中 華 大 學

中 華 大 學 博 士 論 文

服務缺失辨識與服務修復管理系統之建構 Building an Integrative Service Management System for Service Failures Identification and

Service Recovery

系 所 別 ： 科技管理博士學位學程 學號姓名 ： D 0 9 6 0 3 0 0 1 詹 雅 慧 指導教授 ： 林 淑 萍 博 士 楊 振 隆 博 士

中 華 民 國 100 年 5 月

摘要

近年來隨著產業結構的改變，服務業對於國內產業經濟發展而言，占有不可或缺 的地位，因此如何締造成功的服務業營運型態，為持續提升經濟成長之重要關鍵。由 此可知，為了於競爭市場環境中永續經營，提供甚或創造高品質的服務以降低消費者 抱怨，並進而增加消費者滿意度以及知覺價值，為服務產業之長期策略目標。然而，

對於服務業者而言，動態以及高度不確定性的消費者市場，導致服務缺失的產生往往 無法避免，因而如何針對已發掘之服務缺失進行完善的服務修復，便成為近年來學術 界以及實務界所共同關注之議題。有鑑於品質機能展開(quality functional deployment;

QFD)為一項能將消費者需求轉換至產品規劃與設計需求的決策方法，本研究之主要 目的即在於延伸QFD之主要概念，建構有助於服務缺失辨識，以及後續服務修復方案 規畫之系統性管理模式(後簡稱為SFR管理模式)。

為了提升SFR管理模式之決策效度，本研究首先延伸Kano模式之二維概念，提出 價值缺口分析法(valuable-gap-analysis; VGA)以及服務缺失改善率函數(KIR function)，

以做為辨識服務缺失以及建構服務缺失改善潛能矩陣(improvement opportunity matrix) 之基礎。此外，為進一步了解服務修復方案與服務缺失間之對應關係，並基於各服務 修復方案執行時彼此相依性之實務性考量，本研究進一步導入決策實驗分析法 (Decision-Making Trial and Evaluation Laboratory; DEMATEL)，以強化服務修復方案投 資決策之實務有效性。最後，本研究以台灣行動通訊產業為研究個案進行實證分析。

整體而言，研究結果顯示QFD之應用以及各項理論之整合，實有助於協助管理者擬定 有助於修復服務缺失之最佳/核心方案，透過本研究所獲得之相關研究以及實務意涵 亦說明如後。

關鍵字：服務缺失、服務修復、品質機能展開、Kano 理論、行動通訊產業

ABSTRACT

As service sector plays a decisive role in economic development, it is widely recognized that the success of the service sector is an essential factor in measuring an economy’s progress. Especially in today’s competitive environment, delivering superior service to decrease customer complaints and further increase customer satisfaction and value is critical to a firm’s sustainability. As it is difficult for firms to avoid all service failures, the service recovery implementation has then become an important issue. In view of quality functional deployment (QFD) is useful for ensuring that the customer voice is systematically deployed throughout all the stages of product planning and designing, the aim of this study is to apply QFD to develop a SFR model to connect service failures and recovery solutions systematically.

To complete the SFR analysis, two advanced methodologies, such as valuable-gap-analysis and KIR function, are proposed with the involvement of Kano’s two-dimensional theory for service failures identification and improvement opportunity matrix construction. Then the Decision-Making Trial and Evaluation Laboratory method is applied to understand the interdependence among recovery solutions as well as the relationship between service failures and recovery solutions, in order to better prioritize recovery solutions operations. The SFR model proposed in this study is illustrated and validated using data collected from mobile telecom industry in Taiwan. Overall, results show that QFD is useful for identifying the core/optimal solutions for service recovery.

More implications are discussed in this study.

Keywords: service failure, service recovery, quality functional deployment, Kano theory, mobile telecom industry

ACKNOWLEDGEMENT

回顧研讀博士的四年期間，我深深體會求學之路是多麼艱辛，但卻也因此享受著 學習與成長所帶來的滿足!!

博士論文的完成意味著我將踏上人生旅途的另一階段，於此同時，我必需感謝指 導教授林淑萍博士以及楊振隆博士的悉心教誨；老師豐沛的學養及嚴謹的治學態度，

奠定我深厚的學術基礎以及素養，加上老師平時在待人處事上的諄諄教誨，更是讓我 受益匪淺。對於老師們的栽培與提攜，在此謹致上最崇高的敬意與最誠摯的謝意。

其次，承蒙美國堪薩斯州立大學許淳教授以及中華大學田效文教授給予多方面的 關懷及鼓勵，讓我了解學習的歷程中，除了學術的成長外還有更加重要的事情；我將 永遠銘記老師曾說過的每一句話，作為時時提醒自己的座右銘。而於論文審查及口試 期間，也幸蒙交通大學袁建中教授、台北商業技術學院張世佳教授、台中技術學院陳 柏亮教授、中華大學謝玲芬教授以及林錦煌教授不吝指正賜予許多寶貴建議，使本論 文更趨周延與慎密，在此致上最誠摯的謝忱。

此刻，謹懷著一顆感恩的心，感謝我的知心好友由安、怡君、于珈、品憲，博班 同學秋月、碧雲、美蘭、銘杰、林澤、明海、浩峰，以及最可愛的學妹馨文、力嘉和 系助理俐苓，因為你們時時的關心與鼓勵，讓我能在學習路上不覺孤單，大家對我的 好，我一定會銘記於心。此外，感謝親愛的家人一直以來對我的支持和祝福，讓我在 緊鑼密鼓論文著述的最後階段有了最強大的後盾。最後謹於此榮耀，呈現給所有關心 我及我所關心的人，願美好的祝福讓大家永遠健康和快樂。

詹雅慧 謹識於中華大學科技管理學系 中華民國 100 年 4 月 31 日

CONTENTS

摘要 ... i

ABSTRACT ... ii

ACKNOWLEDGEMENT ... iii

CONTENTS ... iii

TABLES ... vi

FIGURES ... vii

CHAPTER 1 INTRODUCTION ... 1

Section 1 Research Background and Purpose ... 1

Section 2 Research Processes ... 4

CHAPTER 2 LITERATURES REVIEW ... 6

Section 1 Service Failure and Recovery ... 6

Section 2 Identification of Service Failures ... 10

Section 3 Resource Based View Theory ... 21

Section 4 Quality Function Deployment ... 23

Section 5 Resource Dependence Theory ... 31

CHAPTER 3 NEW MODEL FOR CONTINUOUS SERVICE MANAGEMENT ... 38

Section 1 Valuable Gap Analysis ... 39

Section 2 Improvement Rate Function ... 42

Section 3 Implementation Procedures of SFR Model ... 44

CHAPTER 4 EMPIRICAL CASE STUDY ... 48

Section 1 Background of the Empirical Case... 48

Section 2 Questionnaire Design ... 48

Section 3 Sampling and Empirical Case Profile ... 49

Section 4 Reliability and Validity Analysis ... 51

CHAPTER 5 EMPIRICAL CASE RESULTS ... 55

Section 1 Service Failures Identification ... 55

Section 2 Improvement Rate Calculation... 60

Section 3 Resource Investment Matrix Construction ... 61

Section 4 Service Recovery Matrix Construction ... 66

CHAPTER 6 CONCLUSION ... 69

Section 1 Research and Practical Implications ... 69

Section 2 Research Limitations ... 70

REFERENCE ... 71

APPENDIX A ... 82

APPENDIX B ... 83

APPENDIX C ... 86

TABLES

Table 1 Summary of service failure definition ... 7

Table 2 Kano et al’s classification scheme ... 14

Table 3 The comparisons between different classification methods ... 17

Table 4 Summary of application and extension of Kano theory ... 19

Table 5 Summary of QFD-based literatures ... 27

Table 6 Summary of DEMATEL-based literatures ... 36

Table 7 The respondents’ demography ... 50

Table 8 Results of reliability and validity analysis for service quality scale ... 51

Table 9 Results of reliability and validity analysis for web quality scale ... 52

Table 10 Descriptive statistics and VGA results of service quality practices ... 55

Table 11 Descriptive statistics and VGA results of web quality practices ... 56

Table 12 Goodness-of fit testing of the BPNN for service quality scale ... 58

Table 13 Goodness-of fit testing of the BPNN for web quality scale ... 59

Table 14 Results of factor classification and improvement rate ... 61

Table 15 Recovery solutions identified by expert interview ... 62

Table 16 Total direct-relation matrix from DEMATEL of C-company ... 64

Table 17 Total direct-relation matrix from DEMATEL of Y-company ... 65  

FIGURES

Figure 1 Research processes ... 5

Figure 2 A service recovery framework ... 9

Figure 3 Continuous service recovery management system ... 9

Figure 4 IPA grid ... 12

Figure 5 Kano’s two-dimensional model ... 13

Figure 6 Vavra’s importance grid ... 15

Figure 7 The house of quality ... 24

Figure 8 Linked houses convey the customer's voice through to manufacturing ... 25

Figure 9 Formation of independent and dependent management systems ... 31

Figure 10 Example of IRM ... 35

Figure 11 From service/product design to service recovery... 39

Figure 12 Conception of VGA approach ... 40

Figure 13 Algorithm logistic of VGA approach ... 41

Figure 14 Procedures of SFR model ... 47

Figure 15 The sample of the DEMATEL questionnaire for constructing RI matrix ... 63

Figure 16 Results of SFR analysis of C-company ... 67

Figure 17 Results of SFR analysis of Y-company ... 68

CHAPTER 1 INTRODUCTION Section 1 Research Background and Purpose

As service sector plays a decisive role in economic development, it is widely recognized that the success of the service sector is an essential factor in measuring an economy’s progress. In practice, service organizations are continuously endeavoring to improve their quality of service as it is of paramount importance to be competitive (Frei, Kalakota, Leone, & Marx, 1999; Saccania, Johanssonb, & Perona, 2007; Soteriou &

Zenios, 1999; Yieh, Chiao, & Chiu, 2007). This is such as Zeithaml, Bitner, and Gremler (2009) indicated that delivering superior service to decrease customer complaints and further increase customer satisfaction and value is critical to a firm’s sustainability.

According to the statistic of the global TARP Worldwide Inc. reported in 2007, when some types of failures occur, only a portion, 45% of customers actually complain to the employees serving them. In addition, a quick response to a service failure can only retain 49% of customers who experience a problem with service delivery. Thus, it is important to identify the service failures and then response to these failures to reach the target of service recovery or even prevention for businesses.

Eventually, numerous service failure identification approaches are proposed from

service quality perspective, and assume a linear relationship between the achievement of

service attributes and customer satisfaction. However, the seminal study by Kano, Seraku, Takahashi, and Tsuji (1984) thoroughly addressed the non-linear relationship between service attribute performance and customer satisfaction. And this viewpoint leads the assumption of linearity to be challenged and proven inaccurate (Kano et al., 1984;

Schvaneveldt, Enkawa, & Miyakawa, 1991). Thus, the integration of Kano’s two-dimensional theory with the service failure identification processes may be imperative

in service recovery (Chan & Lin, 2010).

On the other hand, each firm is constrained by the resources they have available, which may bring some challenges for managers on determining how resources are best deployed for various service failure recoveries (Ba & Johansson, 2008; Matzler, Bailom, Hinterhuber, Renzl, & Pichler, 2004; Panizzolo, 2008). This argument means that identifying the service failures has to be considered as the antecedent when improving customer satisfaction, and aligning a firm’s available resources and the SFs is an extensive issue has to be conducted. According to Resource-Based View (RBV), firms gain sustainable competitive advantages by ensuring appropriate access to a bundle of idiosyncratic resources or capabilities having four characteristics, such as valuable, rare, inimitable, and non-substitutable (Barney, 1991; Eisenhardt & Martin, 2000; Wernerfelt, 1984). Mowery, Oxley, and Silverman (1998) further indicated that such resources/capabilities often are based on tacit knowledge and are subject to considerable uncertainty concerning their quality and performance. In other words, the same strategies that enable a firm to extract a sustainable advantage from its resources may become difficult for the other firm to transfer or apply them in their market transactions. This such as Zhuang and Lederer (2006) stated that RBV may be used to explain why firms might use the same technologies with different results. Referring to the RBV viewpoint, the research question here is that how the firm uses appropriate resources or capabilities for service failure recovery? Especially in a market environment with limited resources, it is difficult for firms to recovery all the failures with the same level of completeness (Ba &

Johansson, 2008; Panizzolo, 2008).

Fortunately, quality functional deployment (QFD) which was originally used to ensure that the customer voice is systematically deployed throughout all the stages of product planning and designing is a useful decision making technique. The implementation

of QFD is through a matrix called the House of Quality which is constructed by several components with their respective scores, such as customer needs, customer assessment, technical requirement/description, relationship matrix of customer needs and technical requirements, correlation matrix of customer needs, technical matrix, etc. In which the technical matrix integrates different types of information for making decisions by utilizing the relative priority of each technical requirement in achieving the collective customer needs. Although QFD has been widely used in various areas due to its high practicality, recent researches (e.g., Partovi, 2007; Ramanathan & Yunfeng, 2009; Sener & Karsak, 2011) start to propose other QFD based techniques by integrating other methodologies in order to extract more reasonable scores for final decision making. Following this research trend, the purposes of this study are as follows:

1. To propose a service failures identification and service recovery model (SFR model) by applying the House of Quality of QFD.

(1) The valuable gap analysis (including performance function and valuable gap function) is proposed based on the redefinition of service quality gap through Kano’s two-dimensional theory with the application of neural network technique for service failures identification and ‘severity’ calculation.

(2) A transformation function (namely the KIR function) with the involvement of Kano’s two-dimensional theory is proposed for ‘improvement rate’ calculation of service failures.

(3) The DEMATEL methodology is adopted to understand the relationship between recovery solutions and service failures with the consideration of resources interdependence.

2. The SFR model is illustrated and validated by the empirical case on mobile telecom service industry.

(1) Two empirical cases are used to present the moderated effect of the resources the service companies have available on service recovery to evident the RBV.

(2) The empirical implications from the SFR analysis are provided for follow-up service recovery.

Section 2 Research Processes

According to the research purpose, the rest of this study is organized as follows.

Section 2 reviews the relevant literatures, including the introduction of service failures and service recovery, the relevant methodologies of service failures identification, and the discussion on RBV Theory, QFD, and Resource Dependence Theory. To elucidate the actual determinant of service failure recovery, section 3 introduces the novel SFR model proposed in this study. To this end, some advanced methodologies, such as valuable-gap-analysis, and KIR function are proposed, while the Decision-Making Trial and Evaluation Laboratory is adopted. To confirm the suitability and practicality of this novel SFR model, an empirical study was implemented in section 4 by using data collected from mobile telecom industry in Taiwan. Finally, section 5 draws the conclusions. The research processes of this study are shown as Figure 1.

Figure 1 Research processes

Empirical case study SFR model building

Service failures identification

Severity calculation Improvement rate calculation

Recovery solutions identification Interdependence

matrix Correlation matrix

QFD (is applied for connecting SERVICE FAILURE and SERVICE RECOVERY systematically)

Empirical results

Conclusions Literature Review Service failure /

recovery Kano theory QFD Resource

dependence theory RBV

Research purpose identification

First step: survey

Second step: interview Data collection

CHAPTER 2 LITERATURES REVIEW Section 1 Service Failure and Recovery

In a highly competitive environment, retaining existing customers takes on increased importance (Halstead, Morash, & Ozment, 1996). A common assumption of service marketing studies is that a satisfied customer has a greater propensity to engage in favorable behavioral intentions, such as repurchasing behavior and positive word-of-mouth, and a greater tolerance when experiencing a failure in the performance of the product or service (Bearden & Teel, 1983; de Matos, Rossi, Veiga, & Vieira, 2009; Richins, 1983).

Furthermore, Wong (2008) indicated that retaining and satisfying customers can be much less costly and more profitable than obtaining new customers. All of these arguments demonstrate the necessity of controlling the frequency of service failure occurrence.

Bitner, Booms and Tetreault (1990) stated that service failure occurs when a customer-requested service is not fulfilled, is un-reasonably delayed, or fails to reach the customer’s expected standard. Palmer, Beggs and Keown-McMullan (2000) defined service failure as a situation where customers find the service to be flawed and irresponsible. Maxham III and Netemeyer (2003) defined service failure as any service related misshape, whether real or perceived, that occurs while the customers has contact with the company in question. Furthermore, Zeithaml et al. (2009) describe service failure as service performance that falls below a customer’s expectation, and as service failure left unfixed can result in customer dissatisfaction and leaving. More details about service failure definition are summarized in Table 1.

Table 1

Summary of service failure definition

Author (year) Definition

Bitner et al. (1990) Service failure occurs when a customer-requested service is not fulfilled, is delayed, or fails to reach the customer’s expectations.

Zeithmal, Berry, and Parasuraman (1993)

Service failure involves activities that occur as a result of customer perceptions of initial service delivery behaviours falling below the customer’s expectations, or “zone of tolerance”.

Palmer et al. (2000) Service failure is a situation where customers find the service to be flawed and irresponsible.

Maxham III and Netemeyer (2003)

Service failure means any service related misshape, whether real or perceived, that occurs while the customers have complains with the company.

Zeithaml et al. (2009) Service failure represents the service performance that falls below a customer’s expectations.

Hedrick, Beverland, and Minahan (2007)

Service failure is commonly defined as a mistake, problem or error that occurs in the delivery of the service.

Sousa and Voss (2009) Service failure is defined as the real or perceived breakdown of the service in terms of either outcome or process.

Although service providers cannot avoid all service failures, they can learn how to

respond to service failures, which known as service recovery (Wang, Wu, Lin, & Wang,

2010) because customer loyalty may actually increase if the service problems are satisfactorily rectified in the retention case of recovery from service failures (Halstead et al., 1996). Kelley and Davis (1994) described that the service recovery comprises the actions that a service provider takes to respond to service failures and the process by which the firm attempts to rectify the failures. Similarly, Johnston and Hewa (1997) defined the service recovery efforts as “the actions of a service provider to mitigate and/or repair the damage to a customer that results from the provider’s failure to deliver a service as is designed” (p. 467). Furthermore, de Matos et al. (2009) stated that service recovery encounters are considered critical “moments of truth” in the relationship between service

provider and customers (Grönroos, 1988), that will be critical both for satisfying its customers and strengthening its relationships with them. In other words, favorable recoveries can work as a stimulus for customers to update their negative experience produced by the failure might have an influence on future assessments of satisfaction by the customers. In this study, service recovery is defined as the process by which service

providers attempt to correct a service failure in order to turn customers’ dissatisfaction into satisfaction and thereby retain customer loyalty based on previous studies (e.g.,

Andreassen, 2000; de Matos et al., 2009; Hart, Heskett, & Sasser, 1990; Kelley & Davis, 1994; Johnston & Hewa, 1997; Maxham III, 2001).

Recently, endeavors to address issues surrounding service failure have centered mainly on the topic of service recovery (Choi & Mattila, 2008). According to a rich stream of research has been conducted in the area of service marketing, service recovery efforts lead to positive attitude toward satisfaction with the service encounter (e.g., Maxham III, 2001; McCollough, Berry, & Yadav, 2000; Webster & Sundaram, 1998) and increase customers’ propensity to return to the same service provider (e.g., Maxham III, 2001;

Sousa & Voss, 2009; Webster & Sundaram, 1998). In short, the relationship of service failure and recovery create is such as an exchange in which the consumer experiences the service failures and the firm attempts to remedy it in the form of a recovery (Smith, Bolton,

& Wagner, 1999). In their empirical study, Miller, Craighead, and Karwan (2000) further outlined a framework for examining the service recovery based on previous services marketing and management literatures. As shown in Figure 2, outcome measures related to customer satisfaction and retention, antecedents to successful/unsuccessful recovery, the phases of recovery, types of recovery activities, and the delivery of service recovery are identified as the key elements associated with recovery.

Figure 2 A service recovery framework

Note: From “Service recovery: A framework and empirical investigation,” by J. L. Miller, C. W. Craighead, and K. R. Karwan, Journal of Operations Management, 18(4), p.388.

Accordingly, understanding how to provide appropriate service recovery is important for the establishment or maintenance of sustainable customer relationships, and a well-designed service recovery must be regarded as a continuous and long-term improvement effort for service companies (as shown in Figure 3).

Figure 3 Continuous service recovery management system

Service design

Service failure occurs Failure prevention

A well designed detection mechanism

Failure control / recovery

Failure

identification Solution identification

Monitoring recovery Service

re-design Service re-design Service

re-design

Feedback

Patronage Service

recovery expectations

Service recovery

Follow-up service recovery Severity

failure of

Perceived service

quality

Psycho- logical aspect

Tangible aspect

Psycho- logical aspect

Tangible aspect Customer

loyalty Service

guarantee Speed of recovery

Front line empower- ment

Pre-recovery phase Immediate recovery phase Follow-up recovery phase

Loyalty satisfaction

retention

Service failure occurs Provider aware of failure Fair restitution provided to customer

Section 2 Identification of Service Failures

Eventually, numerous service failures (SFs) identification methods are proposed from

service quality perspective. This study classifies these methods into two folds: (1) linear

category; and (2) two-dimensional category, based on their different consideration of the possessed impacts of service failures on overall customer satisfaction.

1. SFs identification through linear perspective

As for the linear perspective, research assumed a linear relationship between the achievement of service attributes and customer satisfaction. The PZB model proposed by Parasurman, Zeithaml, and Berry (1985) was the representative that has been extensively applied in various fields (e.g., Lee & Chen, 2009; Pakdil & Harwood, 2005; Rohini &

Mahadevappa, 2006; Winch, Usmani, & Edkins, 1998). In PZB model, five kinds of gaps are used to present why service provider could not meet customers’ needs. Some extended implications of those five gaps for the PZB model are as follows.

1. Gap 1: It means the difference between consumer expectations and management perceptions of consumer expectations. The gap may emerge due to not knowing what customers expect.

2. Gap 2: It means the difference between management perceptions of consumer expectations and service-quality specifications. The gap may emerge due to not selecting right service design.

3. Gap 3: It means the difference between service-quality specifications and the service actually delivered. The gap may emerge due to not delivering in accordance with service standards.

4. Gap 4: It means the difference between service delivery and what is communicated about the service to consumers. The gap may emerge due to not matching performance

to promises.

5. Gap 5: It means the difference between customer expectations and perceptions. The gap may emerge due to not meeting customer’s expected service.

Since Gap 5 is resulting from the sum of degree and direction of Gap 1 to 4 (i.e., Gap 5 = f (Gap 1, Gap 2, Gap 3, Gap 4)), the following research has generally been concentrated on the discussion of Gap 5, which then named as gap analysis (GA), to help make service decisions. In short, if the service performance meets or exceed the customer’s expectations, the customer will be satisfied and perceive good quality to present that improving this service attribute is not a priority for increasing customer satisfaction; in contrary, the customer will be dissatisfied and perceive poor quality to present that this service attribute needs to be improved immediately (Holmlund, 2008).

On the other hand, Martilla and James (1977) proposed a graphical technique to further the development of marketing strategies based on the same materials of GA. By using the importance measure to represent the vertical axis, and the performance measure to represent the horizontal axis, a management matrix, namely importance-performance

analysis (IPA) grid, can be formed. Furthermore, these two axes divide the IPA grid into

four quadrants with the crosshairs setting at the measure of central tendency of importance and performance (Figure 4). The implication for different quadrants were illustrate below:

1. Quadrant I composed of high performance and high importance, and situated at the upper right area of the management grid. It indicated the key attributes that consumers will be more emphasized, and the enterprise performance has satisfied consumers, thus it corresponded to “Keep up the Good Work”.

2. Quadrant II composed of low performance and high importance, and situated at the upper left area of the management grid. It indicated the key attributes that consumers will be more emphasized, but the enterprise performance has not yet satisfied

consumers, thus it corresponded to “Concentrate Here”.

3. Quadrant III composed of low performance and low importance, and situated at the lower left area of the management grid. It indicated the key attributes that consumers will be less emphasized, and the enterprise performance has not yet satisfied consumers, thus it corresponded to “Low Priority”.

4. Quadrant IV composed of high performance and low importance, and situated at the lower right area of the management grid. It indicated the key attributes that consumers will be less emphasized, but the enterprise performance has already exceeded the consumer expectation, thus it corresponded to “Possible Overkill”.

Figure 4 IPA grid

Note: From “Importance-performance analyses,” by J. A. Martilla and J. C. James, Journal

of Marketing, 41(1), p.78.

It can be seen that IPA is a decision making model which depended on the relative relationship between the importance and satisfaction with drafting the service attributes to

Extremely importance

Slightly importance

Excellent performance Fair

performance

Keep up the good work

Q I Concentrate here

QII

QIV

Possible overkill QIII

Low priority

describe their degree of failures. In consideration of its convenience and effectiveness, IPA model has been broadly applied to various fields.

2. SFs identification through two-dimensional theory

The study proposed by Kano et al. (1984) was the first to thoroughly address the

non-linear relationship between quality attribute performance and customer satisfaction.

They suggested that quality attributes can be classified into five categories (i.e., basic, performance, excitement, indifference, and reverse) based on the different influence of individual attributes on overall customer satisfaction. According to Kano model (Figure 5), the influence of performance factor and reverse factor on customer satisfaction is linear. In contrast, the influences of both excitement factor and basic factor are non-linear.

Figure 5 Kano’s two-dimensional model

Prior to the development of Kano model, all quality attributes were treated as performance factor with linear contribution to customer satisfaction. However, this conception was challenged by Kano’s two-dimensional theory, indicating that traditional

Satisfied

Excitement factor

Performance factor

No Satisfied

With quality attribute Without quality

attribute

Basic factor

Reverse factor Indifference factor

service quality proposed based on assumption of linear relationship does not seem to reflect service quality performance correctly (Matzler & Sauerwein, 2002; Matzler et al., 2004). Therefore, it is critical to consider the linear and nonlinear contributions of service practices on improving customer satisfaction while clarifying potential service failures.

In substance, the relevant articles of Kano theory can be organized into two groups: (1) factor classification of quality attributes, and (2) application or extension of Kano theory.

Following are the introduction based on these two research trends.

(1) Factor classification of quality attributes

The producing of Kano’s two-dimensional theory leads to a necessity of clarifying quality factors of attributes in order to extract more serious service failures. The use of

functional (with positive questions) and dysfunctional (with negative questions) questionnaires proposed by Kano et al. (1984) is the pioneer for quality factor

classification. In terms of this approach, respondents are asked to select one of the following five options: (1) I dislike it that way; (2) I can live with it that way; (3) I am neutral; (4) It must be that way; (5) I like it that way. Then, the quality attributes can be classified into one of six categories based on Kano et al.’s classification scheme (Table 2).

Table 2

Kano et al’s classification scheme

Customer requirement Dysfunctional

Like Must-be Neutral Live with Dislike

Functional

Like Q A A A O

Must-be R I I I M

Neutral R I I I M

Live with R I I I M

Dislike R R R R Q

Notes: A = Attractive; O = One-dimensional; M = Must-be; I = Indifferent; R = Reverse; Q = Questionable

In view of its over-complex and difficult to implement, Brandt (1988) developed the

dummy-variable regression method to simplify the categorization of quality attributes.

Under this method, the level of performance of each quality attribute is coded as a pair of dummy variables , , where = low performance and = high performance.

Then a regression model is formulated by using customer satisfaction as the dependent variable and two levels of performance of quality attribute as the independent variables. By doing so, quality attributes are classified according to the direction and level of significance of the strength of the regression coefficients.

Figure 6 Vavra’s importance grid

Note: From “Improving Your Measurement of Customer Satisfaction: A Guide to Creating,

Conducting, Analyzing, and Reporting Customer Satisfaction Measurement Program,” by

T. G. Vavra. Milwaukee, WI: ASQ Quality Press.

Additionally, Vavra (1997) turned to apply the “Grid” to classify product/service attributes. This method assumes every quality measure has explicit importance and implicit importance. And the intersection of these two importance measures purports to differentiate three quality factors. As shown in Figure 6, Basic factor (quadrant IV) are described as factors that respondents consciously acknowledge as being important (high

High explicit importance

Low explicit importance

Low implicit importance High implicit importance

Quadrant IV Basic factor Quadrant II

Excitement factor

Quadrant I Performance factor

Quadrant III Performance factor

explicit importance) but do not necessarily have a significant impact on feeling satisfied with a product experience (low implicit importance); in contrary, Excitement factors (quadrant I) are those attributes with high implicit importance, although respondents do not seem to consciously recognize their importance; finally, attributes falling into quadrants II and III to reflect characteristics of both are classified into the Performance factor.

By comparing the dummy regression analysis with Grid method subsequently, several studies supported the capability of the former one on producing more accurate results studies (e.g., Busacca & Padula, 2005; Matzler & Sauerwein, 2002) duo to the application of Grid method underlines the assumption that the implicit importance is based on linear estimates of the attribute performance-satisfaction relationship. However, Lin, Yang, Chan, and Sheu (2010) indicated that despite its simplicity and the superiority over the grid method, dummy regression has two major flaws: (1) the exclusion the information associated with the common level of performance; (2) the possible distortion resulted from the sample skewness toward one of the two extreme performance levels. Thus, they proposed a new approach with two-step procedure by applying moderated regression

method produce more accurate attribute classification, which are summarized as follows.

1. (STEP 1) Examining the moderated effect of “perception level” on the relationship between attribute performance and customer satisfaction.

(1) If the moderated effect does not exist, the subject attribute is considered to be a performance or inverse factor;

(2) Otherwise, go to next step to determine whether this attribute is an excitement or a basic factor.

2. (STEP 2) Using the sign of the regression coefficient to classify the subject attribute as an excitement factor or basic factor.

(1) If the coefficient of interaction effect is statistically significant and positive, this

attribute can be defined as an excitement factor because the attribute performance, at a high perception level, has greater impact on overall satisfaction than at a low or common perception level;

(2) On the other hand, the statistically significant and negative coefficient of interaction effect indicates that the attribute performance has greater impact on overall satisfaction at a low perception level than at a high or common level. This attribute would, therefore, be considered a basic factor.

Overall, each classification method has its fitness under respective contingency. Table 3 summarizes the comparisons between different classification methods.

Table 3

The comparisons between different classification methods

Classification Methods Author (year) Disadvantage 1. Functional and dysfunctional

questionnaires

Kano et al. (1984) ^z over-complex and difficult to implement

2. Dummy regression method Brandt (1988) z the exclusion the information of the common level performance

z the possible distortion resulted from the sample skewness 3. Importance Grid Vavra (1997) ^z the implicit importance is based

on linear estimates

4. Moderated regression method Lin et al. (2010) z the significance of the moderated effect is extensively based on the sample size

z the application has followed the condition of parametric statistics

(2) Application and extension of Kano theory

The Kano theory was originally applied in understanding customer needs for quality ensuring, while it has recently been widely adopted to discuss the customer satisfaction

strategies due to its additional consideration of non-linear effects possessed by quality attributes. Additionally, numerous literatures tend to proposed more advanced methodologies from Kano theory to draw out more elaborating optimal marketing strategy.

For example, Matzler and Hinterhuber (1998) proposed a methodology, based on Kano theory, to ensure a better understanding of customers’ needs and requirements, as well as procedures and processes to enhance communication by focusing on the voice of the customer within a product development project. In order to achieve the customer satisfaction in an effective way, Tan and Shen (2000) proposed an integrative approach by incorporating the Kano theory into QFD, in which an approximate transformation function is proposed to adjust the improvement ratio of each customer requirement based on the Kano theory analysis. Moreover, Shahin (2004) used the Kano theory as an advanced technique for customer satisfaction/dissatisfaction evaluation and then integrated it with the failure mode and effect analysis (FMEA) in order to overcome the limitation the FMEA inherited, which is that the severity rates are determined only with respect to company’s point of view, as well as to make it customer oriented.

Similarly, Chen, and Chuang (2008) proposed a design approach by incorporating the Kano theory in order to obtain the optimal combination of design form elements. In their robust approach, the Kano theory is used to understand the relationship between performance criteria and customer satisfaction, and to resolve trade-off dilemma in multiple-criteria optimization, while grey relational analysis with the Taguchi method is adopt to optimize subjective quality with multiple-criteria characteristics. On the other hand, to cope with the vague nature of product development processes, Chen and Ko (2009) proposed fuzzy nonlinear programming model based on Kano theory to determine the fulfillment levels of parts characteristics of QFD. Then, to deal with the design risk, the fuzzy-FMEA was incorporated into QFD processes.

Additionally, some Kano advanced approaches are proposed for customer satisfaction improvement through service quality viewpoint. For example, to avoid the misleading from linear assumptions as well as to consider the nature of fuzziness in human perception, Deng and Pei (2009) presented a fuzzy neural based IPA in identifying critical service attributes for service quality and customer satisfaction improvement. Furthermore, to provide a more accurate indicator to the manager about the importance and priority of the service development, Lee and Chen (2009) integrated Kano theory with QFD to categorize service attribute and proposed a revised improved ratio to strengthen the sensitivity of the customer satisfaction improvement, then the service gaps are evaluated through the revised GA approach. Table 4 provides a more elaborate summary of relevant empirical literatures of Kano theory.

Table 4

Summary of application and extension of Kano theory

Author (year) Application domain Advanced methodology Matzler and

Hinterhuber (1998)

product development Kano theory is applied to explore customers’

stated needs and unstated desires and to resolve them into different categories. Then, this categorization was used as a basis for QFD implementation for product development.

Tan and Shen (2000) web pages design Kano theory is incorporated into QFD to adjust the improvement ratio of each customer requirement.

Shahin (2004) service/product development

A new approach is proposed to enhance FMEA capabilities through its integration with Kano theory to involve the customers viewpoints.

Füller and Matzler (2007)

innovation process on new product development

Kano theory is reviewed to explain why it is difficult for customers to express their latent needs as well as those which are taken for granted.

Table 4 (con.)

Author (year) Application domain Advanced methodology Yadav and Goel

(2008)

product development (automatic industry)

The Kano theory is used to gain a profound understanding of vehicle attributes that are capable of achieving higher customer satisfaction if improved further.

Chen and Chuang (2008)

mobile phone design A robust design approach is presented by incorporating the Kano theory with grey relational analysis to enahnce product quality and customer satisfaction with multiple-criteria characteristics.

Arbore and Busacca (2009)

customer satisfaction analysis (retail

banking)

A revised methodology is proposed by the intuitions of the Kano theory to understand the non-linear and asymmetric relationship between attribute performances and overall customer satisfaction.

Chen and Ko (2009) new product development

To cope with the vague nature of product development processes, Kano theory and FMEA are enhanced by fuzzy approaches and then incorported into QFD processes.

Deng and Pei (2009) service quality improvement (hot

spring hotel)

A Fuzzy-IPA integrating fuzzy set theory, back-propagation neural network and Kano theory is proposed in identifying critical service attributes for customer satisfact improvement through service quality.

Lee and Chen (2009) new service development from quality perspective

Kano theory is involved into improved ratio calculation, then the service gaps are evaluated by revised GA approach. All the information are finally integrated into QFD implemention.

Li, Tang, and Luo (2010)

mature-period product improvement

(washing machine)

Kano theory is incorporated into customer competitive analyses to help accurately and deeply understand the nature of the customer requirements.

In short, previous researches focus on proposing advanced methods based on Kano theory for customer satisfaction management towards service quality management (e.g.,

Deng & Pei, 2009) and (new) product development/design (e.g., Chen & Chuang, 2008;

Matzler & Hinterhuber, 1998; Shahin, 2004; Tan & Shen, 2000). Relevant literatures applying the Kano theory in service recovery are still lack. This gap motivates the

necessary of including the potential two-dimensional influences of service failures when implementing the service recoveries. Additionally, the more important argument here is

that how resources are best deployed for service failure recovery for a firm with limited resources available. In the following section, the RBV is reviewed to response the necessity for discussing this argument.

Section 3 Resource Based View Theory

The emphasis of Resource Based View (RBV) is on internal resources available and developed within the firm, rather than those acquired externally (Penrose, 1959). The resource based perspective of the firm states that the firm’s strategy and success is based on its resource profile (Coates & McDermott, 2002), in which resources are not homogeneous within firms and provide unique services or abilities. Similarly, Barney (1991) published an influential article and put forward two fundamental assumptions for RBV theory: (1) resources and capabilities are heterogeneously distributed among firms and (2) resources are imperfectly mobile (Wong & Karia, 2010). In other words, the resources obtained by the firms are the source of economic rents and may be bundled. The bundling of resources is then referred to as capabilities which are linked to competitive advantage when they are a source of abnormal profits (Wernerfelt, 1984).

Competitive advantage, from the RBV perspective, is achieved by focusing on and exploiting the firm’s internal characteristics, specifically its resource profile (Rumelt, 1994;

Hamel & Prahalad, 1994). Adopting this perspective, therefore, this study emphasizes the necessity for defining the available resources and core competence, as well as the characteristics of these resources that make mobile telecom service operator strategically

relevant. In this study, a competence in the RBV is defined as “A bundle of assets,

capabilities, processes and knowledge that the firm performs better than its competitors, and is difficult to imitate and provides an advantage in the marketplace”. Such a definition

is consistent with Penrose (1959), which includes both tangible and intangible assets as firm resources. In the resource and competence based perspective these resources must meet the following conditions:

1. The resources are difficult to imitate: The imitability criterion is concerned with considering the ease with which competitors can replicate a valuable and rare resource possessed by a firm (McIvor, 2009);

2. The resources are rare: The rarity criterion is related to the number of competitors that possess a valuable resource. Clearly, when a number of competitors can easily posses a valuable resource, then it is unlikely to be a source of competitive advantage (McIvor, 2009);

3. The resources are valuable: Resources and capabilities are considered valuable if they allow an organization to exploit opportunities and counter threats in the business environment (McIvor, 2009);

4. The resources are non-substitutable.

Additionally, Barney (1991) argues that a firm must have capabilities of exploiting its resources and capabilities by organizing the organization criterion, such as the reporting structure, management control systems and compensation policies, etc (McIvor, 2009). In short, how to leverage resources in creating and sustaining competitive advantage for a firm has become the central focus for marketing scholars that link various types of market-based assets and capabilities with the performance of a firm (Wu, Yeniyurt, Kim, &

Cavusgil, 2006). Thus, the QFD which is a planning and problem-solving tool by translating customer requirements into technical characteristics is illustrated in the next

section.

Section 4 Quality Function Deployment

Quality function deployment (QFD) was invented in the late 1960s at the Kobe

shipyard of Mitsubishi Heavy Industries. As it evolved it become a well-known planning methodology for translating customer needs into relevant technical requirements (Govers, 1996; Hauser & Clausing, 1988). Generally, the QFD is implemented through a matrix called the House of Quality (HoQ) which is a kind of conceptual map that provides the means for interfunctional planning and communications (Hauser & Clausing, 1988). Figure 7 shows the elements for constructing HoQ which are introduced as follows.

1. Customer requirements: The construction of the HoQ starts with the determination of the customer requirements (WHATs) which is on the left side of the HoQ. To deploy available resources effectively, all WHATs must be rated against each other to quantify their importance to help set priorities for the product or service development process and provide guidelines.

2. Customer assessment: On the right hand side of the HoQ is the customer assessment part, which contains information on the customer’s perception of the company’s performance compared to competitor’s performance.

3. Technical requirements: The upper side of the HoQ is the technical requirements (TRs/HOWs) part, which gives a technical description of how to realize the consumer requirements.

4. Relationship matrix: The centre part of the HoQ depicts the relationship and strength between each WHAT and TR. This relationship matrix also can be used to indicate that a WHAT has inadequately been translated into a TR.

Figure 7 The house of quality

Note: From “Incorporating cost and environmental factors in quality function deployment using data envelopment analysis,” by R. Ramanathan and J. Yunfeng, Omega, 37(3), p.713.

5. Technical correlation matrix: The correlation matrix, put in the roof of the HoQ, is used to specify the various TRs that have to be improved collaterally, providing a basis to calculate to what extent a change in one feature will affect other features (Karsak, Sozer, & Alptekin, 2002). In other words, its use is to show whether trade-off decisions have to be made.

A. Customer Requirements

CR1 CR2 CR3 CR4 CR5 CR6

Importance of WHATs

♁

♁

△

◎

◎

◎

◎

△

♁

♁

♁

△

◎

◎

◎

△

♁

TR1 TR2 TR3 TR4

C. Technical Requirements

B. Competitive Bench-marketing

E. Technical Correlation Matrix

Absolute importance Relative importance

Rank Cost

Relative importance/cost Rank

F. Technical Matrix

D. Relationship Matrix of WHATs and TRs

♁ Strong relationship (9)

△ Medium relationship (3)

◎ Week relationship (1)

6. Technical matrix: The bottom of the HoQ contains different types of information for making decisions by utilizing the relative importance of each TR in achieving the collective WHATs. The derived importance score of WHATs are then multiplying the weightings in the cells of each row of the interrelationship matrix. Further, the cell scores are summed down each column to assign a relative priority to each TR, and the TR with the greatest score is allocated the highest priorities.

Eventually, to convey the customer's voice through to manufacturing, a serious of houses has to be linked (Hauser & Clausing, 1988). As shown in Figure 8, this process continues to a third and fourth phase as the “HOWs” of one stage become the “WHATs” of the next.

Figure 8 Linked houses convey the customer's voice through to manufacturing

Note: From “The House of Quality,” by J. R. Hauser and D. Clausing, Harvard Business

Review, 66(3), p.73.

Even if QFD is originally used by manufacturing organizations to assist in obtaining a balance between customer needs and the organizations’ actual ability to fulfill the requirements, recent researches started to apply QFD in service industries. For example, on basis of the evaluation of existing service concept design methods, Simons and Bouwman (2006) developed an alternative service concept design method based on QFD to focus on achieving multi-channel synergy by using data from six cases in the insurance and telecommunication industry. Wang (2007) applied the QFD to assist management in

Product requirement

Key process operations IV

Engineering characteristics

Customer attributes I

Parts characteristics

Engineering characteristics II ^Part

characteristics III

Key process operations

understanding the voice of outside consumer and to better prioritize internal operations in China Airlines. To help analyze the imprecise and subjective problem information, Su and Lin (2008) present a systematic process based on QFD to resolve the problems in the e-commerce industry and develop a systematic procedure to overcome the discrepancies in the problem-formulating process. Moreover, Andronikidis, Georgiou, Gotzamani, and Kamvysi (2009) discussed the integration of quantitative techniques with QFD and promoted successful application of QFD in finical service organizations.

Additionally, to better address the practical implications, in some cases, QFD was adapted or applied together with other methods by recent studies. For example, by arguing there is a paucity of methodologies for deriving the relative importance of design requirements when several additional factors are considered, Ramanathan and Yunfeng (2009) adopted DEA to facilitate QFD computations and applied this new methodology to the design of security fasteners for a Chinese company. Lee and Chen (2009) in another way integrated Kano theory with QFD, and proposed a revised improved ratio to strengthen the rise in customer satisfaction.

In order to clarify the thread of recent QFD-based literatures, the “Abstract” search is employed by using the term “Quality Function Deployment” in ScienceDirect/Online (SDOL) database. 87 articles are found from 2007 to 2011. Retaining the articles from

Social Science related journals (Table 5), the meta-analysis scheme (referred to Appendix

A) proposed by Carnevalli and Miguel (2008) is then adopted for articles classification.

Overall, QFD has been confirmed as a useful technique for business to systematically translate customer requirement into available technical requirements, while its application on service recovery is rare. Thus, this study extends the QFD’s balance translating conception into service failure and service recovery issue to fit this research gap.

Table 5

Summary of QFD-based literatures

Author (year) Kind of study (T1) QFD application (T9) QFD advanced methodology

Partovi (2007)

F: a manufacturer in the painting and adhesive products segment of the

chemical industry

A1, A6: for process selection and evaluation for manufacturing systems

[C]The AHP will be used to determine the intensity of the relationship between row and column variables of each matrix, and to evaluate the cost associated with each decision.

[D]The ANP is used to determine the intensity of synergy effects among column variables.

[E]A cost-benefit analysis will be conducted in order to determine the best production strategy.

Wang (2007) E: China Airlines A1: to better prioritize internal service operations

[G]Original QFD is used to assist management in understanding the voice of outside consumer and to better prioritize internal operations.

Iranmanesh and Thomson (2008)

D: a loud speaker (example product)

A2, A3: to assists designers to allocate the correct expenditure on designing in

order to optimize product value

[B]To achieve the customer desired product attributes at minimum cost, a multi-objective programming model is proposed to determine design characteristics.

Yadav and Goel (2008)

A, D: simulation is used to generate the data in

automatic industry

A3, A6: for customer satisfaction driven quality improvement target

planning on vehicle development

[G]QFD and Kano model are included in the product development processes in order to identify and prioritize potential attributes for further.

Chin, Wang, Yang, and

Poon (2009) A, D

A2: to prioritize design requirements with both CRs and customers’

preferences taken into account

[G]An evidential reasoning (ER) based methodology is applied for synthesizing various types of assessment information provided by a group of customers and multiple QFD team members.

Table 5 (con.)

Author (year) Kind of study (T1) QFD application (T9) QFD advanced methodology

Zhang and Chu (2009)

A, F: a horizontal directional drilling

machine

A3: for product development

[C][E]A group decision-making approach incorporating with

“logarithmic least squares model” and “weighted least squares model” is proposed to aggregate the multi-format and multi-granularity linguistic judgments of decision-makers, in which Fuzzy set theory is utilized to address the uncertainty in the decision-making process.

Utne (2009) F: the Norwegian fishing fleet

A6: for fisheries management regarding sustainability/environment,

and for design of fishing vessels

[B][E]Eco-QFD is proposed by extending the scope of QFD, and combining with Life-Cycle Cost and Life-Cycle Analysis to evaluate environmental effects and costs in the system development process.

Ramanathan and Yunfeng (2009)

A, F: the major producers of security

fasteners in China

A3: to develop a more competitive product design

[E]Data envelopment analysis is suggested for deriving the relative importance of design requirements (DRs) when several additional factors (in this study, including cost and environmental impact) are considered.

Liu (2009)

A, F: a small-medium manufacturer in Taiwan

that produce double layered stainless

vacuums

A3: to apply the proposed E-QFD method in the process of designing and

developing the stainless thermos

[E]The original part deployment table is modified by adding “Bottleneck level of PCs” and “RPN value of PCs”; a fuzzy ranking method is applied to determine the bottleneck rankings, and then PCs are classified into several groups through a modified fuzzy clustering method.

Bottani (2009) F A1, A6: to identify the most

appropriate enablers to achieve agility

[A][B]By building two subsequent HoQs, competitive bases, agile attributes and enablers are directly linked.

[C][D]The whole scaffold exploits fuzzy logic to deal with linguistics judgments expressing relationships and correlations required in the HoQ.

Table 5 (con.)

Author (year) Kind of study (T1) QFD application (T9) QFD advanced methodology

Karsak and Özogul (2009)

F: a Turkish automotive parts manufacturer

A7: to develop a novel decision framework for ERP software selection

[A]AHP method is used to determine the relative importance weights of corporation’s requirements.

[C][D]The relationships between corporation requirements and ERP characteristics and among ERP characteristics are determined by fuzzy linear regression.

[E]A weighted zero-one goal programming model is proposed to determine the most suitable ERP system alternative.

Li, Tang, Luo, Yao, and Xu (2010)

A, F: a novel type of two-cylinder washing

machine

A2: to demonstrate the feasibility of the proposed approach

[B]A series of algorithms are proposed for acquiring the set of engineering characteristics used in product planning house of quality.

Li et al. (2010) F: fully automatic washing machine

A2, A3: for mature-period product improvement

[F]A new methodology is proposed to rate the final importance of CRs by the integration of (1) improved maximal deviation approach for dealing with the corporations’ performance estimations, (2) importance rating for CRs by AHP, and (3) importance ratings of achieving the targets of CRs on the basis of Kano model.

Chen and Ko (2010)

F: a semiconductor packing case of the turbo

thermal ball grid array

A2, A3: for new product development with the consideration of MEC concept

[C]Based on the means-end chain (MEC) concept, a set of fuzzy linear programming models are built to determine the contribution levels of each “HOWs” for customer satisfaction.

[E]The risk analysis, FMEA, is incorporated into the QFD process as the constraint in the models.

Table 5 (con.)

Author (year) Kind of study (T1) QFD application (T9) QFD advanced methodology

Lin, Cheng, Tseng, and Tsai (2010)

F: a Taiwanese original equipment manufacturing firm

A1, A6: to get a better understanding of a strategic operating decision area in

improvement of production process toward sustainability

[A][B]The ‘Whats’ question of environmental production requirements (EPRs) and ‘Hows’ problem of the sustainable production indicators (SPIs) are made.

[D]The crisp value of importance with interdependence of EPRs and SPIs are calculated by fuzzy-ANP.

Khademi-Zare, Zarei, Sadeghieh, and Owlia

(2010)

E: Iran mobile telecommunication

A1, A2: to develop the strategic actions to meet the users’ expectations

[A]Model I: Fuzzy-TOPSIS method is applied to prioritize CRs; Model II: AHP method is utilized to prioritize CRs.

[E]The cost and benefit are included in final strategic actions implementation.

Wang, Lin, and Huang (2010)

F: a drug development project in the pharmaceutical industry

A3, A6: to propose a

performance-oriented risk management framework for R&D project

[B]QFD is adapted to transform organizational performance measures into project performance measures.

[E]Utility theory is used to determine R&D risks.

[F]The balanced scorecard is used to identify major performance measures of an R&D organization.

Sener and Karsak (2011)

A, F: a washing machine in Turkish home laundry

appliances industry

A3: to improve product quality and increase customer satisfaction

[C][D]Fuzzy regression is introduced in the model to identify the functional relationships between CNs and ECs, and among ECs. Then considering the multi-objective nature of the product design problem, fuzzy multiple objective programming is included.

Note: [A] represent the modification on WHATs of HoQ; [B] represent the modification on HOWs of HoQ; [C] represent the modification on Relationship Matrix of HoQ;

[D] represent the modification on Correlation Matrix of HoQ; [E] represent the modification on Technical Matrix of HoQ; [F] represent the modification on Competitive bench-marketing Matrix of HoQ; [G] represent the application of the original QFD method.

Section 5 Resource Dependence Theory

In recent years, several studies indicated that not all the subjects will be able to completely understand the interrelationship between variables of a management system (e.g., Lee, Li, Yen, & Huang, 2010). As shown in Figure 9, as a service management system is implemented with the assumption of resource independence, more waste would be produced because of the overlook of synthesis effect or interactive relations between available resources. In other words, when some variables do not meet the independent assumption, decision makers will not be able to correctly analyze the service failures and their impacts on customer satisfaction improvement, which results in the wrong conclusion.

Thus, Interpretive Structural Modeling and Decision-Making Trial and Evaluation Laboratory, which are the most popular methods applied on clarifying the dependence/interrelationship among variables, are discussed in turns.

Figure 9 Formation of independent and dependent management systems

1. Interpretive structural modeling

Interpretive structural modeling (ISM) is a computer-assisted methodology proposed

by Warfield (1974) to construct and understand the relationships between the elements in

Management system Indicator / variable 1 Indicator / variable 2 Indicator / variable 3

….

Management system Indicator / variable 1 Indicator / variable 2 Indicator / variable 3

.. ..

(a) Independent system (b) Dependent system

Recovery performance Recovery performance

complex systems or situations (Huang, Tzeng, & Ong, 2005). The implementation of ISM is dependent on through individual or group mental feedback to calculate relation matrix to present the interdependences between the elements composing a management system.

Referring to Huang et al.’s (2005) study, the procedures of ISM are illustrated as follows:

The initial relation matrix is formed by asking the question like “Does the element influence the element ?” If the answer is “Yes”, then 1, otherwise 0. In other words, the initial relation matrix is equal to a binary matrix. Equation 1 presents the general form of the relation matrix:

0 0

0

(1)

where means the ith element in the system, denotes the direct impact of ith element on jth element. Thus, is the initial relation matrix.

After constructing the initial relation matrix, the reachability matrix, , need to be calculated by using Equation (2) and (3) as follows:

(2)

, 1 (3)

where is the unit matrix, and denotes the powers. Note that the reachability matrix is under the operators of the Boolean multiplication and addition, and is obtained by powering the matrix, , to satisfy the Equation (3).

Unfortunately, ISM based on using binary matrix just can be used for understand the

existence of the causal relationship between the elements, but not for the degree of the

impact of a certain element on the other ones. Thus, Decision-Making Trial and Evaluation

Laboratory method that can be used to present the impact of each element on others is

introduced in next section.

2. Decision-making trial and evaluation laboratory

Decision-Making Trial and Evaluation Laboratory (DEMATEL) was proposed by

Battelle Geneva Institute to demonstrate the causality between practices in a system by gathering collective knowledge. Referring to previous relevant literatures, the operational procedure of DEMATEL is summarized and introduced in the following steps:

Step 1: Building the initial average direct-relation matrix

Suppose is the number of experts consulted, and is the number of practices that each expert considers. The integer score refers to the degree that practice i affects practice j for the kth expert, where the scores 0, 1, 2, 3, 4 represent the range from ‘no influence’ to ‘very high influence.’ The average matrix is realized by averaging all the experts’ scores and can be represented mathematically by the following equation:

1 (4)

Step 2: Calculating the normalized direct-relation matrix

The normalized direct-relation matrix is obtained by normalizing the direct-relation matrix and can be represented mathematically by the following equation:

, where ∑ , ∑ (5)

Since the sum of each row i of matrix represents the direct effects that practice i gives to the other practices, and the sum of each column j of matrix represents the direct effects that practice i receives from the other practices, therefore, ∑ ,

∑ represents the direct effects of the practice with the most direct given and received effects on others.