamental 1 9 scale of AHP (1 ¼ equally, 3 ¼ moderately, 5 ¼ strongly, 7 ¼ very strongly, 9 ¼ extremely) and were our interpretation of survey results. We also used intermediate values where appropriate (2 ¼ equally to moderately, 4 ¼ moderately to strongly, 6 ¼ strongly to very strongly, 8 ¼ very strongly to extremely). The factors that will be used to evaluate the alternatives are described in Table 1. Most of them coincide with the factors considered in the literature, such as worker participation, work satisfaction, upper management support, and zero-defects. We expanded these factors with expert judgments (the survey participants quality managers/directors, experts who have been through TQM implementation experience) and captured some additional constructs which are exclusive to the Turkish manufacturing industry. Of those Turkish manufacturing industry-specific constructs, these experts reported that the number of certified consultant companies who were assisting with TQM

ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 85 Table 1 The decision criteria Complaint reduction Quality improvement Price reduction On time delivery Higher revenue Market share Internal costs External costs Appraisal costs Work-in process inventory (WIP) Zero-defects Workforce (WF) quality The companies that are applying TQM reported an increase in customer satisfaction and a decrease in customer complaints. It was reported that the quality is improved. The companies that are applying TQM achieved higher quality with lower price. The companies that are applying TQM delivered their products to markets when their customers want to receive them. It was reported that on-time delivery frequency increased markedly. The respondents reported that highquality products are priced higher than comparable lower quality ones and yield a greater return. Because of higher quality the companies that are applying TQM reported an increase in market share. The costs include the costs of scrap, the cost of repair, rework, paperwork, rescheduling, delays caused by defective products and all the costs of delays, paperwork, rescheduling caused by defective products. Because of the principle is to make it right the first time, the respondents cited that total internal quality costs decreased. These costs include warranty costs, the cost of returns or recalls, and lost business and good will. Since the product quality is improved, companies reported that total external quality costs decreased. This includes the cost of inspection, testing and other quality control activities. The respondents reported that total appraisal costs decreased. The companies that are applying TQM achieved reduction in work-in process inventory. In the TQM approach it is very important to design a production process that does the job right the first time. The companies that are applying TQM reported a decrease in defects. Although their ultimate aim is to reach zero-defects, only 5% of respondents achieved this goal. Survey participants reported that they gave workers the responsibility for improvements and the authority to make changes to accomplish them. That way they placed decision making into the hands of those closest to the job with considerable insight into problems and solutions. As a result Table 1 (continued ) Worker participation Work satisfaction Upper management support Quality education and training Dynamic structure of Turkey Workers support they achieved improvement in workforce quality. The respondents reported that the use of teams for problem solving achieves consensus, gets people involved, and promotes a spirit of cooperation and shared values among the employees. The survey participants reported that giving workers the responsibility for improvements and the authority to make changes to accomplish them provided strong motivation and improved morale for employees. It gave them as sense of job ownership. They cited that all these factors increased work satisfaction. The respondents reported that many problems on the managerial side result from a lack of top management support. Top management must be committed and involved. If they are not, TQM will become just another fad that quickly dies and fades away. Even if upper management support is there initially, it might be lost later as the process progresses. In TQM philosophy, all employees from the shop floor to the board room as well as suppliers and customers should participate in a comprehensive training program. These programs were aimed not only at statistical quality control techniques but also at the broader concepts of TQM. The respondents reported that lack of quality education and training is a very common shortcut. On the other hand, quality education and training may become too expensive and finding qualified instructor may be challenging. The survey participants reported that the number of certified consultant companies who are assisting with TQM implementation is very limited and mostly uncertified and nonauthorized consultants are willing to exercise TQM. Although quality education and training is a major disadvantage as pointed out above, the respondents consider it as an expected shortcoming as well. Turkey is a dynamic country with up and downs in its economy. Since it creates risk, the survey participants reported that it was very difficult for the companies to implement TQM in this kind of environment. There may be silent resistance from the workers. Even if there is a support

86 ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 Table 1 (continued ) Table 1 (continued ) Change in perception Company supplier relationship Knowledge in statistics Costly and a longterm study Unavailability of country-specific models Family companies from the workers initially, the respondents cited that workers support might be lost later on when the process progresses. The respondents reported that there might be an expected change in managers perception of teamwork. Selecting and developing vendors that fit into the TQM system is an important issue. Long-term relationships are cultivated so that vendors deliver parts of perfect quality. Currently vendors are generally regarded as adversaries. Companies often use multiple vendors, playing one off against another, and there is a heavy emphasis on price. The respondents reported that lack of vendors as long-term partners creates an important risk to successfully implement TQM in organizations. The respondents reported that not enough familiarity with statistical ways of thinking and limited use of statistical methods throughout the organization are another common failing. The survey participants reported that implementing TQM is costly. Preventing defects, the cost of training, charting of quality performance to study trends, revising product designs, making changes to production processes, working with vendors and other activities aimed at improving quality and preventing defects is costly. They also cited that adopting TQM is a long-term process. It may take 2 6 years to become fully operational in the organizations. The respondents reported that they adopted Deming, Juran, and Japanese models in their organizations. Yet they believe that due to differences between countries, specific TQM models are necessary. Since country-specific models were unavailable, they had to use models developed mainly for other countries. It was reported that approximately 90% of the companies in Turkey are family partnerships and not managed by professionals. The main reason is that company owners do not want to lose their power at every level of management. This exacerbates the conflict between traditional Turkish managers and those who would implement TQM. TQM expenses Conflict of Turkish managementstructure and TQM Cooperation level Difficulty of achieving teamwork Long-term competitive power Workforce harmony Achieving quality culture The respondents reported that TQM expenses are considered unnecessary by upper management. The respondents reported that in many companies in Turkey the system is autocratic and top down even if the company is professionally managed. On the other hand, TQM requires a democratic bottom up approach. Thus, it creates conflict. The respondents reported that the cooperation level between main and subsidiary industries is very low. The respondents described the working style of employees as individualistic meaning they are perceived to prefer acting independently rather than joining teams. Although the implementation of TQM requires worker participation, the need to have it is viewed as a limitation of TQM implementation as well. The survey participants reported that TQM could generate a sustainable competitive advantage in the long term, since TQM is capable of producing a cost-based advantage. The companies that are successfully applying TQM concepts in their organizations expect to achieve workforce harmony in the long term. The companies that are successfully applying TQM expect to achieve a quality focused culture and consensus in their organizations. implementation was very limited. One of these Turkish manufacturing industry-specific constructs is the limited number of certified consultant companies who assist with TQM implementation, as reported by these experts. Additionally, they stated that mostly uncertified and non-authorized consultants were willing to exercise TQM. Quality education and training is therefore, a major disadvantage and was considered by the respondents as an expected shortcoming as well. Unavailability of country-specific models is another factor affecting TQM implementation in Turkey. The respondents reported that they adopted Deming, Juran, and Japanese models in their organizations. Yet they believe that due to differences between countries, specific TQM models are necessary. Since countryspecific models were unavailable, they had to use models developed mainly for other countries. Although we developed 32 factors, Table 1 shows

ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 87 31 factors since the quality education and training criterion were considered both a risk and a disadvantage by the survey respondents. 4. The ANP decision model The ANP-based framework seems to be suitable to identify the relative importance of different factors on TQM implementation, since there is feedback and dependence among them. The framework also assesses the readiness of the Turkish manufacturing industry to adopt TQM. In this section we describe the ANP decision model we used. Step 1 (BOCR weight development): Unlike traditional cost/benefit analysis, this ANP model considers different weights for the merits. We changed the original terms to ones we thought better described the TQM adoption issue so that benefits became advantages, opportunities stayed the same, costs became disadvantages, and risks stayed the same. The strategic criteria we used to determine the priorities of the BOCR merits are shown in Fig. 1. These weights are obtained by using the Rating approach of AHP (Saaty, 2001b). The strategic criteria are cost of implementing TQM, amount of time required for implementation, and effect on product quality. These are the main criteria needed when a company makes a decision about implementing TQM. Cost of implementing TQM refers to the required cost of implementing TQM such as training costs, etc. Amount of time required for implementation refers to how much time the alternatives will require for implementation. Effect on product quality means how TQM implementation is likely to affect product quality. The four merits of: advantages, opportunities, disadvantages, and risks were rated according to five intensities (very high, high, medium, very low, low) listed below along with their priorities. For example, effect on product quality creates several advantages to the company but has neither risk nor disadvantage. Cost of implementing TQM represents investments and creates disadvantages and risks to the company. The BOCR priority calculations are summarized in Table 2 below and the priorities in bold at the bottom of the table are used in the main top-level structure to synthesize results. Priorities for the ratings intensities, shown listed across the top of Table 2, were derived from pairwise comparisons. The results for each cell are computed by multiplying the weight of the strategic criteria in the left column by the priority of the rating selected. For example, the (cost, advantages) cell in the table below is assigned a rating of very low. So the value for very low, 0.06 is multiplied by the value for cost, 0.635, to give the value for that cell. The values are summed for each column and the totals thus obtained are normalized to yield the priorities at the bottom of Table 2. Step 2 (Model construction): In applying ANP, the next step is to structure the model to be evaluated. The overall objective of this ANP model is to Fig. 1. BOCR merit criteria. Table 2 Priority ratings for the merits: advantages, opportunities, disadvantages and risks Advantages Opportunities Disadvantages Risks Cost (0.635) Very low Low Very high High Quality (0.287) Very high High Very low Very low Time (0.078) Very low Low Very high High Priorities 0.321 0.288 0.195 0.195 Very high (0.420), high (0.260), medium (0.160), low (0.10), very low (0.06).

88 ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 evaluate the ultimate relative importance of different factors that impact the implementation of TQM and assess the readiness of the Turkish manufacturing industry to implement TQM. The factors that will be used to evaluate the alternatives were developed earlier in the paper. Two alternatives, Turkish industry should apply TQM and Turkish industry should not apply TQM are determined and will be evaluated according to these factors. There are four feedback networks one for each of four general controlling factors (the merits of the decision): advantages, opportunities, disadvantages, and risks. The advantages reflect the current benefits of TQM implementation. The opportunities reflect the potential benefits of TQM implementation. The disadvantages reflect the limitations of TQM implementation whereas the risks reflect the expected shortcomings of TQM implementation. First, the factors listed above that affect TQM implementation are classified into advantages, opportunities, disadvantages, and risks. Then they are grouped into clusters in the networks under their respective merits. The clusters in the advantages network are: advantages to customers, advantages to workforce, operational advantages, and financial advantages. There is an alternatives cluster in every network. The opportunities network consists of only one other cluster, potential benefits. Since some of the disadvantages are inherent to TQM whereas some others originate from the current Turkish management structure, the disadvantages network is comprised of the clusters: inherent to TQM, and originates from current Turkish management structure and alternatives. The risks network has the clusters: managerial risks, and technical risks and alternatives. A graphical summary of the overall ANP model is shown in Fig. 2. If links exist from at least one element of a cluster to the elements of another cluster, there is an arrow connecting the clusters. A Fig. 2. The ANP-based framework for successful TQM implementation and for assessing the readiness of Turkish industry for TQM.

ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 89 looped arc on a cluster indicates inner dependence among the elements of the cluster. The directions of the arcs signify dependence. Step 3 (Formulating the interdependencies and performing pairwise comparisons between clusters/ factors): We then formulated interrelationships among all the factors. The question asked when formulating these relationships was: With respect to a specific factor, which of a pair of factors influences it more? For example, with respect to complaints reduction which one affects it more, quality improvement or price reduction; quality improvement or on time delivery? After creating the links between the factors, pairwise comparisons are performed with respect to all those factors that have an impact on other factors within their own cluster or other clusters of the network. In this case, the factors in a cluster are compared according to their influence on a factor in another cluster to which they are connected (or on factors in their own cluster to which they are connected). To establish the interdependencies in the networks, pairwise comparisons among all the factors are conducted and these relationships are evaluated. The next step is to weigh the clusters. When clusters are connected, because an element in one is connected to several elements in another, or vice versa, then logically they have an influence on each other. A series of pairwise comparisons are conducted on the clusters themselves with respect to impact on clusters to which they are linked and the weighted priorities are then calculated resulting in the cluster matrix of priorities used to weigh the corresponding blocks of the unweighted supermatrix. Let us examine in detail the network that belongs to the advantages merit. Tables 3 7 are for that network. The cluster matrix, composed of the eigenvectors derived from making pairwise comparisons of the clusters within the advantages network, is shown in Table 3. It shows how much a cluster (as the column heading) is influenced by the other clusters it connects to. For example, the cluster of advantages to workforce and operational advantages are not influenced by the cluster of advantages to customer because those entries have zeros indicating no effect or dependence among the clusters. On the other hand, the cluster of advantages to workforce influences the cluster of advantages to customers (0.2375). The cluster of financial advantages is influenced by the cluster of advantages to customer (0.3149). The cluster of alternatives is influenced by all the clusters. Step 4 (Constructing supermatrix): As an example, Tables 4 6 illustrate some parts of unweighted, weighted and limit supermatrices of the factors within the advantages network. Table 4 shows the pairwise comparisons of the factors. The weighted supermatrix (Table 5) is obtained by weighting the blocks in the unweighted supermatrix by the corresponding priority from the cluster matrix shown in Table 3. The entries of the weighted supermatrix itself give the direct influence of any one factor on any other factor. The weighted supermatrix has some zeros indicating no interaction. For example, complaints reduction is influenced by quality improvement (0.1990), price reduction (0.3093), on-time delivery (0.0676), workforce quality (0.2903), and zero defects (0.1256). On the other hand, worker participation and WF satisfaction do not affect the complaints reduction. Decrease in internal quality costs (0.2022), external quality costs (0.5709), and appraisal quality costs (0.1075) influence price reduction. Table 6 shows the stable priorities of all the factors. From it the priorities of all the factors and alternatives are extracted and normalized. In the limit matrix, the columns are all the same. To determine the final local priorities the priorities of the factors for each cluster in the columns of the limit matrix are normalized to one. As an example, within the cluster of advantages to customer the factor quality improvement is considered to be the most important with 0.605 or 60.5% as shown in Table 7 in the column labeled local priorities Table 3 Cluster matrix for the advantages network Clusters Advto cust Advto workf Operadv Fincadv Alt Advto cust 0.4532 0.0000 0.0000 0.3149 0.5523 Advto workf 0.2375 0.9000 0.0000 0.0000 0.0985 Operadv 0.1029 0.0000 0.9000 0.1594 0.0633 Fincadv 0.1818 0.0000 0.0000 0.4980 0.2857 Alt 0.0246 0.1000 0.1000 0.0316 0.0000

90 ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 Table 4 Unweighted Supermatrix Advto cust Advto workf Operadv... Comp. red Quality Price On-time WF qu Worpart WF sat Zero WIP Adv to cust Comp 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Quality 0.3196 0.0000 0.0000 1.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Price 0.5584 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 On-time 0.1219 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Advto workf WF qu 1.0000 0.0000 0.0000 0.0000 0.0000 0.3333 0.2500 0.0000 0.0000 Worpart 0.0000 0.0000 0.0000 0.0000 1.0000 0.0000 0.7500 0.0000 0.0000 WF sat 0.0000 0.0000 0.0000 0.0000 0.0000 0.6667 0.0000 0.0000 0.0000 Operadv Zero 1.0000 1.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 WIP 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Fincadv Higher rev 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Market 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Intcosts 0.0000 0.0000 0.2296 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Extcosts 0.0000 0.0000 0.6483 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Apprcosts 0.0000 0.0000 0.1220 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Alt Apply TQM 0.8889 0.9000 0.7500 0.6667 0.8000 0.7500 0.7500 0.8750 0.8750 Not apply TQM 0.1111 0.1000 0.2500 0.3333 0.2000 0.2500 0.2500 0.1250 0.1250.... (normalized by cluster). The global priorities of those factors are then calculated by weighting their local priorities by the priority of the advantages merit, 0.321. For example, for quality improvement we have 0:321 0:605 0:194. When we again normalized, we found 0.084. Table 7 shows the limit priorities of the factors in the advantages network as they appear in the limit supermatrix, with the priorities normalized to one for each cluster and for the global priorities. We then calculated the global priorities of all the factors as explained above. The global priorities indicate that zero defects is the most important factor with a global weight of 13.3%. The second is costly and long-term study with a global weight of 10.2%. It shows that zero defects and costly and long-term study are the factors most affecting TQM implementation. Table 8 shows the global priorities of all the factors in the decision-making model. Step 5 (Obtaining the overall outcome): As Saaty (2001b) suggested; we used additive synthesis to evaluate the alternatives in the final decision. In additive synthesis, we have for example for Apply TQM 0:854 0:312 þ 0:841 0:288 þ 0:178 0:195 þ0:210 0:195 ¼ 0:592. We also compared multiplicative synthesis outcome with that of additive synthesis. We have for Apply TQM: 0:854 0:321 0:841 0:288 0:178 0:195 0:210 0:195 which is then normalized by dividing by the sum of all two such expressions to obtain 0.621 for its priority. Tables 9 and 10 give the necessary information to construct the overall synthesized results, which indicate Turkish industry should apply TQM is chosen by the model, primarily with an overall priority of 0.592. 4.1. Sensitivity analysis To ensure the stability of the outcome of our analysis, we conducted sensitivity analysis. Our results seem to be not very sensitive to how important we consider the advantages to be. As shown in Fig. 3, even if the rating goes down to 13% from 32%, which is the original rating, Apply TQM is still more preferable than Not Apply TQM.

ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 91 Table 5 Weighted supermatrix Advto cust Advto workf Operadv... Comp. red Quality Price On-time WF qu Worpar WF sat Zero WIP Advto cust Comp 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Quality 0.1990 0.0000 0.0000 0.9484 0.0000 0.0000 0.0000 0.0000 0.0000 Price 0.3093 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 On-time 0.0676 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Advto workf WF qu 0.2903 0.0000 0.0000 0.0000 0.0000 0.3000 0.2250 0.0000 0.0000 Worpar 0.0000 0.0000 0.0000 0.0000 0.9000 0.0000 0.6750 0.0000 0.0000 WF sat 0.0000 0.0000 0.0000 0.0000 0.0000 0.6000 0.0000 0.0000 0.0000 Operadv Zero 0.1256 0.8067 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.9000 WIP 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Fincadv Higrev 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Market 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Intcosts 0.0000 0.0000 0.2022 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Extcosts 0.0000 0.0000 0.5709 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Apprcosts 0.0000 0.0000 0.1075 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Alt Apply TQM 0.0268 0.1740 0.0895 0.0344 0.0800 0.0750 0.0750 0.8750 0.0875 Not apply TQM 0.0033 0.0193 0.0298 0.0171 0.0200 0.0250 0.0250 0.1250 0.0125.... If opportunities were to be decreased from their original priority of 0.288 0.092, Apply TQM is still preserved as the best alternative. However, as the priority of opportunities decreases, the best alternative turns out to be Not Apply TQM. The alternatives are more sensitive to the priorities of opportunities. If disadvantages were to be increased from its original priority 0.195 0.329, Apply TQM is still preserved as the best alternative. As the priority of disadvantages increases, the best alternative turns out to be Not Apply TQM. If risks were to be increased from its original priority of 0.195 to 0.357, Apply TQM is still preserved as the best alternative. As the priority of risks increases, the best alternative turns out to be Not Apply TQM. Our sensitivity analysis indicates that for the final priorities of the alternatives to change, we would need to make extreme assumptions on the priorities of BOCR. 5. Conclusion In this paper, we have developed a framework based on ANP to identify the degree of impact of factors affecting TQM implementation and investigated the readiness of the Turkish manufacturing industry to adopt TQM practices based on a survey carried out among 250 large manufacturing companies in Turkey. We used the ANP for decision making with dependence and feedback based on four major factors as mapped to Saaty s benefits, costs, opportunities, risk (BOCR) model. ANP is a new methodology that incorporates feedback and interdependent relationships among decision attributes and alternatives. It leads to fresh insights about issues. Since we included only manufacturing companies in our research, this study indicates manufacturing industry readiness to adopt TQM. Manufacturing companies have been and will continue to be, for a while, a mainstay of the economy in Turkey and hence our results are most likely a good indicator of Turkish industry readiness in general. Based on our model we found that in Turkish industry, conditions for implementing TQM were 59.2% favorable as opposed to not implementing TQM. In the literature, only a few studies addressed the TQM readiness of an industry (Arasli, 2002; Weeks et al., 1995; Aksu, 2003).

92 ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 Table 6 Limit supermatrix Advto cust Advto workf Operadv... Comp. red Quality Price On-time WF qu Worpar WF sat Zero WIP Advto cust Comp 0.0251 0.0251 0.0251 0.0251 0.0251 0.0251 0.0251 0.0251 0.0251 Quality 0.1231 0.1231 0.1231 0.1231 0.1231 0.1231 0.1231 0.1231 0.1231 Price 0.0405 0.0405 0.0405 0.0405 0.0405 0.0405 0.0405 0.0405 0.0405 On-time 0.0148 0.0148 0.0148 0.0148 0.0148 0.0148 0.0148 0.0148 0.0148 Advto workf WF qu 0.0911 0.0911 0.0911 0.0911 0.0911 0.0911 0.0911 0.0911 0.0911 Worpar 0.1608 0.1608 0.1608 0.1608 0.1608 0.1608 0.1608 0.1608 0.1608 WF sat 0.1033 0.1033 0.1033 0.1033 0.1033 0.1033 0.1033 0.1033 0.1033 Operadv Zero 0.1279 0.1279 0.1279 0.1279 0.1279 0.1279 0.1279 0.1279 0.1279 WIP 0.0057 0.0057 0.0057 0.0057 0.0057 0.0057 0.0057 0.0057 0.0057 Fincadv Highrev 0.0215 0.0215 0.0215 0.0215 0.0215 0.0215 0.0215 0.0215 0.0215 Market 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 Intcosts 0.0155 0.0155 0.0155 0.0155 0.0155 0.0155 0.0155 0.0155 0.0155 Extcosts 0.0405 0.0405 0.0405 0.0405 0.0405 0.0405 0.0405 0.0405 0.0405 Apprcosts 0.0087 0.0087 0.0087 0.0087 0.0087 0.0087 0.0087 0.0087 0.0087 Alt Apply TQM 0.1774 0.1774 0.1774 0.1774 0.1774 0.1774 0.1774 0.1774 0.1774 Not apply TQM 0.0303 0.0303 0.0303 0.0303 0.0303 0.0303 0.0303 0.0303 0.0303.... Table 7 Factors and their priorities for the advantages network Merits Clusters Factors Priorities from limit matrix Local priorities (normalized by cluster) Global priorities Advantages (0.321) Advantages to customer Advantages to workforce Operational advantages Financial advantages Complaints 0.0251 0.123 0.017 reduction Quality 0.1231 0.605 0.084 improvement Price reduction 0.0405 0.199 0.028 On-time delivery 0.0148 0.073 0.010 WF quality 0.0911 0.257 0.036 Worker 0.1608 0.257 0.045 participation Work satisfaction 0.1033 0.291 0.040 Zero defects 0.1279 0.957 0.133 WIP 0.0057 0.043 0.006 Higher revenue 0.0215 0.215 0.029 Market-share 0.0139 0.139 0.019 Internal costs 0.0155 0.155 0.022 External costs 0.0405 0.404 0.056 Appraisal costs 0.0087 0.087 0.012

ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 93 Table 8 Factors and their global priorities Factors Complaints reduction 0.017 Quality improvement 0.084 Price reduction 0.028 On-time delivery 0.010 WF quality 0.036 Worker participation 0.036 Work satisfaction 0.040 Zero defects 0.133 WIP 0.006 Higher revenue 0.029 Market-share 0.019 Internal costs 0.022 External costs 0.056 Appraisal costs 0.012 Long-term competitive power 0.006 WF harmony 0.059 Achieving quality culture 0.060 Costly and a long-term study 0.102 Unavailability of country-specific models 0.011 Conflict Mgt./TQM 0.022 TQM expenses 0.011 Quality education and training 0.008 Family companies 0.024 Company supplier relationship 0.013 Difficulty of achieving teamwork 0.005 Upper management support 0.029 Quality education and training 0.031 Dynamic structure of Turkey 0.004 Workers support 0.012 Change in perception 0.005 Cooperation level 0.002 Knowledge of statistics 0.085 Global priorities In a decision problem, decision-makers might intuitively feel that some factors are more important than others in affecting their final preference among alternatives. If there is some feedback and interdependency among the factors, an unimportant factor may turn out to be far more important than even the most intuitively important one. So, there needs to be a methodology like ANP to capture more realistic results. In our research, we have identified 32 factors affecting TQM implementation. Some of the factors initially stated by survey participants to be the most important ones were not, interestingly enough. Because of interdependencies among the factors others turned out to be more important in the decision model. For example, according to the survey respondents quality improvement and higher revenue were the most important factors. But our ANP model showed quality improvement to be the fourth most important factor with higher revenue one of the least important factors. On the contrary, in our analysis, due to inner and outer dependencies, zero defects and costly and long-term study came up as the most crucial ones in TQM implementation. The factors affecting TQM implementation could be qualitative as well as quantitative. There are many qualitative concerns when assessing the factors critical to the TQM implementation process. Some of the factors in our decision model, for example, workers support, dynamic structure of Turkey, family companies, etc., were difficult to quantify yet we were able to include them. ANP Table 9 Inverting disadvantages and risks priorities for use in an additive formula Diasdv. 1/Disad 1/Disad. normalized Risks 1/Risks 1/Risks normalized Apply TQM 0.822 1.216 0.178 0.790 1.266 0.210 Not apply TQM 0.178 5.618 0.822 0.210 4.762 0.790 SUM 1.000 6.830 1.000 1.000 6.028 1.000 Table 10 Overall results Advantages (0.321) Opportunities (0.288) Disadvantages (0.195) Risks (0.195) Final outcome (Additive) Final outcome (Multiplicative) Apply TQM 0.854 0.841 0.178 0.210 0.592 0.621 Not apply TQM 0.146 0.159 0.822 0.790 0.408 0.379

94 ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Apply TQM Not apply TQM Apply TQM Not apply TQM Experiments Priority: ADVANTAGES 0.139473685526 Fig. 3. Sensitivity analysis for advantages. 0.504 0.496 enabled us to incorporate both quantitative and qualitative factors, which are very important in assessing factors affecting a TQM implementation. Although we believe our model provides a comprehensive framework for TQM implementation models, there are some limitations. One is that ANP requires many comparisons and a good bit of effort. However, it should not be surprising that complex decisions may require complex methodology. There are some shortcuts. For example, clustering the factors helps lessen the number of pairwise comparisons. In particular, this limitation might exist if there are several alternatives rather than just two in the decision model. In terms of making a large number of pairwise comparisons, this would be quite demanding. There is an alternative method that can be used when there are a large number of alternatives: the Ratings approach. It can be applied in a first cut to rate all the alternatives rather quickly, and then the most preferable alternatives can be more precisely evaluated by comparing them against each other. To our knowledge no prior study has been done to assess Turkish industry readiness for TQM, even using statistical methods. It would be interesting for future research to see how the results from evaluations using statistical methods compare with the ANP (or AHP) approach. Also, we may investigate further how to control the important factors affecting TQM implementation to increase the degree of readiness. Again to the best knowledge of the authors there is no framework yet considering potential benefits and risks of TQM implementation. Chin et al. (2002b) used AHP to investigate the critical factors and sub-factors that determine the adoption and implementation of TQM in the stateowned enterprises and foreign joint ventures in China. This study appears to be the closest endeavor to ours, yet our framework is an enhancement to the study of Chin et al. (2002b), since ANP is capable of dealing with all kinds of feedback and dependence when modeling a complex decision environment. Therefore, we contend that our results are more accurate. In our study the respondents stated that quality improvement and higher revenue were the most important factors. However, due to the inner and outer dependencies, zero defect and costly and long-term study turned out to be the most important factors with ANP. Chin et al. (2002b) found the soft factors of TQM emphasized by both state-owned enterprises and foreign joint ventures. They might have reached a different result had they used ANP instead of AHP. Models we have reviewed that consider benefits of TQM do not distinguish between current and potential benefits. However, as it was in our model, current and potential benefits may not have the same level of importance in the eye of experts hence will have different level of impact in TQM implementation. ANP enables us to capture the complexity of dealing with this uncertainty. This research contributes to both TQM body of knowledge and ANP implementation. From TQM perspective we propose an ANP-based framework for assessing the impact of different factors on TQM implementation. From an ANP implementation point of view it is the most comprehensive one. Based on a review of the articles in literature, only Saaty used the BOCR approach. In addition to that we have not seen any application of sensitivity analysis. In real life problems we contend that sensitivity analysis of the results is almost as important as finding what the influence of different factors is on TQM. TQM transformation is a long-term process requiring a fundamental shift in management practices and culture. This may explain why there have been an overabundance of studies questioning the value of TQM, many by the consulting firms who work with their clients to implement TQM. There is little question that when implemented

ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 95 properly, TQM can have a dramatic impact on the performance and culture of an organization (Beer, 2003). If we identify factors affecting TQM implementation and then control for these factors, we can facilitate and hopefully increase the degree of success achieved by individual firms. So we believe our model can be used in other countries, too, as a framework to measure the degree of TQM readiness of industry there with some modifications to account for country specific criteria. Sousa and Voss (2002) suggest, for future research, producing guidelines on what practices should be emphasized by organizations at different stages of TQM maturity and what might be the best TQM practice implementation sequence to reach the end result. Our model does not address this issue completely yet it can be used by Turkish companies who want to initiate TQM programs. By controlling the most influential factors, the newcomers may increase the success of TQM implementation in their companies. For future research, we suggest improving on our research to identify factors that need to be emphasized by organizations at different stages of TQM maturity. We agree with Sousa and Voss (2002) that how to implement TQM research stream has taken for granted that all TQM practices are universally applicable. Implicit in their view is that it is always possible and worthwhile to change an organization s setting to accommodate all TQM practices. However, research on what to do suggests that there may be innate organizational characteristics resulting from factors such as the nature of markets, business strategy, and country of operation that cannot or are very difficult to change in order to accommodate standard TQM. Our framework provides a Turkish manufacturing industry-specific model. As a future research, we suggest the generalization of the framework for other nations, developing an ANP-based framework. This new framework requires addition and elimination of country-specific constructs. Most of the existing models that investigate the factors affecting TQM implementation have used statistical methods. We have shown an alternative approach using expert judgment. Since ANP is capable of dealing with all kinds of feedback and dependence when modeling a complex decision environment, we contend that our results are more accurate. ANP deals with uncertainty and complexity and provides insights that other, more traditional methods could miss. Acknowledgment The authors would like to thank Rozann Whitaker, for her helpful comments and suggestions. References Agarwal, A., Shankar, R., 2002. Analyzing alternatives for improvement in supply chain performance. Work Study 51 (1), 32 37. Agus, A., Abdullah, M., 2000. Total quality management practices in manufacturing companies in Malaysia: An exploratory analysis. Total Quality Management 11 (8), 1041 1051. Aksu, M.B., 2003. TQM readiness level perceived by the administrators working for the Central Organization of the Ministry of National Education in Turkey. Total Quality Management and Business Excellence 14, 563 591. Allen, R.S., Kilmann, R.H., 2001. Aligning reward practices in support of total quality management. Business Horizons 77 84. Arasli, H., 2002. Gearing total quality into small- and mediumsized hotels in North Cyprus. Journal of Small Business Management 40, 350 360. Ashayeri, J., Keij, R., Broker, A., 1998. Global business process re-engineering: A system dynamics-based approach. International Journal of Operations and Production Management 18 (9/10), 817 831. Bayazit, O., 2003. Total quality management practices in Turkish manufacturing organizations. The TQM Magazine 15 (5), 345 350. Beer, M., 2003. Why total quality management programs do not persist: The role of management quality and implications for leading a TQM transformation. Decision Sciences 34 (4), 623 642. Black, S.A., Porter, L.J., 1996. Identification of the critical factors of TQM. Decision Sciences 27, 1 21. Chan, T.H., Quazi, H.A., 2002. Overview of quality management practices in selected Asian countries. Quality Management Journal 9 (1), 172 180. Chin, K.S., Tummala, V.M.R., Chan, K.M., 2002a. Quality management practice based on seven core elements in Hong Kong manufacturing industries. Technovation: International Journal of Technical Innovation and Entrepreneurship 22 (4), 213 230. Chin, K.S., Pun, K.F., Xu, Y., Chan, J.S.F., 2002b. An AHP based study of critical factors for TQM implementation in Shanghai manufacturing industries. Technovation: International Journal of Technical Innovation and Entrepreneurship 22 (11), 707 715. Chin, K.S., Tummala, V.M.R., Chan, K.M., 2003. Quality management practice in Hong Kong industries: A comparison between electronics and toy products manufacturing industries. International Journal of Quality and Reliability Management 20 (9), 1051 1083. Chong, C.L., 1998. TQM and ISO in Singapore: Strategy or ideology. In: Ho, S.K. (Ed.), Proceedings of the Third ICIT, HKBU, Hong Kong, pp. 421 426. Conca, F., Llopis, J., Tari, J.J., 2004. Development of a measure to assess quality management in certified firms. European Journal of Operational Research 156, 683 693.

96 ARTICLE IN PRESS O. Bayazit, B. Karpak / Int. J. Production Economics 105 (2007) 79 96 Dayton, N.A., 2001. Total quality management critical success factors, a comparison: The UK versus the USA. Total Quality Management 12 (3), 293 298. Flynn, B., Schroeder, R.O., Sakakibara, S., 1995. The impact of quality management practices on performance and competitive advantage. Decision Sciences 26 (5), 659 692. Fok, L.Y., Fok, W.M., Hartman, S.J., 2001. Exploring the relationship between total quality management and information systems development. Information and Management 38, 355 371. Ghobadian, A., Gallear, D., 2001. TQM implementation: An empirical examination and proposed generic model. Omega 29, 343 359. Hamalainen, R.P., Seppalainen, T.O., 1986. The analytic network process in energy policy planning. Socio-Economic Planning Sciences 20 (6), 399 405. Harker, P.T., Vargas, L.G., 1990. Reply to Remarks on the analytic hierarchy process. Management Science 36, 269 273. Hellsten, U., Klefsjo, B., 2000. TQM as a management system consisting of values, techniques and tools. The TQM Magazine 12 (4), 238 244. Huang, Y., Lin, B.M.T., 2002. An empirical investigation of total quality management: A Taiwanese case. The TQM Magazine 14 (3), 172 180. Karpak, B., Bayazit, O., 2001. Should Turkish industry apply TQM or not: A dynamic model. ISAHP 2001, Bern, Switzerland, August 2 4, 205 212. Laszlo, G.P., 1999. Implementing a quality management program Three Cs of success: Commitment, culture and cost. The TQM Magazine 11 (4), 231 237. Lee, J.W., Kim, S.H., 2000. Using analytic network process and goal programming for interdependent information system project selection. Computers and Operations Research 27, 367 382. Meade, L., Sarkis, J., 1998. Strategic analysis of logistics and supply chain management systems using the analytical network process. The Logistics and Transportation Review 34 (3), 201 215. Motwani, J., 2001. Critical factors and performance measures of TQM. The TQM Magazine 13 (4), 292 300. Nilsson, L., Johnson, M.D., Gustafsson, A., 2001. The impact of quality practices on customer satisfaction and business results: Product versus service organizations. Journal of Quality Management (6), 5 27. Partovi, F.Y., 2001. An analytic model to quantify strategic service vision. International Journal of Service Industry Management 12 (5), 476 499. Partovi, F.Y., Corredoira, R.A., 2002. Quality function deployment for the good of soccer. European Journal of Operational Research (137), 642 656. Radovilski, Z.D., Gotcher, J.W., Slattsveen, S., 1996. Implementing total quality management: Statistical analysis on survey results. International Journal of Quality and Reliability Management 13 (1), 10 24. Reed, R., Lemak, D.J., Mero, N.P., 2000. Total quality management and sustainable competitive advantage. Journal of Quality Management 5, 5 26. Saaty, T.L., 1999. Fundamentals of the analytical network process. Proceedings of ISAHP 1999, Kobe, Japan, August 12 14, pp. 48 63. Saaty, T.L., 2001a. Decision Making in Complex Environments The Analytic Network Process for Decision Making with Dependence and Feedback. RWS Publications, USA. Saaty, T.L., 2001b. Decision Making with Dependence and Feedback: The Analytic Network Process, second ed. RWS Publications, USA. Saaty, T.L., 2001c. Decision making with the ANP and the national missile defense example. Proceedings of the Sixth International Symposium on the AHP, ISAHP 2001, Bern, Switzerland, pp. 365 382. Sarkis, J., 1998. Evaluating environmentally conscious business practices. European Journal of Operational Research 107, 159 174. Sarkis, J., 1999. A methodological framework for evaluating environmentally conscious manufacturing programs. Computers and Industrial Engineering 36, 793 810. Sarkis, J., 2003. Quantitative models for performance measurement. International Journal of Production Economics 86, 81 90. Sarkis, J., Talluri, S., 2002. A synergistic framework for evaluating business process improvements. The International Journal of Flexible Manufacturing Systems (14), 53 71. Scharitzer, D., Korunka, C., 2000. New public management: Evaluating the success of total quality management and change management interventions in public services from the employees and customers perspectives. Total Quality Management 11 (7), 941 953. Sousa, R., Voss, C., 2002. Quality management re-visited: A reflective review and agenda for future research. Journal of Operations Management 20, 91 109. Tsang, J.H.Y., Antony, J., 2001. Total quality management in UK service organizations: Some key findings from a survey. Managing Service Quality 11 (2), 132 141. Ugboro, I.O., Obeng, K., 2000. Top management leadership, employee empowerment, job satisfaction, and customer satisfaction in TQM organizations: An empirical study. Journal of Quality Management 5, 247 272. Wadsworth, H.M., Stephens, K.S., Godfrey, A.B., 2002. Modern Methods for Quality Control and Improvement, second ed. Wiley, New York. Wang, G., Huang, S.H., Dismukes, J.P., 2004. Product-driven supply chain selection using integrated multi-criteria decisionmaking methodology. International Journal of Production Economics 91, 1 15. Weeks, B., Helms, M.M., Ettkin, L.P., 1995. A physical examination of health care s readiness for a total quality management program: A case study. Hospital Material Management Quarterly 17, 68 75. Woon, K.C., 2000. Assessment of TQM implementation benchmarking Singapore s productivity leaders. Business Process Management Journal 6 (4), 314 324. Young, G.J., Charns, M.P., Shortell, S.M., 2001. Top manager and network effects on the adoption of innovative management practices: A study of TQM in a public hospital system. Strategic Management Journal 22, 935 951.