University of Nottingham

Measuring Service Quality in Distribution Logistics Using SERVQUAL and AHP: A Case Study in a Pharmaceutical Wholesaler in Turkey

HARIKA KARPUZCU

MSc Operations Management

University of Nottingham

Measuring Service Quality in Distribution Logistics Using SERVQUAL and AHP: A Case Study in a Pharmaceutical Wholesaler in Turkey

by

Harika KARPUZCU

2006

A Dissertation presented in part consideration for the degree of “MSc Operations Management”

Contents List of Tables.............................................................................................................................iii List of Figures ...........................................................................................................................iii Acknowledgements ................................................................................................................... iv Abstract ...................................................................................................................................... v CHAPTER 1: Introduction ..................................................................................................... 1 1.1. SERVQUAL............................................................................................................ 2 1.2. AHP......................................................................................................................... 4 1.3 Research Objectives ................................................................................................. 6 CHAPTER 2: Literature Review............................................................................................ 9 2.1. Service Quality........................................................................................................ 9 2.1.1. Criticisms of SERVQUAL..................................................................... 11 2.2. Physical Distribution Service Quality ................................................................... 13 2.3. Analytic Hierarchy Process ................................................................................... 15 2.3.1. Criticisms of AHP .................................................................................. 16 CHAPTER 3: Research Methodology……………………………………………………...18 3.1. Focus Company..................................................................................................... 18 3.2. Sample Selection ................................................................................................... 19 3.3. Development of the Questionnaires ...................................................................... 20 3.3.1. Business Insight...................................................................................... 20 3.3.2. PWSQ..................................................................................................... 20 3.3.3 AHP model.............................................................................................. 26 3.4. Data Collection...................................................................................................... 28 CHAPTER 4: Data Analysis ................................................................................................. 30 4.1. AHP Results .......................................................................................................... 30 4.1.1. Point allocations ..................................................................................... 30 4.1.2. Sample calculations................................................................................ 31 4.1.3. P/E scores ............................................................................................... 33 4.1.4. Regression Analysis ............................................................................... 34 4.2. PWSQ Results ....................................................................................................... 36 4.2.1. Unweighted analysis .............................................................................. 36

i

4.2.1.1. P-E scores ................................................................................ 37 4.2.1.2. Regression Analysis ................................................................ 39 4.2.2. Weighted analysis .................................................................................. 39 4.2.2.1. P-E scores ................................................................................ 40 4.2.2.2. Regression Analysis ................................................................ 41 CHAPTER 5: Comparison of the Three Approaches......................................................... 42 5.1. Behavioural Comparison....................................................................................... 42 5.2. Numerical Comparison ......................................................................................... 44 5.2.1. Unweighted PWSQ versus weighted PWSQ ......................................... 45 5.2.1.1. Correlation analysis................................................................. 46 5.2.2. PWSQ versus AHP................................................................................. 47 5.2.2.2. Correlation analysis................................................................. 50 5.3. The Effect of Gender............................................................................................. 51 CHAPTER 6: Summary and Conclusions ........................................................................... 56 References ............................................................................................................................... 61 Appendices .............................................................................................................................. 68 Appendix A - The PWSQ questionnaire.......................................................................68 Appendix B - The AHP questionnaire..........................................................................72 Appendix C - The modified section of the AHP questionnaire……………………... 77 Appendix D - Sample AHP questionnaire....................................................................78 Appendix E - Sample PWSQ questionnaire.................................................................81

ii

List of Tables Table I – SERVQUAL dimensions …………………………………………………..............10 Table II – Allocation of items to dimensions ………………………………………………...21 Table III – References to papers on logistics service quality ………………………………..23 Table IV – PWSQ dimensions ……………………………………………………………….24 Table V – Sources of PWSQ dimensions ……………………………………………………25 Table VI – The AHP quality measurement scale …………………………………………….28 Table VII – A sample pairwise comparison matrix ………………………………………….28 Table VIII – The most and least important dimensions ……………………………………...31 Table IX – A sample response matrix ………………………………………………………..31 Table X – Sample P/E score calculation ……………………………………………………..32 Table XI – Category-based satisfaction scores for the seven dimensions …………………...33 Table XII – Distribution of satisfied and dissatisfied customers …………………………….34 Table XIII – Unweighted gap scores for the seven dimensions ……………………………..37 Table XIV – Distribution of satisfied and dissatisfied customers for unweighted PWQ analysis ……………………………………………………..39 Table XV – Category-based overall satisfaction scores…………………………………….. 45 Table XVI – Distribution of satisfied and dissatisfied customers for the two models…………………………………………………………………….. 48 Table XVII – Overall satisfaction percentages for the three analyses………………………..48 Table XVIII – Comparison of the three correlations ………………………………………...51 Table XIX – Sample data for unweighted analysis …………………………………………..52 Table XX – Sample data for weighted analysis ……………………………………………...53 Table XXI – Sample data for AHP analysis …………………………………………………54

List of Figures Figure 1 – Gaps model ………………………………………………………………………...3 Figure 2 – A decision hierarchy with k levels ………………………………………………...5 Figure 3 – Hierarchy of PW service quality ………………………………………………... 26 Figure 4 – Normal distribution curve for hypothesis testing ……………………………….. 52

iii

Acknowledgements There are many people to whom I am extremely grateful for their help and support during the preparation, research and execution of this dissertation and must therefore acknowledge. I would like to express my sincere appreciation to my supervisor, Dr. Ramakrishnan Ramanathan, for his helpful advice and comments throughout the preparation of this dissertation. This study would not have been such interesting and instructive for me without his invaluable instructions. Many thanks to Nevzat Pharmaceutical Wholesaler, Ankara for letting me to conduct this pilot study there and the management support that helped me a lot to develop the service quality measurement scales. Special thanks to customer services representatives for helping me to conduct the interviews with customers and obtain various data. Appreciation is also extended to all the pharmacists who spared their time to accept my interviews. The information they provided is valuable to this work. Loads of thanks to my mother and father for their never-ending love and support; I would not have accomplished a postgraduate study in England without their help and encouragement. I must also send thanks to all my family members for their love and caring while I were away from home. Last but not least, I am grateful for his love and words of encouragement; the beloved boyfriend, Bora, I would have never achieved anything in this hard year without you.

iv

Abstract Service quality has gained quite a lot attention in the last two decades from both academics and private sector, due to its significant influence on consumer perceptions of performance. Various tools have been developed to measure service quality accurately, yet there seems not to be a single universally agreed method that serves properly to the purpose. This study attempted to adapt SERVQUAL and AHP techniques for developing two instruments for service quality measurement in a physical distribution service setting. The tools are exercised by a sample of customers of the focus company and several results are obtained. Service quality attributes important for the business context are identified through detailed analysis of relevant literature and gathered into the service quality measurement tools. Sample responses are compared on various bases; responses to PWSQ (the SERVQUAL variation model) and AHP questionnaires are found to be quite different. Gap scores obtained from PWSQ instrument suggested that most of the customers were dissatisfied with the service quality, whereas only a small number of respondents seemed to be dissatisfied with the service quality as a result of AHP model analysis. The significant differences between the results of the two methods suggested that the approaches differed in terms of their capabilities in reflecting customer opinions accurately and depending on the various comparisons made, the AHP variation model is claimed to be a more proper tool in measuring service quality.

v

CHAPTER 1: Introduction Today’s business environment can be characterized by increased competition, uncertain economic conditions and shifts in global trading relationships (Kong and Mayo, 1993). Organizations are under great pressure to understand and satisfy the market requirements in order to remain in this competitive environment. As consumers become more and more demanding, it is only essential that companies meet this increasing demand, and even delight their customers by going beyond their expectations in order to survive in 2000’s (Clow and Vorhies, 1993; Kong and Mayo, 1993). Besides vast studies about total quality management and supply chain management in order to perform better on product-based issues, it is now realized that performing well in services is also very important for remaining competitive in business (Fung and Wong, 1998; Clow and Vorhies, 1993). Despite the fact that services, or customer service, are mostly intangible (performance or actions that cannot be seen, felt, tasted or touched in the same manner as tangible goods), heterogeneous (no two services can be precisely alike since services are generally performed by humans), produced and consumed simultaneously, and are perishable (services cannot be saved, stored, resold or returned) (Zeithaml et al., 2006), its importance in consumers’ overall evaluation of a company is significant. Richard and Allaway (1993) investigated the importance of service quality as a predictor of actual choice behaviour and, concluded that managers could emphasize certain service attributes in order to attract customers and influence their choice behaviour, since they identified a positive correlation between numerous service features and choice. Service quality has gained popularity and has become an important research topic in the last two decades, mainly due to its close relationship with customer satisfaction (Daugherty et al., 1998; Innis and La Londe, 1994; Stank et al., 2003). It has now been realized that companies (public or private sector, profit or non-profit every organization) can obtain increased business (Anderson et al., 1996; Zeithaml et al., 2006) and market share (Daugherty et al., 1998; Innis and La Londe, 1994; Stank et al., 2003) through maintaining loyalty among customers by increased customer satisfaction, which can only be possible by better customer service (Innis and La Londe, 1994; Stank et al., 2003). Zeithaml et al. (2006) point out to the importance of word-of-mouth communication: post-experience evaluations of service quality would significantly affect what customers tell others about the service; this means that customers who are satisfied with the service they received can easily influence other consumers to purchase that service, or avoid that particular service firm if they were dissatisfied. Alexandris et al. have shown in their 2002 study that customer perceptions of 1

service quality dimensions explain most of the purchase intentions of customers (Alexandris et al., 2002). Daugherty et al. (1998) explain the importance of achieving high levels of service quality as a company as follows:
“The intent is to keep customers so satisfied with the service that they remain loyal, repurchasing the product again and again, which is expected to have a positive effect on financial indicators, such as market share and profitability” (Daugherty et al.,

1998). Measuring customer satisfaction is therefore critical to understanding current situation and then to responding faster and better to customers than the competitors (Kong and Mayo, 1993). In order to improve customer satisfaction, it is essential that current service quality be measured first. Due to its intangibility and nonstandardized nature (Zeithaml et al., 2006), services are quite difficult to measure and mostly there is not any one particular method with which service quality can be measured accurately in every setting. There has been a growing body of literature in the last two decades on how to measure service quality and many researchers have developed different tools and methodologies for this purpose, however the variety of papers cause potential users a serious difficulty in choosing a proper tool for their particular needs (Franceschini et al, 1998). 1.1. SERVQUAL Many researchers in this area claim that service quality of an organization can best be measured through that organization’s customers’ points of view, as they are the ones receiving the service at the end of the day. The following words of Tucker might best describe the main drive for academics and companies to investigate how customers evaluate service quality:
“The key to customer service is understanding the customer and his perceptions. It doesn’t matter what a supplier does, but rather what customers think the supplier does in the area of customer service” (Tucker, 1980).

One of the most popular instruments for measuring service quality through customer opinions is named SERVQUAL, which was first developed by A. Parasuraman, Valarie A. Zeithaml and Leonard L. Berry in 1988. This instrument has its roots in a previous study by the authors. Parasuraman et al. published a conceptual paper in 1985 which introduced a framework for understanding five quality gaps, four on the service provider’s side and one on customer’s side. These gaps can be viewed in Figure 1.

2

Following are brief definitions of the five gaps in Parasuraman et al.’s model: Gap 1: Difference between customer expectations and the management perceptions of customer expectations. Gap 2: Difference between the management perceptions of customer expectations and service quality specifications. Gap 3: Difference between service quality specifications and the service actually delivered. Gap 4: Difference between service delivery and what is communicated about the service to customers. Gap 5: Difference between customer expectations and perceptions. (Grapentine, 1998)

Figure 1. Gaps model The focus of SERVQUAL is mainly the fifth gap, the difference between customer expectations and perceptions about service quality. Customer expectations are beliefs about service delivery that serve as standards or reference points against which performance is judged (Zeithaml et al., 2006). Zeithaml et al. (2006) point out to the importance of customer expectations as follows:
“Knowing what the customer expects is the first and possibly the most critical step in delivering quality service. Being wrong about what customers want can mean losing a customer’s business when another company hits the target exactly. Being wrong can also mean expending money, time, and other resources on things that do not count to the customer. Being wrong can even mean not surviving in a fiercely competitive market.” (Zeithaml et al., 2006)

Parasuraman et al. claimed that customers evaluate the quality of a service mainly by comparing their initial expectations with actual experience (1988); that is, they think the service quality was poor if the service company has performed worse than their expectations. Customers approve the service quality, however, if they receive a service that match with their expectations or can even be delighted if the company outperforms their initial expectations. Based on this gap theory (Parasuraman et al., 1985; Clow and Vorhies, 1993; Mehta and Durvasula, 1998), Parasuraman et al. (1988) have developed an instrument that questions

3

customer expectations and perceptions separately and then allows to measure the gaps, if any, between these two and conclude if the customer was satisfied or not. Consumers evaluate the quality of a particular service according to several dimensions; they compare the actual service and their initial expectations for different criteria such as responsiveness, courtesy and tangibles. SERVQUAL was first constructed around the ten dimensions identified by Parasuraman et al. (1985) along which consumers perceive and evaluate service quality. The dimensions were reduced to only five later on, being tangibles, reliability, responsiveness, assurance and empathy (Parasuraman et al., 1988). SERVQUAL essentially consists of two sections; the first section basically questions customers’ expectations from companies in general within a specific industry, and the second part measures customers’ perceptions of a particular company’s service in that business (Parasuraman et al., 1988). The 22 items in each part of the instrument are sub-items of the predefined five dimensions; the statement “XYZ has modern-looking equipment” is, for example, under the dimension tangibles. Corresponding statements in the two parts, for example, fifth question in the expectations part and fifth question in the perceptions part, refer to the same decision criteria. Utilization of the same scale in both parts enables the surveyor to directly measure the gap between one customer’s perceptions and expectations. The straightforward nature of SERVQUAL has made it popular for measuring service quality in both academic and business environments and enabled it to be utilized in many different settings. 1.2. AHP Another technique that has been used, although less common, to measure service quality is the analytic hierarchy process (AHP). The AHP technique, first developed by Saaty (1980, 1990), is a powerful and reputable tool generally used for multiple criteria decision making purposes (Vaidya and Kumar, 2006; Chow and Luk, 2005). It uses a process of pairwise comparisons to determine the relative importance (and thus the priority) of alternatives in a multi-criteria decision-making problem. AHP involves decomposing a complex and unstructured problem into a set of variables that are organized into a hierarchy (Chow and Luk, 2005). At the top of the hierarchy is the most macro decision objective, such as the objective of selecting the best inventory management software package for a company. The lower levels of the hierarchy contain attributes that contribute to the decision, such as the price, speed, complexity and applicability of the software packages. Details of these attributes

4

may increase at the lower levels of the hierarchy. The final level in the hierarchy has the alternatives being considered, different inventory management software packages, in this case. The decision maker would have to consider each alternative in terms of the attributes in hierarchy and choose the alternative that has the highest overall rating. A standard form for the AHP hierarchy is depicted in Figure 2.

Figure 2. A decision hierarchy with k levels (Zahedi, 1986) AHP basically enables decision-makers to choose from a number of alternatives by formulating priorities and making a series of tradeoffs. The logic of the AHP tool can be described in the following five steps: (1) Break down the criteria under consideration into a manageable number (5-8) of subcriteria and attributes. Then structure these criteria and attributes in hierarchical form, (2) Make a series of pairwise comparisons among the sub-criteria according to their importance for the overall aim being examined, (3) Estimate relative weights for the sub-criteria and attributes based on customer comparisons, (4) Get customer ratings for each alternative on the performance of each sub-criteria, (5) Aggregate weights with performance scores for each alternative and synthesize them for measurement of overall performance. (Saaty, 1990; Ahmed and Rafiq, 1998; Min and Min, 1996)

5

The literature on using the AHP tool for service quality measurement purposes is scarce; most of the present studies on this subject had the objective of benchmarking, involving at least two companies (Chow and Luk, 2005; Min and Min, 1996; Min and Min, 1997; Min et al., 2002). This paper would contribute to the present literature by addressing the gap in utilization of analytic hierarchy process for measuring the service quality of a single organization. 1.3 Research Objectives The purpose of this study is threefold: 1) The primary objective is to measure the service quality of a Turkish company in distribution logistics sector; 2) Another objective is to modify the SERVQUAL instrument (Parasuraman et al., 1988) to obtain a better fit with the specific industry in question and develop a new service quality measurement tool, PWSQ; 3) Because SERVQUAL has its own criticisms, the last objective of this study is introducing AHP as an alternative methodology for measuring service quality that compensates some of SERVQUAL’s shortcomings. The study is performed within the pharmaceutical wholesaling industry, which can be considered as a sub-sector of distribution logistics. The study is conducted at a focal company in Turkey that is well-known and experienced in the business. The author interviewed a large sample of customers of the focal company and presented them modified versions of SERVQUAL and AHP tools in order to get their opinions about the company’s service quality. The reason why physical distribution was selected as the research area is that service quality in logistics has been a subject covering a large portion of businesses around the world; suppliers have to distribute raw materials to factories, manufacturers have to distribute finished goods to central or regional distribution centers, and those centers have to distribute goods to retailers, and so on. As such, physical distribution and logistics constitute an important part of a majority of businesses and high quality performance in these services gains even more importance as consumers in general have been increasing expectations. Various features of physical distribution service can provide a competitive advantage by differentiating companies with superior levels of service (Xing and Grant, 2006). According to Mentzer et al. (2001), the ability to deliver the right amount of the right product at the right 6

place at the right time in the right condition at the right place with the right information (Shapiro and Heskett, 1985) is crucial in providing satisfactory customer service (Mentzer et al., 2001). Daugherty et al. (1998) link better distribution service to increased customer satisfaction, and try to show that increased customer satisfaction would bring increased loyalty and larger market share in return (Daugherty et al., 1998). This implies that distributors have to increase their service performance continuously in order to remain competitive among rivals, and it is only necessary that they measure their current service quality first. SERVQUAL has been used for measuring service quality in distribution logistics industry in numerous studies (Mehta and Durvasula, 1998; Mentzer et al., 1999; Bienstock et al., 1997). Researchers generally modified the instrument according to the specific needs of the specific industry and this way obtained a more reliable tool for understanding customer satisfaction. A similar approach will be taken in this paper such that the original SERVQUAL instrument (Parasuraman et al., 1988; 1991) will be modified in terms of its dimensions in order to consider the most important issues for customers in evaluating physical distribution service quality. Utilizing AHP for distribution logistics service quality measurement will be relatively new in the research area; however the results of the study would easily be used for not only distribution logistics area but customer services in general. After applying both tools on a selected sample of customers, the author will compare he results on behavioural and numerical bases and make suggestions about which tool is more appropriate in service quality measurement. This study will contribute to previous research on service quality measurement in the sense that it will suggest a tool customized for measuring physical distribution service quality. In addition, this will be the first study in the literature to apply logistics service quality measurement techniques to a company in Turkey. This paper is organized as follows: In the second chapter a literature review is provided covering past studies on measuring service quality and distribution logistics service quality, as well as explaining how SERVQUAL and AHP had been utilized for these purposes before. In addition, major criticisms and advantages of the SERVQUAL and AHP tools are going to be addressed towards the end of Chapter 2. Next, the research methodology of the study will be introduced in Chapter 3, including details about the focus company, sample selection and development of service quality measurement instruments. Once all data are collected from the 100 selected customers, 3 sets of results will be analyzed in Chapter 4 in terms of customer satisfaction; unweighted and weighted SERVQUAL variation model analyses, as well as the

7

AHP variation model analysis will take place in this part. The analyzed data are to be compared on behavioural and numerical bases in Chapter 5 and inferences will be made regarding which instrument would be more appropriate for measuring service quality. The last chapter of the paper will include concluding remarks from the study as well as identification of the study’s limitations and opportunities of future research.

8

CHAPTER 2: Literature Review This section reviews the relevant areas of research associated with the measurement of service quality. The literature review includes studies about development of a measurement scale for service quality perceptions, and for physical distribution service quality measurement, in particular. 2.1. Service Quality Since it has been realized that achieving higher service quality would lead to increased customer satisfaction, and that would yield an increase in business, there has been a growing interest for the research field of measuring service quality. Because services are heterogeneous across time, organizations and people, ensuring consistent service quality is challenging (Zeithaml et al., 2006). According to Zeithaml et al. (2006), service supplier may not always fully control the quality of service, since it is dependent on many factors such as the ability of the customer to express his or her needs, the ability and willingness of service personnel to satisfy those needs, or the presence (or absence) of other customers in the setting. However, despite the difficulty to measure service quality accurately and precisely, academics have written a large number of research papers on how to measure service quality. Some researchers attempted to measure a company’s service quality internally (Novack et al., 1994), that is, by questioning the executives and personnel of the company itself; however there seems to be a consensus around the idea that service quality of a company can best be understood and measured through the opinions of that particular company’s customers (Parasuraman et al., 1985; Zeithaml et al., 1990). As mentioned above, SERVQUAL is one of the most famous tools for measuring service quality from customer’s perspective. It is founded on the view that customers assess a particular company’s service quality depending on the gap between what they expect from a class of service providers (say, all cargo carriers), and their evaluations of the performance of that particular company (say DHL) (Buttle, 1996). Parasuraman et al. identified ten components of service quality in their 1985 work: (1) (2) (3) (4) (5) Reliability, Responsiveness, Competence, Access, Courtesy,

9

(6) (7) (8) (9)

Communication, Credibility, Security, Understanding/knowing the customer,

(10) Tangibles. Then, in 1988, they constructed a service quality measurement scale, SERVQUAL, and reduced the number of dimensions to five as a result of scale purification studies. All the important items in SERVQUAL were gathered into the dimensions tangibles, reliability, responsiveness, assurance and empathy (Parasuraman et al., 1988), as shown in Table I.
Dimensions Reliability Assurance Tangibles Empathy Responsiveness Definition The ability to perform the promised service dependably and accurately The knowledge and courtesy of employees and their ability to convey trust and confidence The appearance of physical facilities, equipment, personnel and communication materials The provision of caring, individualized attention to customers The willingness to help customers and to provide prompt service Items in scale 4 5 4 5 4 (Buttle, 1996)

Table I. SERVQUAL dimensions Finally, as a follow-up study Parasuraman et al. refined their SERVQUAL scale in terms of wording of the statements; researchers thought that evoking unrealistically high expectations by statements such as “Telephone companies should keep their records accurately” would be misleading and, instead of aiming at customers’ normative expectations, the authors decided to direct attention to what customers would expect from companies delivering excellent service. As a result they changed the wording of the above sentence and the rest as “Excellent telephone companies will insist on error-free records” (Parasuraman et al., 1991). In addition to this modification, authors changed the wording of some statements in order to make all of them positively worded, instead of some being negatively worded and affecting the variation and reliability of the scale. Finally, a point-allocation part in which customers would allocate a total of 100 points to the service quality dimensions according to each of their importance was attached at the end of the instrument so that it would be possible to combine them with the individual attribute ratings to obtain a composite, weighted estimate of overall service quality (Parasuraman et al., 1991). The finalized version of SERVQUAL can be found in the work of Parasuraman et al., 1991.

10

SERVQUAL has been utilized quite many times by different researchers for the purpose of measuring service quality and understanding how customers perceive and evaluate a particular service. Among the different industries that SERVQUAL was applied to are health care services (Youssef et al., 1995; Vandamme and Leunis, 1993; Babakus and Mangold, 1992), tourism (Gabbie and O’Neill, 1996; Juwaheer, 2004; Alexandris et al., 2002), education (Sahney et al., 2004; Pariseau and McDaniel, 1997), banking (Chiu et al., 2003; Angur et al., 1999), retailing (Long and McMellon, 2004; Gagliano and Hathcote, 1994), and police services (Donnelly et al., 2006). An extended list of the application areas of SERVQUAL can be obtained from the work of Buttle (1996). Through the utilization of the SERVQUAL instrument, various companies learned their customers’ expectations of a service as well as their opinions about the service they provide, and aimed at improving the particular dimensions with which customers were dissatisfied (Zeithaml et al., 2006). North Lanarkshire Council utilized SERVQUAL in a public services context; internal customer opinions were gathered in catering and ground maintenance services, revealing that the gaps were negative (expectations higher than perceptions) for all the 5 dimensions. The resultant quality gaps in both services formed the basis for future service developments (Brysland and Curry, 2001). Zeithaml et al. (2006) point out to the study performed in a large U.S. manufacturing company where the SERVQUAL instrument was administered each year in order to assess the impact of corrective actions taken as a result of previous years’ dissatisfaction scores. Although it was only a pilot application conducted for a single multinational customer of the company, it became an ongoing activity in the company due to the success of the SERVQUAL survey and its implementation was expanded to other divisions and customer groups later on (Zeithaml et al., 2006). 2.1.1. Criticisms of SERVQUAL Although very popular, SERVQUAL is not without criticisms. One of the most common issues raised by academics has been whether a scale to measure service quality can be universally applicable across industries (Asubonteng et al., 1996; Buttle, 1996). There seems to be a consensus around the idea that the quality dimensions in SERVQUAL need to be modified to consider the important service quality issues specific to different business environments (Mentzer et al., 1999; Rafele, 2004; Kuei and Lu, 1997; Bienstock et al., 1997; Stank et al., 1999; Franceschini and Rafele, 2000; Durvasula et al., 1999). Supporting this claim, Kettinger and Lee (1994) tried to apply SERVQUAL to information services industry and decided to omit the tangibles dimension, since tangibles are not visible to the customers

11

in IS industry. Furthermore, Kettinger et al.(1995) attempted to utilize SERVQUAL in a cross-national study for information services quality and concluded that, besides the fact that the original five dimensions of the SERVQUAL scale have to be modified for the IS industry, important IS quality dimensions might differ from country to country as a result of cultural differences (Kettinger et al., 1995). Another example might be Juwaheer’s paper that identifies 9 factors, some of which are similar to the original SERVQUAL dimensions, specific to and important in hotel industry (Juwaheer, 2004). There are also studies in which nine, seven, four, three, and two-dimensional service quality constructs had been introduced (Mentzer et al., 2001; Emerson and Grimm, 1996; Mentzer et al., 1989; Maltz and Maltz, 1998; Xing and Grant; 2006; Lehtinen and Lehtinen, 1982; Grönroos, 1984). After trying to apply the 22 items of SERVQUAL in four different service settings, Carman (1990) concluded that it is not appropriate to use all the service dimensions as suggested by Parasuraman et al. in 1985 in every setting but instead suggested the use of five to nine dimensions depending on the type of service sector in question. Durvasula et al. (1999) also point out that the RATER dimensions (Reliability, Assurance, Tangibles, Empathy, Responsiveness) cannot be used universally across business sectors. A major criticism has been over the use of gap scores in identifying service quality (Babakus and Boller, 1992; Cronin and Taylor, 1992; Teas, 1993; Brown et al., 1993). Although SERVQUAL proves successful in understanding customer expectations from an excellent service provider, calculating the perceptions-expectations score might yield misleading results. Zeithaml et al. (2006) argue that a measurement program capturing only perceptions of service is missing a critical part of the service quality equation, since customers’ expectations serve as standards or reference points for them and therefore should also be included in the measurement instrument. Although this view is quite reasonable, many researchers claim that perception-only models have higher convergent and predictive validity than models like SERVQUAL (Brown et al., 1993; Teas, 1993) where both expectations and perceptions are considered. Cronin and Taylor (1992) conducted a study in which they compared expectation-perception gaps versus perceptions only, which they call SERVPERF, and concluded that measurement of service performance (perceptions) alone is adequate in understanding service quality. Brown et al. (1993) point out to the problems that may arise as a result of using difference scores in service quality measurement such as poor reliability and variance restriction, as well as to potential problems with that may arise about discriminant validity. Moreover, they show that the perceptions component of SERVQUAL performs

12

nearly as well as SERVQUAL itself (Brown et al., 1993); the empirical findings even support the argument of Cronin and Taylor (1992) that perceptions-only evaluations outperform SERVQUAL in predicting behavioural intentions. Furthermore, SERVQUAL has been criticized for focusing only on the process quality of service delivery and lacking any implications about the outcome quality of the service encounter (Buttle, 1996; Richard and Allaway, 1993; Cronin and Taylor, 1992; Kang, 2006). Grönroos (1984) has suggested that service quality is composed of two sets of attributes, being technical quality and functional quality. Technical or outcome quality considers what the customer actually receives from the service, and functional or process quality refers to the way the service is delivered to the customer (Richard and Allaway, 1993). Mehta and Durvasula (1998) refer to a number of studies that emphasize the service encounter, or process of service delivery, as the critical factor shaping perceived level of customer satisfaction. Although Parasuraman et al. (1985) found out from their research on service quality literature that quality evaluations are made on both the outcome of a service and the process of service delivery, their SERVQUAL seems to include dimensions that refer to only the process of service delivery. Some academics deem it a certain failure since the choice of behaviour (to continue business with the service provider in future or not) is affected by the combination of both service process and outcome quality (Richard and Allaway, 1993). Another important limitation of SERVQUAL is that it has no contribution to assessing a service company’s comparative service performance with respect to its competitors (Min and Min, 1996); it only evaluates the customer expectations and perceptions of a single company’s service performance. As such, unlike constructs like analytical hierarchy process, it fails to contribute to any benchmarking efforts of companies. An extensive list of other criticisms to which SERVQUAL has been subjected can be obtained from Buttle (1996). 2.2. Physical Distribution Service Quality The aim of this study is to measure service quality in a relatively un-researched area, the distributor/wholesaler channel. These channel systems generally include manufacturers, distributors, and distributors’ customers. Manufacturers sell products to distributors, sometimes called wholesalers, who sell the products to business customers, or retailers (Maltz and Maltz, 1998). The main interest of this study is on measuring the quality of service to the distributors’ customers, and therefore it is necessary to give an overall review about relevant literature on measuring physical distribution service quality. Service quality in distribution logistics has interested many researchers (Fung and Wong, 1998; Seth et al., 2006; Rahman,

13

2006; Collins et al., 2001; Rafele, 2004; Bienstock et al., 1997; Mentzer et al., 1989; Mentzer et al., 1999; Mehta and Durvasula, 1998; Chow et al., 1994; Stank et al., 2003) and numerous papers in literature provide different tools for measuring service quality in logistics. Some of them are Rafele’s reference framework for evaluating logistic service performance in a hierarchic structure (2004), and PDSQ by Bienstock et al. (1997) who developed a service quality measurement scale specific to physical distribution service. According to Bienstock et al. (1997) overall channel capabilities such as responsiveness can best be measured based on perceptions, and that using perceived performance allows for common scaling across industries and comparison to the consumer service quality literature (Bienstock et al., 1997; Maltz and Maltz, 1998). Stank et al. (2003) conducted a survey in which they modified SERVQUAL by enhancing the instrument with other dimensions as well as questions regarding loyalty of the customers and financial performance of the service provider (Stank et al., 2003). Mentzer, Gomes and Krapfel, in their 1989 study which integrates logistics and marketing customer service in a conceptual model, argued that availability, timeliness and delivery quality are the major dimensions of logistics service (Mentzer et al., 1989). Mentzer et al. (1999) interviewed a logistics company’s customers in order to find out the factors important in their logistics service quality evaluations. They developed a logistics service quality scale that attempts to measure customer perceptions in logistics sector by identifying nine dimensions specific to industry, being information quality, ordering procedures, ordering release quantities, timeliness, order accuracy, order quality, order condition, order discrepancy handling, and personnel contact quality (Mentzer et al., 1999). In another related study, Mehta and Durvasula (1998) investigated organizational customers in Singapore, who use the services of a shipping line for exporting their products, on their expectations and perceptions of the service quality offered by their preferred shipping lines (Mehta and Durvasula, 1998). They used SERVQUAL as the measurement tool and found out that preferred shipping lines were delivering high quality service and that these levels may serve as standards for other shipping lines (Mehta and Durvasula, 1998). In the mentioned studies which utilized SERVQUAL in order to measure logistics service quality, the authors needed to modify the tool and tailor it to fit into the specific needs of the particular business (Mentzer et al., 1999; Bienstock et al., 1997; Stank et al., 1999; Franceschini and Rafele, 2000; Rafele, 2004). In line with the above criticisms, many researchers have concluded that SERVQUAL dimensions have to be modified according to

14

the properties of the business in question, that is, distribution logistics (Bienstock et al., 1997; Mentzer et al., 1999; Durvasula et al., 1999). Kong and Mayo (1993) also argued that the gaps model of Parasuraman et al. (1985) has to be modified before it can be applied in a business-to-business context. 2.3. Analytic Hierarchy Process It has now been stated that SERVQUAL has a number of limitations and criticisms. In order to avoid being caught in a single perspective by using only SERVQUAL, this paper will also consider an alternative tool for measuring service quality, the analytic hierarchy process. AHP is basically a multicriteria decision making approach in which factors are arranged in a hierarchic structure (Saaty, 1990). It helps to evaluate different alternatives according to a number of decision criteria and their pairwise comparisons. Service quality of a company can be assessed by utilizing the AHP logic explained in the introduction part in a slightly different way. Moreover, in order to apply AHP to service quality measurement, comparing more than one company (alternatives) is not necessary all the time; in fact, customers’ expected and perceived service quality levels could be the ones to be compared. After structuring of the problem as a hierarchy and identifying the criteria that contribute to the overall goal (service quality), the criteria can be pairwisely compared by the customers regarding their relative importance with respect to the overall goal (Saaty, 1990). Then the customers can be asked to compare the company’s actual performance with their initial expectations on all sub-criteria in the hierarchy. Finally, by combining the previous criteria weights and expected and perceived service levels on those criteria, and summing them up at the end the company can understand its overall performance as well as the particular issues that it has to improve in order to increase its customer satisfaction. Common application themes of AHP include selecting a best alternative, resource allocation, evaluation, benefit-cost analysis, priority and conflict, planning and development, and optimization (Vaidya and Kumar, 2006). One of the most common applications of AHP, selection of the best alternative, has been utilized in the areas of warehouse site selection (Korpela and Tuominen, 1996), software selection (Lai et al., 2002; ), project management for selection of the best contractor (Al Harbi, 2001), and vendor selection (Tam and Tummala, 2001). The tool has been used in manufacturing, engineering, management, education and government areas as well as being used in social studies (Vaidya and Kumar, 2006). AHP has also been utilized in combination with other methods such as quality function deployment

15

(QFD) (Armacost et al., 1994; Koksal and Egitman, 1998), goal programming (Badri, 2001; Suresh and Kaparthi, 1992; Ramanathan and Ganesh, 1995), and linear programming (Saaty et al., 2003). Zahedi (1989) and Vaidya and Kumar (2006) provide comprehensive surveys on AHP applications in the literature. Unlike SERVQUAL, the AHP technique allows for the comparison of service qualities of different companies (Chow and Luk, 2005) and also an importance/superiority aspect between quality dimensions. Min and Min (1996) explain why AHP should be used instead of SERVQUAL for service quality assessment and competitive benchmarking purposes. Besides pointing out to SERVQUAL’s lack of competitive analysis, they also say that using AHP would allow a company to investigate the sensitivity of the overall quality measure to any changes in customer judgments (Min and Min, 1996; Ahmed and Rafiq, 1998). Utilization of the AHP technique for service quality measurement purposes has gained interest in the last decade. A number of studies have been conducted that utilized analytic hierarchy process for benchmarking purposes in the hotel industry; they assessed luxury hotels’ service qualities through customer opinions and identified their competitive position among rivals (Min and Min, 1996; Min and Min, 1997; Min et al., 2002). Min and Min (1997) propose competitive benchmarking as an important service improvement tool and use AHP for obtaining customer opinions about the importance of various service attributes and comparing different service providers’ performances. Their proposed methodology could be used for investigating trade-offs among different service attributes associated with changes in importance weights (Min and Min, 1997). Chow and Luk (2005) exercise measuring service quality in fast-food industry and develop an AHP approach that would help managers identify which service dimensions (reliability, assurance, tangibles, empathy, responsiveness) require attention to create a sustainable competitive advantage, and also act as a comparative service improvement technique (Chow and Luk, 2005). Currently no recognized papers utilizing Saaty’s AHP method for measuring physical distribution service quality exists in the literature. 2.3.1. Criticisms of AHP The popularity of AHP stems from its simplicity, flexibility, intuitive appeal and its ability to mix quantitative and qualitative criteria in the same decision framework. Despite its popularity, some shortcomings of AHP have been reported in the literature, which have limited its applicability. One of the two prominent limitations of AHP is about the fact that

16

the number of pairwise comparisons to be judged increases rapidly as the number of alternatives and criteria increases. This is often a tiresome exercise for the decision-maker, which was actually observed in the present study. Some modifications were carried out in the AHP analysis in order to make it easier for the respondents to rank the importance of criteria. Details are presented in appropriate sections. Another criticism about AHP is the issue of rank reversal, as pointed out first by Belton and Gear (1983). The ranking of alternatives determined by AHP may have to be changed by the addition or deletion of another alternative for consideration. Certain precautions, such as ensuring that all the alternatives are considered in the exercise from the beginning and that no alternative needs to be added or deleted in between, can limit the problems of rank reversal. However, rank reversal problem does not affect the analysis presented in this study since the AHP model considers only two alternatives, namely expected service quality (ESQ) and perceived service quality (PSQ). It has to be noted that, in spite of its criticisms, AHP technique continues to find many applications in the literature and might serve as a better alternative to other methods in service quality measurement.

17

CHAPTER 3: Research Methodology This paper aims to utilize SERVQUAL and AHP tools for measuring service quality in a physical distribution service setting. Results of the two instruments will be compared on various bases and a decision will be made about which of the two tools is more appropriate and useful for service quality measurement. In order to conduct this study, the first thing to do was to identify a specific industry in which the surveys were going to be applied, and then to choose a particular company in that business. The author’s selection of industry was physical distribution, or more specifically, pharmaceutical wholesaling. A successful company in this industry in Turkey was selected then. It has already been stated that offering quality service in logistics is quite important for distributor companies since it helps to increase customer satisfaction and loyalty, having the probability that it eventually would yield increased market share and business (Daugherty et al., 1998; Stank et al., 2003). Distributors have to increase their service performance continuously in order to remain competitive among rivals, and it is only necessary that they measure their current service quality first. This study would contribute to previous research on service quality measurement in the sense that it will suggest an appropriate tool that is customized for measuring physical distribution service quality. As discussed in the literature review part, the original SERVQUAL instrument has to be modified according to the specific needs of the industry, since some of the service quality dimensions might not be applicable to every sector. The situation is the same with physical distribution business (Bienstock et al., 1997). Numerous research papers on logistics service quality have been analyzed and most relevant issues to pharmaceutical wholesaling business in Turkey are gathered in this modified version of SERVQUAL. Apart from literature review, meetings with management and personal observations has also contributed to the development of both SERVQUAL and AHP variation models. 3.1. Focus Company Name of the selected distribution company for which SERVQUAL and AHP tools were exercised is Nevzat Pharmaceutical Wholesaler (NPW). Founded in 1962, NPW delivers medicines purchased from pharmaceutical companies to pharmacies and hospitals all over the country. As such, it can be thought of a regional distribution center delivering goods to retailers which are then going to be sold to end customers, that is, patients. NPW holds an average of 4300 SKUs and serves over 5000 customers in Turkey. It is ranked the 4th largest

18

pharmaceutical wholesaler in Turkey with annual revenues reaching up to £125 million in 2005. The company has the headquarters in Ankara with 335 employees and also has 11 subsidiaries and 15 contact offices in different parts of the country. 3.2. Sample Selection Headquarters in Ankara, which contributes a £52 million to overall revenues, was selected as the test region and the set of two questionnaires (a modified version of SERVQUAL and AHP) were applied to a sample of 100 customers (pharmacists) in the area. The survey was conducted by the author through face-to-face interviews with the customers in order to avoid low return rates resulting from mail surveys (Lambert et al., 1993; Maltz and Maltz, 1998; Innis and La Londe, 1994). By this way, 100% participation was guaranteed and also the author helped any customers who had problems filling in the questionnaires. The aim of these interviews was to understand how the customers of NPW perceive its service quality and whether they were happy with the service they receive or not. Selection of the sample customers involved the important concern of homogeneity. The first step was to categorize all customers of NPW according to their purchasing power from the company. Four categories were created as follows: pharmacies that buy merchandise of value over £10,000 monthly belong to first category and named as NPW’s best customers, pharmacies with a monthly purchasing between £10,000 and £4,000 belong to second category, those between £4,000 and £1,000 belong to third category customers, and pharmacies that buy medicine of value less than £1,000 monthly are fourth category or NPW’s worst customers. After this categorization, an equal number of customers (25) were picked from each category in order to consider both loyal and infrequent customers and to maintain homogeneity. The main reason for this kind of a categorization was to avoid any bias that might have been caused by interviewing only loyal and happy customers, or those who are infrequent and less satisfied. Having an equal number of customers from each category ensured having a homogeneous survey sample. In addition, locations of the selected pharmacies were carefully scattered throughout Ankara in order to involve customers both close and very far from the distribution center. As a result of this two-step selection process, a set of 100 customers that were considered representative of NPW’s customer profile in Ankara were informed about the study and were visited by the author for 45 minutes slots determined by appointments.

19

3.3. Development of the Questionnaires 3.3.1. Business Insight In line with the aforementioned criticisms, the original SERVQUAL instrument of Parasuraman et al. (1988) was to be modified for this study in order to obtain a better fit into the business-to-business context (Kong and Mayo, 1993), considering the specific features of pharmaceutical wholesaling industry. Before explaining the modifications, it would be convenient to describe this specific business in more detail. Pharmaceutical wholesalers in Turkey are basically in the middle of the chain that provides medicine to patients. Pharmaceutical wholesalers (PWs) purchase medicine in bulk from numerous pharmaceutical companies and distribute them to pharmacies all over the country. PWs renew pharmacies’ stocks several times a day as needed. Most of the orders are taken through telephone calls, but ordering by fax or internet is also possible. Telephone operators assigned to specific pharmacies take orders regularly and pass these orders to storage area via intranet. Orders are prepared and packed in minutes and loaded to service cars that deliver ordered medicine to customers (pharmacies) at specified time slots. In addition to these regular deliveries, PWs provide prompt deliveries for orders specified as urgent by pharmacies outside the regular service hours. NPW sells mostly the same products with its competitors, that is, all pharmaceutical wholesalers in Turkey sell the medicine that they purchase from pharmaceutical companies. In other words, none of the PWs can differentiate among competitors in terms of the products they offer to customers, since all PWs more or less have the same range of products in stock and the price of the products are exactly the same in all PWs. Therefore, the only way that a PW can differentiate itself and become more preferable to the customer can be through differentiating in customer service. This study would be beneficial to the company as it is to measure and analyze current service quality of NPW and to bring into light the areas on which it should concentrate its resources and aim at differentiating, according to the customer responses to the two questionnaires. 3.3.2. PWSQ The original SERVQUAL instrument was aimed to be modified in the light of this information about pharmaceutical wholesaling business. Through a detailed review of similar efforts to tailor SERVQUAL to distribution service quality (Bienstock et al., 1997; Rafele, 2004; Franceschini and Rafele, 2000; Mentzer et al., 1999) and the author’s first-hand

20

observations about the business, seven dimensions were decided to be effective in customer judgments about service quality of a PW, and Nevzat Pharmaceutical Wholesaler in particular. These seven dimensions are reliability, responsiveness, flexibility, availability, assurance, personnel contact quality and tangibles. After determining the dimensions, individual items representing the various aspects of each dimension were formed, again depending on previous literature search and personal observations. A total of 23 items were written corresponding to the seven pre-specified pharmaceutical wholesaling service quality dimensions. Table II shows the allocation of items to the seven dimensions. In order to stick with Parasuraman et al.’s SERVQUAL (1988), the same 7-point Likert scale ranging from strong disagreement to strong agreement is used in the instrument. The modified SERVQUAL item also has two sections, being the perceptions and expectations sections, as in the original survey. The resulting service quality measurement scale is named PWSQ, standing for pharmaceutical wholesaling service quality. PWSQ is given in Appendix A.
Dimensions Reliability Responsiveness Flexibility Availability Assurance Personnel Cont. Qty. Tangibles Total Number of questions 6 3 3 2 3 4 2 23 Overall percentage 26 13 13 9 13 17 9 100

Table II. Allocation of items to dimensions Although having a number of similarities with the original SERVQUAL instrument, PWSQ is actually quite different in its dimensions. Empathy, for instance, is omitted from the questionnaire since personal attention to individual customers is not considered very important in the business. The other four dimensions, being reliability, responsiveness, assurance and tangibles, of SERVQUAL were maintained in PWSQ since these were considered as important factors shaping pharmacists’ perceptions about NPW’s service quality. It is easy to understand that timeliness and condition are very important factors that would affect customers’ perceptions of a PW’s service quality, as in general within the distribution logistics sector (Maltz and Maltz, 1998; Mentzer et al., 1989), and these two points are gathered under the dimension reliability. Another item that is put under reliability dimension is cycle time performance, or reliability of the order cycle, which is supported as being a fundamental logistics service construct (Maltz and Maltz, 1998). Maltz and Maltz

21

(1998) pointed out in their study that responsiveness, including effective handling of information requests, after-sales service, flexibility, and provision of emergency services, is taken as a dimension that would go beyond basic services and differentiate the distributor’s service. Although not all of them under the responsiveness dimension, these items are present in different parts of PWSQ. Together with responsiveness, assurance is also an important feature that service organizations should have in order to attract customers. Although not very essential, tangibles are still important factors affecting distribution service quality perceptions. Different from SERVQUAL, availability, a synonym for inventory capability, is one of the dimensions of PWSQ, since it is a fundamental factor in the pharmaceutical wholesaling business. The importance of availability in logistics customer service is emphasized in Mentzer, Gomes and Krapfel’s work (Mentzer et al., 1989). Among the numerous researchers who attempted to use SERVQUAL in logistics sector, Bienstock et al. (1997) claimed that timeliness, availability and condition were the fundamental dimensions to assess physical distribution service quality and so should be included in the framework. Emerson and Grimm (1996) further added to Mentzer, Gomes and Krapfel’s dimensions communication, which basically refers to accurate and timely transfer of appropriate information between the seller and the customer. NPW has informed the author about pharmacists attaching a lot of importance to the quality of the personnel assigned to them; it was agreed that the contact personnel who establishes the communication between customers and the company are quite important for a customer’s willingness to carry on the business with NPW. Pharmacists should feel comfortable and easy when placing orders to their telephone operators and this would only be achieved by contact personnel’s courtesy, knowledge, and ease of communication with the customer. This is why the author decided to add a new dimension called “personnel contact quality” to the questionnaires, similar to Emerson and Grimm’s (1996) study. Another issue that pharmacists consider important is the flexibility that a pharmaceutical wholesaler provides in ordering and payment stages. Emerson and Grimm (1999) shows that increased supplier flexibility contributes to customer satisfaction, thereby supporting other researchers’ claims (Oliver, 1980). Since it is confirmed by the company, flexibility was added as a new dimension to PWSQ as one of the important criteria shaping customer opinions. Table III shows references to some important studies on logistics service quality that were made use of in developing PWSQ dimensions and items. Dimensions corresponding to a particular paper are indicated by a check ( ) sign.

22

Mentzer et al. (1999) 1. 2. 3. 4. 5. 6. 7. 8. 9. Items in PWSQ Delivery of orders within the promised time frame Consistency of lead time Ability to perform a service right the first time Delivery of accurate orders Delivery of orders in good condition (undamaged) Invoices matching with orders Delivery of orders regardless of amount Responding to unexpected/urgent orders Willingness to help customers

Francheschini and Rafele (2000)

Daugherty et al. (1998)

Bienstock et al. (1997)

10. Flexibility and ease in order placement 11. Flexibility and ease in payment methods 12. Flexibility in payment conditions 13. Availability of orders in stock 14. Tracking of new products, high product variety 15. Convenient and applicability of return procedures 16. Convenient sales conditions 17. Customers feeling safe and assured with company 18. Courtesy of employees 19. Adequacy of knowledge of employees 20. Providing information on timing of a service 21. Employees being clear and comprehensible 22. Neatness of appearance of employees

23. Modern-looking equipment Table III. References to papers on logistics service quality 23

Taking a look at Parasuraman et al.’s original list of dimensions (1985) show that knowledge and skill of the contact personnel, clean and neat appearance of the public contact personnel, facilitating communication, and financial security and confidentiality are all important items in service quality; these items are included in PWSQ under different names. Most of the dimensions of PWSQ can also be justified by looking at other researchers’ works that did not utilize the SERVQUAL instrument directly but identified important and common indicators of logistics service performance. A good example might be the 1993 study of Lambert et al., who found out the attributes that shippers deem most important in their carrier selection. Of the 18 most important attributes, 8 were similar to the items in PWSQ, some of them being quality of personnel, consistency of transit times, accurate billing and prompt response to claims (Lambert et al., 1993). Moreover, as a result of Rafele’s investigation among several companies in manufacturing and shipment sectors, reliability, completeness, correctness, harmfulness, productivity, lead time, delay, regularity, flexibility, availability and scrap level were found to be the most commonly used indicators of logistics service (Rafele, 2004). Of these eleven indicators, nine are included in PWSQ, although under different names and dimensions. In a study about measuring physical distribution service quality of internet retailers, Xing and Grant (2006) refer to a number of research papers that point out the importance of order fulfillment, in terms of inventory availability, delivery timeliness and reliability (Xing and Grant, 2006). However, the e-PDSQ framework they develop consists of four dimensions –timeliness, availability, condition and return-, based on a number of previous PDSQ constructs (Xing and Grant, 2006). The fact that PWSQ includes all these dimensions adds to its justification. Moreover, the seven R’s of logistics service, as defined by Shapiro and Heskett (1985), are mostly gathered in one dimension, reliability, in PWSQ. Table IV shows brief descriptions of the seven dimensions of PWSQ.
1. Reliability 2. Responsiveness 3. Flexibility4. Availability 5. Assurance Ability to perform the promised service both dependably and accurately Willingness to help customers and to provide prompt service Flexibility to allow for different transaction options and methods Availability of products in stock Ability to convey trust and confidence into customers and make them feel that they are receiving good service 6. Personnel contact Knowledge and courtesy of employees as well as their ability qualityto ease communication with customers 7. Tangibles The appearance of physical facilities, equipment, personnel, and communication materials.

Table IV. PWSQ dimensions

24

The following table (Table V) shows the sources of individual items in PWSQ; they are mainly products of literature search on physical distribution service and management support as well as personal observations regarding important factors in pharmaceutical wholesaling business.
Measure Reliability Delivery of orders within the promised time frame Consistency of lead time Ability to perform a service right the first time Delivery of accurate orders Delivery of orders in good condition (undamaged) Invoices matching with orders Responsiveness Delivery of orders regardless of amount Responding to unexpected/urgent orders Willingness to help customers Flexibility Flexibility and ease in order placement Flexibility and ease in payment methods Flexibility in payment conditions Availability Availability of orders in stock Tracking of new products, high product variety Assurance Convenient and applicable of return procedures Convenient sales conditions Customers feeling safe and assured with company Personnel contact quality Courtesy of employees Adequacy of knowledge of employees Providing information on timing of a service Employees being clear and comprehensible Tangibles Neatness of appearance of employees Modern-looking equipment Notes: * Used in previous studies on service quality ** Developed in conjunction with management and observations of the industry * * * * * ** * ** ** * ** ** ** ** * ** * * * * * * * Source

Table V. Sources of PWSQ dimensions

25

3.3.3 AHP model It has already been stated that SERVQUAL has a number of shortcomings and criticisms. Therefore it might not be an appropriate tool for understanding customer satisfaction and service quality of a company. The second part of this study is aimed at utilizing analytic hierarchy process (AHP) as an alternative to SERVQUAL in measuring physical distribution service quality and comparing its results with those of SERVQUAL. An AHP questionnaire was developed with the same dimensions and items in PWSQ so that the customers were to be questioned on exactly the same points about service quality, as well as for convenience and ease of comparison. The resulting AHP tool with its hierarchic structure supported the idea that several factors combine to compare two things, expected service quality (ESQ) and perceived service quality (PSQ) in this case. The hierarchy of the problem was structured as depicted in Figure 3.

Figure 3. Hierarchy of PW service quality Including the existing parts of AHP in literature, a new questionnaire was developed. The construct consists of 2 main parts; in the first part the customer is requested to rank the aforementioned seven service quality dimensions according to their importance relative to each other by filling in a 7 x 7 matrix A. The seven columns and rows in A, also known as the pairwise comparison matrix (Winston, 2004), each stand for the seven dimensions and 26

constitute the criteria in deciding whether the quality of service was better than expected service quality, or vice versa. The entry in row i and column j of A (aij) indicates how much more important dimension i is than dimension j. “Importance” was to be measured on a special scale whose interpretation can be viewed in Appendix B. For all i, it is necessary that aii=1. If, for example, a25=7, criteria 2 is very strongly more important than criteria 5. If aij=k, then for consistency it is necessary that aji=1/k (Winston, 2004). Customers should compare two criteria by answering questions of the following kind: “Of the two criteria being compared, which one do I consider more important with respect to the overall goal of satisfaction from the service?” With the help of these customer rankings, individual questions in the survey would be given weights that correspond to the relative importance of the particular dimension the question belongs to, from the customer’s point of view. Necessary directions that explain how to fill in the 7 x 7 comparison matrix as well as brief descriptions of each of the seven decision criteria are provided at the beginning of the survey. This first part of AHP is quite important in the sense that it allows for giving different weights to different decision criteria so that outperforming in more important criteria becomes more effective than outperforming in less important criteria. The reason for this allowance is the idea that not every decision criteria might be of same importance while comparing alternatives. This comparison part is one of the main points that make AHP different from the original SERVQUAL (1988) and PWSQ as well. Realizing the need to rank different dimensions while deciding about service quality, Parasuraman et al. added a point allocation part to SERVQUAL in 1991. In the second part of AHP customers are requested to compare NPW’s performance with their initial expectations on 23 items. This question type resembles the recommendation of Carman (1990) and Babakus and Boller (1992) for revising the SERVQUAL scale by combining the expectations and perceptions into a single question. The scale used in the AHP questionnaire is quite different from that of PWSQ. For any item, a customer could choose one comparison statement from the 9 possibilities displayed in Table VI. By this way customers make a single comparison for any item in the questionnaire rather than specifying expectations first and then perceptions in two different steps, as in PWSQ. The same 23 items as in PWSQ are used in this part of the survey. The complete AHP questionnaire developed for measuring NPW’s service quality is given in Appendix B.

27

Scale value 1/9 1/7 1/5 1/3 1 3 5 7 9

Interpretation Excellent Very much better than your expectations Much better than your expectations Slightly better than your expectations They just met your expectations Slightly worse than your expectations Much worse than your expectations Very much worse than your expectations Worst performance

Table VI. The AHP quality measurement scale 3.4. Data Collection It was planned that the two surveys (PWSQ and AHP) were going to be conducted with 100 customers of NPW through face-to-face interviews by the author. The study was initiated as planned; the author visited 5 customers on the first day. The customers easily understood how to answer the PWSQ questionnaire; it was quite simple in the sense that it required customer opinions about expectations from an “ideal” PW and then perceptions of NPW’s service quality. The customers had certain difficulty with the other survey, though. The first part of the AHP model seemed complicated and boring with 21 pairwise comparisons to be made. Although the customers tried hard and filled all the spaces in the pairwise comparison matrices, it was found out that the entries were inconsistent. Table VII displays a sample matrix A filled in by one of the customers:
Objectives Reliability Respons. Flexibility Availability Assurance Pers.Quality Tangibles Reliability 1 1 3 1/3 1 1/3 1/7 Respons. 1 1 3 1/3 1 1/7 1/9 Flexibility 1/3 1/3 1 7 5 1 1/5 Availability 3 3 1/7 1 3 1/5 1/9 Assurance 1 1 1/5 1/3 1 1/9 1/9 Pers.Quality 3 7 1 5 9 1 1 Tangibles 7 9 5 9 9 1 1

Table VII. A sample pairwise comparison matrix

28

This matrix has significant inconsistencies: for example, the very first customer entry in the first row indicates that reliability and responsiveness dimensions are of the same importance to this customer; however, towards the end of the first and second rows, the same customer has implied that reliability is weakly more important than personnel contact quality but responsiveness is very strongly important than personnel contact quality. In this case, how can two equal dimensions be of different importance values in comparison with a particular dimension? Their rank of importance should also be same according to another dimension. Another inconsistency in the above matrix appears in third row: the customer implied that flexibility and personnel contact quality are two dimensions with the same importance; however the two dimensions received different pairwise comparison scores from reliability, responsiveness, availability, and assurance dimensions. The consistency checks performed at the end of the day revealed that responses in all the 5 matrices filled in were significantly inconsistent according to Winston (2004, p. 802) and Saaty (1990, p. 13) with CI/RI ratios greater than 0.10. Depending on the first day results, the author tried to find an alternative way of obtaining customer opinions regarding the importance of seven service quality dimensions for the AHP survey. Inspired by the additional section at the end of the original SERVQUAL tool (Parasuraman et al., 1991), it was decided to replace the pairwise comparison matrix with a direct rating question which requests that customers give points to the seven service quality dimensions -between a minimum of 0 and a maximum of 100- according to their importance in their service quality evaluation. The resulting new questionnaire was called the AHP variation model. When the new model was tested with the same 5 customers, to whom the pairwise comparison matrices were presented, they agreed that this new model facilitated giving accurate responses. After the AHP variation model proved successful in gathering customer opinions about the importance of various service dimensions, the author decided to use this version of the second questionnaire throughout the entire study. The AHP variation model was then presented to the rest of customers in the test sample, completing to 100 respondents. The modified part of AHP survey is given in Appendix C.

29

CHAPTER 4: Data Analysis 4.1. AHP Results Once all the customers in the prespecified sample of 100 were interviewed with, the responses were analyzed. The main tool used for data analysis was Microsoft Excel, since performing the calculations on spreadsheets would facilitate any statistical tests and also comparative analyses between different spreadsheets. The first series of analyses were conducted for the AHP data. A sample AHP variation model questionnaire filled in by one of the respondents is shown in Appendix D. 4.1.1. Point allocations A very important feature of the AHP variation used in this study is that it gets customer opinions about the importance that they assign to various service dimensions. With the modifications explained above, the first section demands that customers give points between 0 and 100 to the seven dimensions according to their importance to them in evaluating a service provider. As the points over 100 are obtained from customers, they are converted to weights that correspond to each dimension. For example, in the sample rating given in Appendix D, the customer had given 70 points to flexibility, 85 points to responsiveness and tangibles, and 100 points to the rest of the dimensions. The weight of a dimension is calculated by summing all the points for seven dimensions and dividing the rating of a particular dimension by this total number. The weight of flexibility, for example, is therefore equal to 70/(100+85+70+100+100+100+85) = 70/640 = 0.109, and the weight of reliability is equal to 100/640 = 0.156. All other dimensions’ weights can easily be calculated in the same way; the weights of responsiveness, availability, assurance, personnel contact quality and tangibles are 0.133, 0.156, 0.156, 0.156, and 0.133, respectively. All the weights then add up to 1 to represent the service quality evaluation process as a whole. Most of the respondents assigned the highest points to reliability, assurance and personnel contact quality dimensions. Specifically, 63% of them gave the highest points to reliability, 62% gave the highest points to assurance, and 65% gave the highest points to personnel contact quality. It was the general case that one customer gave the same importance points to more than one dimension. It might be useful to give more details about the respondents’ importance point allocations to dimensions. Only 30% of the customers considered availability as the most important dimension; 47% of them assigned the lowest points to availability. 49 customers thought that responsiveness is the most important dimension, 24

30

thought that it was the second most important dimension, and 14 customers assigned the lowest points to responsiveness as being the least important dimension. The dimensions that were mostly considered as the least important dimensions were availability and tangibles. Table VIII provides details on percentages of respondents who chose a dimension to be most important, second most important, and least important for them in assessing pharmaceutical wholesalers’ service quality.
Overall RELIABILITY RESPONSIVENESS FLEXIBILITY AVAILABILITY ASSURANCE PERS.QUALITY TANGIBLES most important 63 49 47 30 62 65 34 Second most important 27 24 21 14 32 12 21 Least important 5 14 25 47 4 9 36

Table VIII. The most and least important dimensions The detailed investigation of importance weights assigned to service quality dimensions by the 100 respondents revealed no significant differences between the four categories of customers; that is, the dimensions ranked highest and lowest were generally the same in different categories. The only interesting point was that the customers in fourth category ranked tangibles as the most important dimension much more often than customers in the other three categories. Ranking of tangibles as first dimension was quite low (about 24%) in the first, second and third categories, whereas this percentage was 64% in the last category. 4.1.2. Sample calculations The second part questions of the AHP variation model involved ratings for the 23 service attributes with respect to customers’ initial expectations. Each marking on AHP surveys were converted to rating numbers; for example, the sample respondent in Appendix D had chosen the statement “Much better than your expectations” that corresponds to a rating of 1/5 for the third question. This rating leads to a comparison matrix as the one in Table IX.
Question 3 EXPECTATIONS PERCEPTIONS Total ANORM EXPECTATIONS PERCEPTIONS EXPECTATIONS 1 5 6 EXPECTATIONS 1/6 5/6 PERCEPTIONS 0.2 1 1.2 PERCEPTIONS 0.2/1.2 = 1/6 1/1.2 = 5/6

WEIGHTS (1/6+1/6)/2 = 1/6 (5/6+5/6)/2 = 5/6

Table IX. A sample response matrix 31

Ratings obtained from the second part questions are transferred to the comparison matrices in Microsoft Excel in order to find out the satisfaction scores for each item in the questionnaire. In the above example expectations have a weight of 0.16667 and perceptions have a weight of 0.83333; taking the ratio perceptions/expectations, we obtain a satisfaction score of 0.83333/0.16667 = 5 which means that NPW had performed much better than the customer’s expectations regarding performing a service right the first time. After all the 23 items are converted into P/E scores, scores of the items that corresponded to the same service quality dimension were averaged to obtain the satisfaction score of the dimension. For the sample customer in Appendix D the satisfaction score for reliability was found to be (3+3+5+7+7+7)/6 = 4.76, meaning a performance much better than the customer’s initial expectations. This number could also be found by calculating the average expectations and perceptions weights of the 6 items of reliability dimension and taking their ratio. Satisfaction scores of all dimensions were calculated in the same fashion. Finally, it was necessary to find out the overall customer satisfaction score which depended on both the importance ratings from the first part and the P/E scores from the second part. The calculations for the sample data are shown in Table X.
Expectations Perceptions Responsiveness matrix Expectations Perceptions Flexibility matrix Expectations Perceptions Availability matrix Expectations Perceptions Assurance matrix Expectations Perceptions Pers.Quality matrix Expectations Perceptions Tangibles matrix Expectations Perceptions Reliability matrix 0.1736 0.8264 0.2917 0.7083 0.1389 0.8611 0.2500 0.7500 0.2639 0.7361 0.1354 0.8646 0.2500 0.7500

RELIABILITY RESPONSIVENESS FLEXIBILITY AVAILABILITY ASSURANCE PERS.QUALITY TANGIBLES Total

Points over 100 WEIGHTS 100 0.156 85 0.133 70 0.109 100 0.156 100 0.156 100 0.156 85 0.133 640 100.000

Overall expectations score = 0.156*0.1736 + 0.133*0.2917 + 0.109*0.1389 + 0.156*0.25 + ‘0.156*0.2639 + 0.156*0.1354 + 0.133*0.25 = 0.2157 Overall perceptions score = 0.156*0.8264 + 0.133*0.7083 + 0.109*0.8611 + 0.156*0.75 + 0.156*0.7361 + 0.156*0.8646 + 0.133*0.7500 = 0.7843 Overall satisfaction score = P/E = 0.7926/0.2074 = 3.6358

Table X. Sample P/E score calculation

32

Having an overall satisfaction score of 3.64, the sample respondent can be said to have received a service that was better than his expectations. All the sample responses were analyzed similarly and their overall as well as dimension-based P/E scores were recorded. 4.1.3. P/E scores As a result of a detailed investigation of the satisfaction scores for the AHP variation model, dimensions that were most and least satisfying were identified. The first category customers were most happy with personnel contact quality, as implied by the score 6.75, whereas availability was the dimension with which they were most dissatisfied with a satisfaction score of 3.44. The situation is the same with second, third and fourth category customers. In general, the dimension that respondents were most satisfied with was personnel quality, while the one that is most dissatisfying was availability. Details of the satisfaction scores in terms of the seven service quality dimensions are given for each category of customers in Table XI.
Categories First category Second category Third category Fourth category Reliability Respons. Flexibility Availability Assurance Pers. Qty Tangibles 5.0778 4.8662 4.1082 4.3525 5.6679 5.4146 4.0084 5.1421 6.0807 4.1418 4.2127 4.7165 3.4422 3.3198 2.2617 3.2169 3.7490 3.7817 3.4084 3.6841 6.7546 6.3109 4.9185 5.8241 4.2406 3.8437 4.0204 4.6204 Overall 3.6090 3.6066 3.1311 3.7186

Table XI. Category-based satisfaction scores for the seven dimensions As implied in the above information, availability was found to be the least satisfying dimension in the overall, with a satisfaction score of 3.06, and personnel contact quality was the most satisfactory dimension in the overall among the others, with the score 5.95 out of 9. A point to note is that first category customers were slightly more satisfied with almost all the seven dimensions than the other 3 categories. The overall AHP service quality measures of the first, second, third and fourth category customers were 3.61, 3.61, 3.13 and 3.72, respectively. As a result of the overall analysis, it was found that NPW’s average customer satisfaction score is 3.5163; in terms of AHP terms, this score means that NPW’s performance were slightly better than its customers’ expectations. The highest score achieved in terms of customer satisfaction was 8.26; this essentially means that in the overall, the service quality of NPW was very close to excellent in the scale, or in other words NPW performed very much better than it’s most happy customer’s

33

initial service expectations. The lowest score observed was 0.43, implying that NPW’s service quality was slightly worse than its most dissatisfied customer’s expectations. 84% of all the respondents were found to be happy with the service they received from NPW and thought that their perceptions were higher than their initial expectations, meaning that 16 of the respondents (representing 16% of the whole) were dissatisfied with the service quality of NPW (satisfaction scores less than 1). In terms of the individual dimensions, 96% of the customers were happy with the personnel quality of NPW and 93% were happy with the flexibility NPW offers. 23% of the customers were dissatisfied with the availability aspect of NPW’s service and 12% were not happy with the tangibles dimension. The full details of customer responses in terms of individual dimensions are given in Table XII.
Satisfied customers 92 92 93 77 89 96 88 Dissatisfied customers 8 8 7 23 11 4 12

RELIABILITY RESPONSIVENESS FLEXIBILITY AVAILABILITY ASSURANCE PERS.QUALITY TANGIBLES

Table XII. Distribution of satisfied and dissatisfied customers The general picture seems to be quite positive for the company in question; it is implied in the AHP survey results that most of the customers are happy and satisfied, and even delighted with the service NPW provides. The happy customers are likely to be continuing business with NPW and even be involved in positive word-of-mouth communication with other pharmacists regarding NPW’s high quality service, causing increased business. 4.1.4. Regression Analysis It would be interesting to investigate if any variables affect the level of customer satisfaction that NPW faces. It is a possibility that a number of factors combine together with customer expectations and perceptions and yield a certain level of customer satisfaction. A good idea might be to look for any relationships between a particular customer’s satisfaction score and the number of years for which that customer has worked with NPW. As a result of the above analyses, most of the respondents to the AHP survey seemed to be happy with the service they received; however there were still some unhappy customers. One

34

of the reasons for this variance in customer satisfaction scores might be the effect of years of association of the customers with the company; that is, older customers of NPW might be more satisfied with the service and as a result they are still continuing to do business with the company. In order to understand whether such a relationship was present between satisfaction and years of association, the author decided to perform a regression analysis. In the regression line equation Y = a + bX, Y stands for the customer satisfaction level (P/E) in AHP data as the dependent variable where X stands for the number of years of association as the independent variable. By using the software called STATA, the regression line equation is obtained as follows: Y = 5.437 + 0.618 * X This equation implies that years of association has a positive impact on customer satisfaction within our sample of 100 customers. It would be convenient to look at population parameters before concluding that this association exists in the whole population of NPW’s customers. The equation of population regression line is Y =A + BX, where A and B are the Y-intercept and regression coefficient, respectively, of the population. The sample regression line, Y = 5.437 + 0.618*X, estimates the population regression line, Y =A + BX, and therefore it is possible to use it to make inferences about the population regression line. The basic thing to be investigated here is whether years of association is a significant factor in explaining the variance in customer satisfaction or not. Therefore the hypothesis to be tested is: H0: B=0 ← Null hypothesis: years of association has no effect on satisfaction H1: B≠0 ← Alternative hypothesis: years of association has significant effect on satisfaction STATA gives the standardized test statistic as 2.47. Testing the hypothesis with a 0.05 level of significance and 99 degrees of freedom, the appropriate t value to reject the null hypothesis is approximately 1.980. Because the main concern here is whether b is equal to zero or not, this is a two-tailed test and the critical value is 1.980. Since the t statistic (2.47) lies outside the acceptance region, we reject the null hypothesis that years of association has no effect on satisfaction and conclude that length of the history of business between a customer and the company significantly effects the satisfaction level of that customer.

35

4.2. PWSQ Results After the 100 AHP questionnaires were completely analyzed, a second series of analyses were set up for the PWSQ data. Question by question, all responses on the 100 questionnaires, for both expectations and perceptions sections, were transferred into Microsoft Excel. Then, analyzing the differences between perceptions and expectations responses, the gap scores (PE) were obtained and the data were ready to be analyzed in terms of customer satisfaction/dissatisfaction. A sample PWSQ survey which is filled in by the same respondent as in Appendix D is displayed in Appendix E. Literature on measuring service quality involves studies in which satisfaction scores are not analyzed assuming all service dimensions are of equal importance, or in other words, unweighted, but instead weighted analyses are conducted (Teas, 1993; Cronin and Taylor, 1992; Mehta and Durvasula, 1998) since customers may assign different importance weights, or priorities, to different factors and mentally evaluate the overall service quality of a service encounter in accordance with those priorities. It might be much more accurate and realistic to assess a customer’s satisfaction at the end of a weighted analysis. In the light of the information from literature, the author decided to conduct both unweighted and weighted analyses of the PWSQ questionnaires, and then to compare the results as which analysis best explains the variation in customer satisfaction scores. 4.2.1. Unweighted analysis To conduct the unweighted analysis for PWSQ it was first necessary to calculate the average gap scores for each of the seven dimensions, by taking averages of the gap scores of the questions corresponding to each dimension. These calculations were done for each of the 100 respondents on an Excel spreadsheet. By doing that, it was possible for one to assess whether any particular customer was satisfied with a particular dimension. Say for the reliability dimension, if the average gap score of the 6 corresponding questions is greater than zero for customer X, than we can say that expectations of customer X from NPW in terms of reliability were lower than his perceptions; in other words, customer X was satisfied or happy with the service he received in terms of reliability dimension. In the sample survey in Appendix E, the respondent had marked point 7 for all the 6 items of reliability dimension in the expectations part, meaning she strongly agreed that the stated features should exist in an excellent PW. In the perceptions part, however, she marked points 5, 5, 6, 6, 6 and 7, respectively, for the first 6 questions. The corresponding gap scores are calculated by simply

36

subtracting the expectations scores from perceptions scores. The sample respondent’s gap scores for the reliability dimension are therefore -2, -2, -1, -1, -1 and 0. These gap scores mean that this customer is not happy with the first five items in the survey and only satisfied with the sixth item, invoices matching purchase orders. The average score for the reliability dimension is therefore (-2-2-1-1-1+0)/6 = -1.16667, implying dissatisfaction with the reliability aspect of NPW’s service. After calculating the average scores on each dimension, these seven averages are summed and then divided by seven in order to obtain the overall satisfaction scores for each respondent. The resulting score of each customer gives the unweighted measure of service quality, since all the seven dimensions were treated as the same. The overall average gap score for the sample survey in Appendix E is -1.3333, which means that the respondent’s perceived service is inferior to his expectations. 4.2.1.1. P-E scores As a result of the unweighted analysis of PWSQ data it was found that the first category customers were most satisfied by the personnel contact quality of NPW, with an average gap score of -0.18. The dimension with which the first category customers were most dissatisfied was availability, with an average gap score of -1.74. Similarly, the second category customers were most satisfied with the personnel contact quality of the service provided by NPW, with a score of -0.36, and were most dissatisfied with availability, with a score of -1.52. Third category customers were most happy with the flexibility dimension (-0.44) and were least happy with availability (-1.56). Finally, the dimensions with which the fourth category customers were most and least satisfied were flexibility (-0.43) and availability (-1.14), respectively. Looking for any similarities in these results, it is easy to say that availability is the dimension that all categories of customers were most dissatisfied with. Also, personnel quality and flexibility are among the dimensions that NPW’s customers are most happy with. Details about category-based satisfaction scores for each dimension are given in Table XIII.
Categories First category Second category Third category Fourth category Reliability Respons. Flexibility Availability Assurance Pers. Qty Tangibles -0.8467 -0.5400 -0.7667 -0.7933 -0.1867 -0.4000 -0.6133 -0.5067 -0.3333 -0.3867 -0.4400 -0.4267 -1.7400 -1.5200 -1.5600 -1.1400 -1.0400 -1.0533 -1.0133 -0.9067 -0.1800 -0.3600 -0.5900 -0.5300 -0.4600 -0.5600 -0.8200 -0.6000 Overall -0.6838 -0.6886 -0.8290 -0.7005

Table XIII. Unweighted gap scores for the seven dimensions

37

In the overall, flexibility is the dimension that received the highest score for customer satisfaction (-0.3967) and availability is the one with the lowest gap score (-1.49). The overall unweighted service quality measures of the first, second, third and fourth category customers were -0.68, -0.69, -0.83 and -0.70, respectively. By analyzing all 100 respondents together, it was found that NPW’s average customer satisfaction score on the overall was -0.7255. It is an interesting point to note that all of the aforementioned average gap scores, no matter if they belong to most satisfactory or least satisfactory dimensions, are negative, that is, below zero. This essentially means that even though a dimension received the highest score among the seven of them, it was still performing worse than customers’ expectations. Unfortunately, there was not any one dimension of NPW’s service quality that was found to be above customer expectations. Wall and Payne (1973) argued that this kind of consistently low deficiency scores would normally appear when people are to compare the ideal thing with the reality. When they respond to “what is desirable” and “how much is there now,” they seldom rate the former lower than the latter (Wall and Payne, 1973). Babakus and Boller (1992) therefore pointed out that expected service quality scores may exceed perceived service quality scores consistently for no other reason than this type of response tendency. This argument is well verified in the present study by the large number of negative gap scores. The highest overall satisfaction score that NPW got from its customers was 0.9524, meaning that at the best case NPW could manage to go beyond the initial expectations of its customers by almost 1 point. The lowest score was -2.5595, meaning that NPW had performed approximately 2.5 points worse than it’s most dissatisfied customer’s expectations. Only 14% of the customers were satisfied with the service of NPW (positive P-E scores) and 86% of the customers implied that the service they received were worse than their expectations (negative P-E scores). It is apparent that there is a consistent gap between customer expectations and NPW’s actual performance which has to be taken care of as soon as possible. In terms of individual dimensions, 48% of the customers were happy with the responsiveness of NPW and 46% were happy with the flexibility NPW offers. 84% of customers were dissatisfied with the assurance aspect of NPW’s service and 80% were not happy with the availability aspect. The full details of customer responses in terms of individual dimensions are given in Table XIV.

38

RELIABILITY RESPONSIVENESS FLEXIBILITY AVAILABILITY ASSURANCE PERS.QUALITY TANGIBLES

Satisfied customers 22 48 46 20 16 39 43

Dissatisfied customers 78 52 54 80 84 61 57

Table XIV. Distribution of satisfied and dissatisfied customers for unweighted PWQ analysis 4.2.1.2. Regression Analysis In this section, the regression analysis performed previously for the AHP data is repeated for PWSQ data. Years of association of the 100 respondents with NPW are regressed with their unweighted satisfaction (P-E) scores; “years of association” was taken as the independent variable whose relationship with “customer satisfaction score” as the dependent variable was sought in the form Y = a + bX. The set of hypotheses to be used here is mostly similar with the above discussions, only with the difference that unweighted customer satisfaction levels are considered here as the dependent variable. H0: B=0 ← Null hypothesis: years of association has no effect on satisfaction H1: B≠0 ← Alternative hypothesis: years of association has significant effect on satisfaction STATA yielded the following regression line equation for the 100 sample data: Y = 7.574 – 0.053 * X This regression line suggests that years of association has a slight negative impact on customer satisfaction. Hypothesis testing is performed with the same level of significance in the previous section and therefore the test statistic is again 0.198. Since this is a two-tailed test, the acceptance region is between -0.198 and 0.198. The standardized regression coefficient resulting from the regression analysis takes on the value -0.059, which is within our acceptance region. As a result, we must accept the null hypothesis that years of association has no effect on the level of customer satisfaction. 4.2.2. Weighted analysis As mentioned above, the PWSQ data were to be analyzed both weighted and unweighted. The idea behind weighted analysis is that consumers generally assign priorities to different

39

aspects of a service and judge the overall service quality they receive according to these mental assignments; that is, they consider satisfaction in an important dimension as dominant to satisfaction in a less important dimension. For example, in a restaurant setting, a customer who thinks that cleanliness of the restaurant and meals is much more important than decoration of the place, he would still be very satisfied in observing that the restaurant is quite clean as well as the dishes, even if the décor is ugly and depressing. In this case the customer mentally assigned weights to his satisfaction scores for the two dimensions and the resulting impression was satisfying since the weight of cleanliness was much higher than that of décor. With this logic, the author decided to include weights for the seven dimensions of PWSQ in overall satisfaction analyses. The 100 respondents have assigned importance points to dimensions in the first section of the AHP variation questionnaire. These points could easily be converted to importance weights for the seven dimensions and used in weighted gap score analyses in this part of the paper. Once the weights are transferred from the AHP questionnaires for each respondent, it is only necessary to multiply the average P-E score for each dimension with its importance weight and then to add up the weighted scores for each dimension to obtain each customer’s overall weighted PWSQ score. For the sample survey given in Appendix E, we have to use the points given by the same respondent to the seven service quality dimension in Appendix D. The weight of reliability among the seven dimensions was 0.156. Multiplying this number by -1.16667, the average gap score for reliability, we obtain -0.182 as the weighted gap score for reliability. After calculating the weighted gap scores of each dimension in a similar fashion, we have to add them up in order to obtain the overall weighted satisfaction score of the respondent. The weighted gap score of the sample respondent is found to be -1.3403, indicating dissatisfaction from the service NPW delivers. 4.2.2.1. P-E scores First category customers had an overall weighted satisfaction score of -0.67. The next category of customers had this score as -0.68. Third and fourth category customers had a weighted gap score of -0.83 and -0.70, respectively. The significant issue of most satisfaction scores being negative in the unweighted analysis appears in this part as well, only meaning that most of the respondents were not satisfied with the service that NPW provided. The highest satisfaction score observed in this study was 0.9372 and the lowest was -2.5846. As a result of the weighted analysis, only 14% of the customers were found out to be happy with

40

the service they receive, and 86% implied that the service they received were worse than their expectations (negative P-E scores). The overall weighted PWSQ score for the 100 respondents was -0.7185. As in the previous section, a substantial gap is present between customer expectations and perceptions, causing NPW mostly to fail to satisfy its customers. 4.2.2.2. Regression Analysis As a last study regarding the effects of years of association on customer satisfaction, the weighted P-E scores of the 100 respondents of the study are regressed this time with their length of business with NPW. The corresponding regression equation is as follows: Y = 7.695 + 0.118 * X This regression line suggests that years of association has a positive impact on customer satisfaction. The significance level is again selected as 0.05 and the test statistic is therefore still 0.196 and the acceptance region is between -0.196 and 0.196. STATA gives the value of standardized regression coefficient as 0.136, which is well within our acceptance region. As a result, we must accept the null hypothesis that years of association has no effect on the level of customer satisfaction. This result is the same with the unweighted analysis; therefore we can say that analyzing PWSQ data in unweighted or weighted forms does not reveal a significant interaction with years of association.

41

CHAPTER 5: Comparison of the Three Approaches This study dealt with measuring service quality of a distributor company through the utilization of two different tools which had in their essence SERVQUAL and AHP techniques. This part of the paper will attempt to compare the two instruments, their results, advantages and disadvantages. The comparison will be made on two bases, behavioural and numerical. 5.1. Behavioural Comparison A homogeneous sample of 100 customers were selected from NPW’s customer pool and visited upon appointment by the author. A set of questionnaires consisting of PWSQ and the AHP variation model consecutively was presented to each customer. Customers were given brief information in advance regarding the purpose and scope of the study, so they were aware that they were to answer a series of questions about service quality of NPW. Because it was the first questionnaire in the set of two, all customers started from answering PWSQ. Customers could easily fill in the first part (expectations from an ideal PW); they usually indicated quite high expectations (6 or 7) on the statements since they were all selected to be important factors in service quality. Having the impression from the first couple of questions of PWSQ that all the items would be dealing with important factors, customers did not seem to put a lot of effort in thinking about each item in depth and quickly selected high points close to “Strongly agree” in the first part. It generally took longer for them to answer the second part questions (perceptions of NPW’s service quality) since they needed to think each item one by one in terms of their past experience with the company. Some of the respondents showed signs of boredom while answering PWSQ and said that it would have been better and easier if they had to deal with each item –delivery of orders within the promised time frame, for example- only once, instead of stating general expectations first and then what they actually though about NPW’s service on the same item. It took approximately 10 to 15 minutes for a respondent to finish answering PWSQ. The next instrument in the set of questionnaires was the AHP model. As explained in the questionnaire development part, the initial AHP questionnaire included a pairwise comparison matrix at the beginning. Unfortunately, all the five customers the author interviewed on the first day of the study filled the matrix in inconsistently. Moreover, they were quite bothered while filling in the matrix; they thought it was far too complicated and difficult as well as time-consuming. They had difficult time in comparing the seven dimensions pairwisely. At

42

that point the author observed the drawback of AHP that increasing the number of criteria to be compared with each other would increase the number of pairwise comparisons to be made significantly. Should only the RATER dimensions of the original SERVQUAL tool were used the number of pairwise comparisons to be made would be 10; increasing the dimensions to seven caused this number rise to 21, more than doubling it. It would have been easier for the respondents to fill in the matrix if the number of dimensions were less; however all the dimensions had been formed and selected very carefully to make PWSQ fit into the specific needs of the pharmaceutical wholesaling industry. Despite the difficulty with the first part, the second part of the AHP model where the customers were to compare their actual service experience with their initial expectations on the provided scale was quite straightforward and nearly no problems were observed in answering the second part questions. Coupled with the inconsistencies of comparison matrices filled in till then, the obvious boredom of respondents in answering the first part of the questionnaire made the author modify the original survey as explained in the Chapter 3. When the modified part of the AHP survey was presented to the first 5 customers who dealt with the original version before, they said this modification to the questionnaire made it significantly easier and they felt they could give more accurate responses this way. The variation model took the respondents approximately 15 minutes to fill in completely, whereas the original AHP model had taken them more than 30 minutes. The ease of use of the direct rating method was obvious in terms of time consumption. As to compare the two tools, PWSQ and AHP variation model, on behavioural basis, it is easy to say that questions in AHP model were more straightforward than those in PWSQ, since they simply attempted to compare perceptions with expectations and revealed at once whether a customer is satisfied or not with a particular service item. With the modification of the first part, respondents found it easier to answer the AHP questions than PWSQ questions; this was mainly due to boredom caused by dealing with the same items two times in the same questionnaire. An important observation regarding customer responses to the two questionnaires is that, as mentioned above, respondents generally implied very high expectations in the 23 items of PWSQ; unintentionally, they did not leave NPW any chance to go beyond their expectations and delight them, at least on the paper. As Babakus and Boller (1992) point out, a number of

43

psychological constraints may be activated to make the resulting gap scores problematic, when people are asked to indicate a “desired level” and “existing level” on a particular attribute. Since the respondents constantly selected the highest expectations scores, the best possible result that NPW could get in terms of customer satisfaction in this questionnaire was satisfying the customer, only if respondents selected high scores in the perceptions section of PWSQ. This was the case in the sample survey given in Appendix E, where the respondent selected point 7 for each and every item in the expectations section. Since his perceptions were not all marked point 7 in the second part, the customer seemed dissatisfied with NPW’s service quality. The same respondent, however, seemed quite happy with the service quality of NPW as a result of AHP data analysis. This may be claimed as a serious drawback of SERVQUAL variation models, of which PWSQ is an example, in terms of revealing customer satisfaction levels accurately. It seemed that customers could express their satisfaction and comparisons more easily with the AHP variation model than with PWSQ. In an appropriate SERVQUAL variation, all the dimensions would be selected carefully in order to include the factors that are quite important to the consumers of that particular service; therefore it is usually expected that the respondents would choose high scores in the expectations part of the questionnaire. In fact, it was expected that the sample customers of NPW would indicate strong agreement with the scale items, since each of the 23 items were developed as a result of detailed literature reviews and meetings with management to identify the most important features in PW service quality. The obvious implication of this logic is that a service provider has nearly no chances of delighting its customers by outperforming their expectations; the best they could do on the paper is to match the expectations. Based on this it might be said that SERVQUAL shortfalls in reflecting the accurate level of customer satisfaction. This view will be supported by numerical comparisons in the following part. 5.2. Numerical Comparison The results of the study were analyzed in three versions; unweighted PWSQ, weighted PWSQ and AHP variation model. Details of the results were explained in Chapter 4. Now the three versions will be compared and some inferences will be made regarding which way is better to obtain accurate customer satisfaction scores. It might be a good idea to first of all compare the results of the unweighted and weighted analysis of PWSQ before moving on to a comparison of PWSQ and AHP results.

44

5.2.1. Unweighted PWSQ versus weighted PWSQ The search for any significant differences between the customer satisfaction levels of the four customer categories had arisen from the idea that customers doing more business with NPW would be more satisfied with the service they receive and that might be the reason why they are better customers of the company. In other words, this investigation attempted to support the claim that increased customer satisfaction would yield to increased business (Anderson et al., 1996; Daugherty et al., 1998). Unfortunately the analysis yielded no such variation among the four categories. As it can be seen from Table XV, first, second and third categories had nearly the same satisfaction levels in both unweighted and weighted PWSQ results, whereas this level decreased for the third category. It is counter-intuitive that customers doing very little business with the company were more satisfied with the company’s service quality than customers having more business relationships with the company.
Unweighted PWSQ 1. category average 2. category average 3. category average 4. category average Overall -0.6838 -0.6886 -0.8290 -0.7005 -0.7255 Weighted PWSQ -0.6729 -0.6774 -0.8256 -0.6982 -0.7185 AHP variation model 3.6090 3.6066 3.1311 3.7186 3.5163

Table XV. Category-based overall satisfaction scores In terms of overall satisfaction, the unweighted and weighted analyses of the PWSQ data returned exactly the same amount of satisfied and dissatisfied customers, 14% and 86%, respectively. Moreover, the 14 happy customers were exactly the same in both studies. The overall satisfaction scores (P-E) for the unweighted analysis was -0.7255, whereas this score was -0.7185 as a result of the weighted analysis. The implication of these results is that applying weighted analysis to SERVQUAL variation models that normally treat every service quality dimension equally does not necessarily yield very different results from the unweighted analysis; therefore analyzing only unweighted SERVQUAL gap scores might be sufficient in understanding customer satisfaction (Mehta and Durvasula, 1998). This was basically the case in the present study; the unweighted and weighted analyses of the PWSQ data returned very close customer satisfaction scores as well as the amount of satisfied and dissatisfied customers. So we can say that the weights given to the seven dimensions did not

45

influence the results of this study very much for PWSQ analyses; they did not even transfer any one customer from the dissatisfied group to the happy group, or vice versa. 5.2.1.1. Correlation analysis Obtaining so similar results at the end of unweighted and weighted analyses brought about the need to make a correlation analysis among unweighted and weighted P-E scores; by this way it would be possible to claim that the two variables move together. The two measures for describing the correlation between two variables are the coefficient of determination (r2) and the coefficient of correlation (r) (Levin and Rubin, 1998). r =
2

a ∑ Y + b ∑ XY − nY

2

∑Y

2

− nY

2

where r2 = sample coefficient of determination a = Y-intercept b = slope of the best-fitting estimating line n = number of data points X = values of the independent variable Y = values of the dependent variable Y = mean of the observed values of the dependent variable

This formula implies that a regression line be found first for the two variables and then a and b are used to calculate the coefficient of determination. The coefficient of determination can take on vales between 0 and 1; an r2 close to 1 means that a strong correlation exists between the two variables whereas an r2 close to 0 indicates little correlation between the two. Now it is time to investigate the amount of correlation between the 100 unweighted and weighted customer satisfaction scores of PWSQ. First of all the equation of the regression line should be obtained. The resulting line has the following equation:
Y = -0.0195 + 1.0201 * X,

where Y is weighted (P-E) score and X is unweighted (P-E) score. Then, using directly the above formula, r2 is: (−0.0195)(−71.85) + (1.0201) * (89.08) − (100)(0.7185) 2 40.6477 r = = = 0.976. 92.29 − (100)(0.7185) 2 41.6586
2

46

This value of coefficient of determination being so close to 1 implies that there is a very strong correlation between unweighted and weighted (P-E) scores; in fact, it can be said that 97.57% of the variation in weighted customer satisfaction scores (the dependent variable Y) can be explained by the variation in unweighted customer satisfaction scores (the independent variable X). The coefficient of correlation can also be calculated here. It is simply:

r = r 2 = 0.976 = 0.988.
This number implies that the relation between the two variables is direct and the slope is positive. The combined consideration of these two coefficients helps to explain why the numerical results obtained from unweighted and weighted analyses in the previous parts were not very different from each other. This result supports Cronin and Taylor’s findings regarding unweighted and importance weighted measures; they found a high correlation between the two measures and concluded that using unweighted measures is sufficient (Cronin and Taylor, 1992). 5.2.2. PWSQ versus AHP The next step is to compare the results of PWSQ and AHP variation models. With a quick overview it is easy to say that the most and least satisfying service quality dimensions were nearly the same in both models; customers were most happy with personnel contact quality and flexibility dimensions and they were least satisfied with the availability dimension in both models. No significant differences between the four customer categories in terms of customer satisfaction scores were detected in the previous data analysis sections; therefore it is not necessary to consider them in this comparison part of the paper. The major points that need to be compared here are the customer satisfaction levels, both in the overall and in terms of separate dimensions. It is quite important here to remind that the scales used in the two models are completely different from each other and therefore the customer satisfaction levels of PWSQ and AHP variation model should not be compared based on magnitude. 5.2.2.1. Distribution of satisfied and dissatisfied customers A solution to this might be comparing the number of satisfied and dissatisfied customers in the two models. Let’s first investigate the results in terms of individual dimensions. As it can be viewed in Table XVI, these numbers show significant difference in the PWSQ and AHP variation model; number of satisfied customers in PWSQ are consistently lower than

47

number of dissatisfied customers, whereas the amount of satisfied customers are much higher than the dissatisfied amount. All seven dimensions achieved significantly higher customer satisfaction in the AHP variation model than in PWSQ. It can easily be said that a lot of customers who seemed not happy with the service quality of NPW in PWSQ questionnaires seemed delighted with the service in the AHP questionnaire. In the overall, 84% of all the respondents implied that they were happy with NPW’s service quality (their perceptions better than initial expectations) in AHP variation model, while this amount dropped to only 14% satisfied customers in the PWSQ instrument (Table XVII).

PWSQ Satisfied customers Reliability Responsiveness Flexibility Availability Assurance Pers. Quality Tangibles 22 48 46 20 16 39 43 Dissatisfied customers 78 52 54 80 84 61 57

AHP variation model Satisfied customers 92 92 93 77 89 96 88 Dissatisfied customers 8 8 7 23 11 4 12

Table XVI. Distribution of satisfied and dissatisfied customers for the two models

Unweighted PWSQ Percentage of satisfied customers Percentage of dissatisfied customers 14 86

Weighted PWSQ 14 86

AHP variation model 84 16

Table XVII. Overall satisfaction percentages for the three analyses It is quite interesting to obtain such diverse results from two different survey instruments, by using exactly the same respondents and service items in both studies. There seems to be enough reason to believe that this discrepancy in the results is due to differences between the natures of two models. PWSQ forces the respondent first to think about his expectations from an ideal service provider and then to rate the service he actually receives from a particular service provider. AHP, on the other hand, demands that the respondent directly compares his

48

expectations and experiences about a service provider. The two scales might have been effective in directing how the customers express their thoughts in the questionnaires. PWSQ uses a 7 point Likert scale which included only numbers from 1 (Strongly disagree) to 7 (Strongly agree) to rate if NPW fulfils a certain feature. This scale might have caused the customers unintentionally express their satisfaction levels as quite low. The scale used in AHP questionnaires is quite different both in wording and magnitude; a respondent can choose from the alternatives listed in Table P that express opinions regarding their satisfaction in any one item of the questionnaire. Direct expressions in the scale such as “Very much better than your expectations” might have facilitated to indicate how the customers feel about NPW’s service quality and more positive results are obtained at the end. The base measure used in PWSQ was the gap score between perceptions and expectations, or simply P-E. This number can take on values between -6 and 6. A gap score greater than zero indicates that the customer is happy with the service quality since he received the same or better level of service that he initially expected. A gap score smaller than zero means that the customer is not satisfied with the service, since he could not get what he had expected. The overall average satisfaction score obtained at the end of the unweighted PWSQ analysis was -0.7255, with sample scores ranging between a minimum of -2.5595 and a maximum of 0.9524. The average score in the weighted analysis was -0.7185 in the overall and individual scores ranged between a minimum of -2.5846 and a maximum of 0.9372. As indicated above, the results of the unweighted and weighted analyses of PWSQ data are not very different from each other. What they actually imply is that the 100 respondents to the questionnaire were a bit dissatisfied on the average with NPW’s service quality. The base measure used in the AHP variation model was the ratio of perceptions over expectations, or P/E. This number can take on values between 1/9 and 9. This number should ideally be greater than or equal to 1 if a particular customer is satisfied with the service quality; a P/E score equal to 1 indicates that expectations and perceptions are exactly the same. A number between 1/9 and 1 would mean that the customer could not get what he initially expected. The overall average P/E score obtained at the end of AHP analysis was 3.5163, meaning that NPW’s service quality had been slightly better than its customers’ expectations. The sample scores ranged between a minimum of 0.4276 and a maximum of 8.2620; NPW had performed slightly worse than its least satisfied customer and had nearly showed an excellent performance by absolutely going beyond its happiest customer. The general picture in these results indicates that NPW satisfied its customers most of the time;

49

although it performed badly with respect to some customers’ opinions, in the overall it provided a service quality that is beyond customer expectations. Despite the fact that PWSQ and AHP questionnaires utilized quite different scales, the results can still be compared for their implications regarding customer satisfaction. The overall average customer satisfaction scores indicated common dissatisfaction (negative P-E score) and low levels of service quality in PWSQ, and common delight (P/E value greater than 1) and superior levels of service quality in AHP study. As indicated above, the significant difference between the results of the two studies might be due to several factors such as wording of the scales and question types. 5.2.2.2. Correlation analysis It might be a good idea to investigate the correlation between PWSQ and AHP results in terms of customer satisfaction levels. Although the scales in two methods were different from each other, sample data can still be used in estimating any relationships. First of all the unweighted PWSQ data and AHP variation model data are to be compared. Following a similar procedure like the one in previous section, a regression line must be obtained first. Analyzing the 100 sample data from both questionnaires, the regression line is found to be Y = 4.488 + 1.418 * X, where Y is AHP (P/E) score and X is unweighted (P-E) score. Then, using the above formula, r2 is: r2 =

(4.488)(351.63) + (1.418) * (−185.55) − (100)(3.516) 2 78.564 = = 0.165. 2 475.849 1712.29 − (100)(3.516)

This r2 value close to zero implies that there is little correlation between AHP and unweighted PWSQ data; only 16.5% of the variation in AHP variation model’s satisfaction scores (the dependent variable Y) can be explained by the variation in unweighted customer satisfaction scores (the independent variable X). The coefficient of correlation here is:

r = r 2 = 0.165 = 0.406.
This number indicates a positive but weak correlation between the two variables. These two coefficients explain why the satisfaction levels obtained from unweighted PWSQ and AHP data in the previous parts showed significant differences from each other. Now that unweighted and weighted PWSQ satisfaction scores and unweighted PWSQ and AHP variation model satisfaction levels are correlated, the last pair of scores to be analyzed is weighted PWSQ and AHP scores. The corresponding regression line is as follows: 50

Y = 4.542 + 1.427 * X,

where Y represents AHP (P/E) scores and X represents weighted (P-E) scores. The coefficient of determination is therefore equal to 0.178. Again, the value of r2 indicates little correlation between AHP and weighted PWSQ data; only 17.8% of the variation in AHP variation model’s satisfaction levels (the dependent variable Y) can be explained by the variation in weighted customer satisfaction levels (the independent variable X). The coefficient of correlation here is r = r 2 = 0.178 = 0.422. This r value points out that a positive but weak correlation exists between AHP and weighted PWSQ satisfaction levels. Together with the previous correlation analyses, these results base the significant differences between the amounts of satisfied people in PWSQ (14%) and AHP (86%) surveys to some statistical grounds. Table XVIII shows the results of the three correlation analyses. The closest relationship is between unweighted and weighted PWSQ results, indicating that most of the variance in weighted PWSQ gap scores can be explained by the variance in unweighted PWSQ gap scores. Either PWSQ data have weaker correlations with the AHP data. This table supports the view that analyzing gap scores by treating every dimension as the same would be sufficient to obtain accurate service quality measures. Weakness of the correlation between gap scores and perceptions-only scores imply that the variation between the two survey results should be explained by other factors.
Unweighted PWSQ vs. weighted PWSQ Coefficient of determination Coefficient of correlation 0.976 0.988 Unweighted PWSQ vs. AHP 0.165 0.406 Weighted PWSQ vs. AHP 0.178 0.422

Table XVIII. Comparison of the three correlations 5.3. The Effect of Gender In addition to the various analyses that are conducted for the unweighted and weighted PWSQ scores as well as the AHP satisfaction scores, yet another analysis was performed. The author decided to investigate whether gender has a role in explaining the variance in customer satisfaction scores of the 100 respondents. The female and male respondents are separated into two groups and their average overall satisfaction scores obtained from the PWSQ data (unweighted) are analyzed. The average score for female respondents was -0.7119, where it

51

was -0.7466 for males. It seems that there is no significant difference between the satisfaction scores of both genders. Although a comment could easily be made by simply looking at the two genders’ average scores, it is best to prove this claim by a statistical test. Therefore a two-tailed test was performed to understand whether the means of two samples (female and male customers) were same or not. First of all the respondents were separated into two groups being females (61) and males (39). A summary of the sample data are given in Table XIX. Then their unweighted gap scores were put into a two-tailed test of the following hypotheses: H0: H1:
1 1

= ≠

2 2

← Null hypothesis: there is no difference between means ← Alternative hypothesis: a difference exists between means ← Level of significance for testing this hypothesis
Mean customer satisfaction score of sample -0.7119 -0.7466 Standard deviation of sample 0.6370 0.6815 Size of sample 61 39

α = 0.05
Group Female Male

Table XIX. Sample data for unweighted analysis When this hypothesis test is illustrated graphically (Figure 4), the significance level of 0.05 corresponds to the two shaded areas, each of which contains 0.025 of the area. The acceptance region contains two equal areas of 0.475 each. Since both of our samples are greater than 30, we can use the normal distribution (Levin and Rubin, 1998). The normal distribution table suggests the critical value of z for 0.475 of the area under the curve to be 1.96.

1.96

1.96

0.255

Figure 4. Normal distribution curve for hypothesis testing

52

Since we do not know the standard deviations of the two populations, we have to estimate them as the sample standard deviations:

ˆ ˆ σ 1 = s1 = 0.637 ; σ 2 = s 2 = 0.681
Now the estimated standard error of the difference between the two means can be calculated by
σ 2 σ 2   (0.637 )2 (0.6815)2 ˆ ˆ ˆ x1 − x 2 =  1 + 2  =  σ  61 +  n 39 n2    1    = 0.01856  

= 0.136 ← Estimated standard error Next the difference of sample means, x1 − x 2 , is standardized. First, the hypothesized difference of the population means, ( 1- 2)H0, is subtracted and then it is divided by the

ˆ estimated standard error of the difference between the sample means, σ x1 − x 2 : z= (x

1

− x 2 − (µ 1 − µ 2 ) H 0 ˆ σ x −x
1 2

)

=

(− 0.7119 + 0.7466) − 0 = 0.255
0.136

Since the z value of 0.255 is well within the acceptance region depicted in Figure 4, the null hypothesis of no difference between sample means is accepted. Thus, the author’s claim that difference in genders does not help to explain the variance in average satisfaction scores is proved for the unweighted PWSQ analysis. Now let’s see if the picture is different with weighted satisfaction scores. A summary table about the two samples (female and male respondents) is provided in Table XX.
Group Female Male Mean customer satisfaction score of sample -0.7106 -0.7309 Standard deviation of sample 0.6365 0.6755 Size of sample 61 39

Table XX. Sample data for weighted analysis The following hypothesis will be tested regarding any differences between sample means: H0: H1:
1 1

= ≠

2 2

← Null hypothesis: there is no difference between means ← Alternative hypothesis: a difference exists between means ← Level of significance for testing this hypothesis

α = 0.05

53

The acceptance region in this case is exactly the same with the previous tests, since the levels of significance are the same, meaning that the critical value of z is again 1.96. The standardized z value is equal to 0.150, which lies within the acceptance region; so the null hypothesis of no difference between sample means is accepted. This time the insignificance of gender effects to explain the variance in average satisfaction scores is proved for the weighted PWSQ data. The final analysis to be conducted regarding gender of customers is about AHP satisfaction scores. The numerical values of (P/E) scores are analyzed with genders of their corresponding customers and any difference between the population means of female and male satisfaction scores is sought. As in the previous analyses, a two-tailed hypothesis testing will be made by using the following data in Table XXI.
Group Female Male Mean customer satisfaction score of sample 3.5639 3.4419 Standard deviation of sample 2.0398 2.4379 Size of sample 61 39

Table XXI. Sample data for AHP analysis The hypotheses to be tested this time are: H0: H1:
1 1

= ≠

2 2

← Null hypothesis: there is no difference between means ← Alternative hypothesis: a difference exists between means

The same level of significance (α = 0.05) is going to be used here as in the previous tests for convenience and ease of comparison across studies. This means that we have the same acceptance region and critical value of z (1.96) for this hypothesis test as well. The standardized z value is found to be 0.2597; again, this value is within our acceptance region, so the null hypothesis is accepted which means that there is no significant difference between the population means of female and male satisfaction scores (P/E). The three separate hypothesis testing efforts yielded the same result; a particular customer’s gender does not help at all to explain the variance in customer satisfaction scores in neither PWSQ nor AHP variation models. The efforts can be taken further by investigating any effects of gender on expectations and perceptions separately.

54

CHAPTER 6: Summary and Conclusions As consumers grow to be more and more demanding and the business environment becoming more competitive every day, it is essential that companies improve their offerings and show themselves to be better than competitors. Besides product quality, service quality has also become critical to customer satisfaction, whose variables have been researched for more than two decades now. The emphasis on the importance of service quality has forced many researchers to develop ways to measure service quality. This dissertation attempted to utilize two different tools to measure the service quality of a company in Turkey engaged pharmaceutical wholesaling business. First tool was a variation model of the famous SERVQUAL instrument of Parasuraman et al. (1988, 1991), which use the gaps between customers’ initial expectations from a service and their perceptions of the actual service as the measure of service quality. A comprehensive literature review is presented about the application areas of SERVQUAL, how it has been used for service quality measurement and how it has been modified by different researchers to obtain better fits with the industries in question. In the present study, modifications have been made to the original service quality scale by developing logistics attributes that fit into the previously defined features of the industry and by identifying additional gaps that could be applied to physical distribution service context. Through a detailed literature review as well as meetings with company management, a service quality measurement scale named PWSQ was developed which had reliability, responsiveness, flexibility, availability, assurance, personnel contact quality and tangibles as service quality dimensions. This scale has the same features with SERVQUAL

such as the seven-point Likert scale ranging from “Strongly Agree” to “Strongly Disagree” and two separate sections questioning customer expectations and perceptions, respectively. The main difference of PWSQ from the original SERVQUAL lies in its dimensions, as explained previously. This tool was tested on a sample of 100 customers of the focal company and service quality gaps existing between customer expectations and perceptions were identified for the seven dimensions. The second tool utilized Saaty’s (1980, 1990) Analytic Hierarchy Process (AHP) to understand customers’ importance attributions to the seven service quality dimensions through pairwise comparisons, and obtain performance ratings for these dimensions as compared to customers’ initial expectations. Due to the difficulty of performing a large number of pairwise comparisons and inconsistent responses obtained from an initial set of sample customers, the first part of the AHP model (pairwise comparison matrix for comparing

56

service quality dimensions) was replaced with a direct rating section which demanded the customer to simply rate the specified service quality dimensions according to their importance in service quality evaluations. The regular AHP principles were followed in second part of the model; respondents were to rate the performance of the focus company in terms of the 23 service items (same items with the PWSQ scale) as compared to their initial expectations. This variation model of AHP has quite a different perspective from the PWSQ model about measuring service quality; instead of using gap scores (P-E), it relied on direct performance ratings to measure service quality. This tool was presented to the 100 sample customers together with the PWSQ scale. Reliability, assurance and personnel contact quality were the dimensions that received the highest point allocations in the AHP variation model. Approximately half of the respondents (47%) assigned the lowest importance rankings to availability, meaning that the service provider’s ability to hold a large variety of items in stock is not as important as its capabilities in other issues such as consistency of order delivery times or conveying trust and confidence into the customer. This study has several implications. First of all, the original SERVQUAL dimensions – reliability, responsiveness, assurance, empathy and tangibles- had to be modified significantly in order to obtain a better fit to the specific industry in question, since some of the original 22 items in SERVQUAL did not apply to pharmaceutical wholesaling business as well as lacking a number of important features of the industry. This effort more or less verifies the claims that SERVQUAL dimensions have to be tailored for the specific needs of the service business in consideration, and for physical distribution service, specifically (Bienstock et al., 1997; Rafele, 2004; Franceschini and Rafele, 2000; Mentzer et al., 1999). A second implication of the study is about the usage of simple performance ratings instead of gap scores in service quality measurement. A number of researchers have argued that using perceptions-only measures would be sufficient to understand levels of customer satisfaction and service quality, instead of gathering expectations and perceptions of customers separately (Babakus and Boller, 1992; Cronin and Taylor, 1992; Teas, 1993; Brown et al., 1993; Mehta and Durvasula, 1998). This study involved the utilization of both gap scores and performance scores in the two separate service quality measurement tools, PWSQ and AHP variation models, respectively. Results of the two studies differed significantly in terms of customer satisfaction levels; responses to PWSQ implied consistently low service quality levels whose average was below customer expectations, whereas customers seemed quite happy with the service quality as reflected in the AHP responses. Moreover, the amount of correlation

57

between PWSQ gap scores (P-E) and AHP satisfaction scores (P/E) was found to be weak, meaning that the two sets of results had little relationship. It was surprising as well as interesting to obtain such diverse responses from the same set of customers for the same service attributes in two different questionnaires. These differences might be explained by the capabilities of the two techniques in reflecting true customer opinions. Based on the comparative analysis made between PWSQ and AHP results, it was felt that simple comparisons of perceptions with expectations reflect the customer opinions about service quality more accurately than the gap scores obtained from PWSQ. This was due to the behavioural aspect of customers selecting high points in the expectations section of PWSQ, and not leaving the service provider any chances to go beyond their expectations and delight them, although unintentionally. Since they constantly selected the highest expectations scores, the best possible result that the focus company could get in terms of customer satisfaction was matching expectations, only if respondents selected high perceptions scores. As a result, overall satisfaction scores obtained from the analysis of PWSQ were consistently lower than zero, indicating dissatisfaction which was confirmed by only 14% of the respondents being satisfied with the service quality and the rest (86%) being dissatisfied. However, the analysis of AHP results revealed that 84% of the sample customers were happy with the service they received from the focus company. Although there is no empirical evidence to prove that one method is definitely better than the other in measuring service quality, intuition and consideration of the survey results suggest that the AHP variation model reflects both the customer satisfaction levels and service quality measures more accurately than the PWSQ instrument. To verify the P-E and P/E scores obtained using the SERVQUAL and AHP models, some regression analyses have been accounted. The P-E or P/E scores have been regressed against the years of association of customers with the company. P/E scores of AHP were significantly positively related to years of association, while the P-E scores of SERVQUAL models had no significance with the years of association. This regression result could reinforce the belief that AHP results are better than SERVQUAL results. The test sample had been selected in a way to reflect the focus company’s customer range; all customers were categorized according to their monthly purchasing power from the company and equal numbers of customers were included in the test sample to maintain homogeneity. It was expected that customers with stronger business with the company would be more satisfied than the ones purchasing low amounts, depending on the idea that increased customer satisfaction would yield increased business (Anderson et al., 1996; Zeithaml et al.,

58

2006; Daugherty et al., 1998; Innis and La Londe, 1994; Stank et al., 2003). However, various analyses on the sample responses from customers belonging to the four categories implied no such relationships; neither PWSQ nor AHP results indicated that customers from higher categories were significantly happier with the service quality of the company than customers from lower categories. This may be due to several reasons such as a wrong categorization of customers or ignoring the differences between the sales capacities of customers. These issues can be addressed in a further study and any relationships between loyalty and business can be investigated. Another important point emphasized by the study results was that analyzing customer responses unweighted or weighted did not make any significant differences in terms of overall satisfaction scores. This issue was tested statistically and the resulting correlation analysis between unweighted and weighted gap scores suggested that 97.57% of the variation in weighted gap scores can be explained by the variation in unweighted gap scores. This was counter-intuitive since consumers may generally be assigning priorities to different aspects of a service and judge the overall service quality they receive according to these mental assignments. Many researchers performed weighted analyses to service quality measures (Teas, 1993; Cronin and Taylor, 1992; Mehta and Durvasula, 1998), including the importance weights of different service attributes, as in the present study. The results obtained from this study confirmed the previous studies implying that analyzing only unweighted SERVQUAL gap scores might be sufficient in understanding customer satisfaction (Mehta and Durvasula, 1998; Cronin and Taylor, 1992). As mentioned earlier, additional regression analyses performed on the sample data showed that a particular customer’s total years of association with the focus company has significant effects on the level of that customer’s satisfaction from the service quality of the company. On the other hand, the hypothesis that years of association has no significant impact on satisfaction scores was accepted in both unweighted and weighted PWSQ data analyses; these efforts did not reveal a considerable interaction between customer satisfaction and years of association. However, the responses of PWSQ surveys were considered to be misleading and AHP responses were taken as accurate reflections of customer satisfaction, where the association between length of business and satisfaction was confirmed. Finally, the effect of gender on customer satisfaction was investigated. Three separate hypothesis tests performed for unweighted and weighted PWSQ scores as well as for

59

responses to the AHP variation model yielded the same result that a customers’ genders does not help to explain the variation in customer satisfaction scores. This study attempted to use SERVQUAL to measure service quality in a physical distribution service setting in Turkey, and also has addressed the gap in the literature regarding the usage of analytic hierarchy process for service quality measurement purposes. The study would contribute to the present literature on service quality by providing two instruments- PWSQ as a variation model of SERVQUAL, and an AHP variation model- to measure physical distribution service quality; in fact, the tools can be modified to be used in other service industries. Implications regarding the usage of unweighted perceptions-only ratings instead of gap scores would enhance the current views about the inferiority of SERVQUAL in measuring service quality accurately (Cronin and Taylor, 1992). Furthermore, this study has contributed to the focus company, Nevzat Pharmaceutical Wholesaler, in a number of ways. First of all, the company gained insight into their customers’ expectations in terms of various service attributes, as well as their invaluable opinions about the company’s current service quality. By this way, NPW had the chance to identify the areas on which it has to focus its resources such as money and time in order to close the existing gaps between its customers’ expectations and perceptions. Current responses to both questionnaires can also be stored to track customer expectations and perceptions over time (Zeithaml et al., 2006). As for the AHP model, NPW can easily use the current methodology to extend the study for benchmarking with competitors. The sample responses obtained from this study would serve as base-line for future efforts in competitive benchmarking.

60

References

Ahmed, P.K. and Rafiq, M. 1998. Integrated benchmarking: a holistic examination of select techniques for benchmarking analysis. Benchmarking for Quality Management &
Technology 5(3): 225-242.

Alexandris, K., Dimitriadis, N. and Markata, D. 2002. Can perceptions of service quality predict behavioral intentions? An exploratory study in the hotel sector in Greece.
Managing Service Quality 12(4): 224-231.

Al Harbi, K.M. 2001. Application of the AHP in project management. International Journal of Project Management 19(1): 19–27.

Anderson, R.D., Crum, M.R. and Jerman, R.E. 1996. Relationships of work improvement program experience and logistics quality management factors. Transportation Journal 36(1): 20-34. Angur, M.G., Nataraajan, R. and Jahera Jr., J.S. 1999. Service quality in the banking industry: An assessment in a developing economy. International Journal of Bank Marketing 17(3): 116-123. Armacost, R.L., Componation, P.J., Mullens, M.A. and Swart, W.W. 1994. An AHP framework for prioritizing customer requirements in QFD: An industrialized housing application. IIE Transactions 26(4): 72–79. Asubonteng, P., McCleary, K.J. and Swan, J.E. 1996. SERVQUAL revisited: a critical review of service quality. Journal of Services Marketing 10(6): 62-81. Babakus, E. and Boller, G.W. 1992. An empirical assessment of the SERVQUAL scale.
Journal of Business Research 24(3): 253-268.

Babakus, E. and Mangold, W.G. 1992. Adapting the SERVQUAL scale to hospital services: an empirical investigation. Health Services Research 26(2): 767-786. Badri, M.A. 2001. A combined AHP-GP model for quality control systems. International
Journal of Production Economics 72(1): 27–40.

Belton, V. and Gear, T. 1983. On a short-coming of Saaty's method of analytic hierarchies. Omega 11(3): 228-230.

61

Bienstock, C.C., Mentzer, J.T. and Bird, M.M. 1997. Measuring physical distribution service quality. Journal of the Academy of Marketing Science 25(1): 31-44. Brown, T.J., Churchill Jr., G.A. and Peter, J.P. 1993. Improving the measurement of service quality. Journal of Retailing 69(1): 127-139. Brysland, A. and Curry, A. 2001. Service improvements in public services using SERVQUAL. Managing Service Quality 11(6): 389-401. Buttle, F. 1996. SERVQUAL: review, critique, research agenda. European Journal of
Marketing 30(1): 8-32.

Carman, J.M. 1990. Consumer perceptions of service quality: an assessment of the SERVQUAL dimensions. Journal of Retailing 66(1): 33-55. Chiu, C.C., Lewis, B.R. and Park, W. 2003. Service quality measurement in the banking sector in South Korea. International Journal of Bank Marketing 21(4): 191-201. Chow, C.C. and Luk, P. 2005. A strategic service quality approach using analytic hierarchy process. Managing Service Quality 15(3): 278-289. Chow, G., Heaver, T.D. and Henriksson, L.E. 1994. Logistics performance: definition and measurement. International Journal of Physical Distribution & Logistics Management 24(1): 17-28. Clow, K.E. and Vorhies, D.W. 1993. Building a competitive advantage for service firms.
Journal of Services Marketing 7(1): 22-32.

Collins, A., Henchion, M. and O’Reilly, P. 2001. Logistics customer service: performance of Irish food exporters. International Journal of Retail & Distribution Management 29(1): 6-15. Cronin, J.J.Jr. and Taylor, S.A. 1992. Measuring service quality: A reexamination and extension. Journal of Marketing 56(3): 55-68. Daugherty, P.J., Stank, T.P. and Ellinger, A.E. 1998. Leveraging logistics/distribution capabilities: the effect of logistics service on market share. Journal of Business
Logistics 19(2): 35-51.

Donnelly, M., Kerr,N.J., Rimmer, R. and Shiu, E.M. 2006. Assessing the quality of police services using SERVQUAL. Policing: An International Journal of Police Strategies &
Management 29(1): 92-105.

62

Durvasula, S., Lysonski, S. and Mehta, S.C. 1999. Testing the SERVQUAL scale in the business-to-business sector: The case of ocean freight shipping service. The Journal of
Services Marketing 13(2): 132-150.

Emerson, C.J. and Grimm, C.M. 1996. Logistics and marketing components customer service: An empirical test of the Mentzer, Gomes and Krapfel model. International Journal of
Physical Distribution & Logistics Management 26(8): 29-42.

Franceschini, F., Cignetti, M. and Caldara, M. 1998. Comparing tools for service quality evaluation. International Journal of Quality Science 3(4): 356-367. Franceschini, F. and Rafele, C. 2000. Quality evaluation in logistics services. International
Journal of Agile Management Systems 2(1): 49-53.

Fung, P. and Wong, A. 1998. Case study: managing for total quality of logistics services in the supply chain. Logistics Information Management 11(5): 324-329. Gabbie, O. and O’Neill, M.A. 1996. SERVQUAL and the Northern Ireland hotel sector: a comparative analysis- part 1. Managing Service Quality 6(6): 25-32. Gagliano, K.B. and Hathcote, J. 1994. Customer expectations and perceptions of service quality in retail apparel specialty stores. Journal of Services Marketing 8(1): 60-69. Grapentine, T. 1998. The history and future of service quality assessment. Marketing
Research 10(4): 4-20.

Grönroos, C. 1984. A service quality model and its marketing implications. European Journal of Marketing 18(4): 36-44.

Innis, D.E. and La Londe, B.J. 1994. Customer service: the key to customer satisfaction, customer loyalty, and market share. Journal of Business Logistics 15(1): 1-27. Juwaheer, T.D. 2004. Exploring international tourists’ perceptions of hotel operations by using a modified SERVQUAL approach- a case study of Mauritius. Managing Service
Quality 14(5): 350-364.

Kang, G. 2006. The hierarchical structure of service quality: integration of technical and functional quality. Managing Service Quality 16(1): 37-50. Kettinger, W.J. and Lee, C.C. 1994. Perceived service quality and user satisfaction with the information services functions. Decision Sciences 25(5/6): 737-766.

63

Kettinger, W.J., Lee, C.C. and Lee, S. 1995. Global measures of information service quality: A cross-national Sstudy. Decision Sciences 26(5): 569-588. Koksal, G. and Egitman, A. 1998. Planning and design of industrial engineering education quality. Computers and Industrial Engineering 35(3/4): 639–642. Kong, R. and Mayo, M.C. 1993. Measuring service quality in the business-to-business context. Journal of Business & Industrial Marketing 8(2): 5-15. Korpela, J. and Tuominen, M. 1996. A decision aid in warehouse site selection. International
Journal of Production Economics 45(1–3): 169–180.

Kuei, C. and Lu, M.H. 1997. An integrated approach to service quality improvement.
International Journal of Quality Science 2(1): 24-36.

Lai, V.S., Wong, B.K. and Cheung, W. 2002. Group decision making in a multiple criteria environment: A case using the AHP in software selection. European Journal of
Operational Research 137(1): 134–144.

Lambert, D.M., Lewis, M.C. and Stock, J.R. 1993. How shippers select and evaluate general commodities LTL motor carriers. Journal of Business Logistics 14(1): 131-143. Lehtinen, J.R. and Lehtinen, O. 1982. Service quality: A study of quality dimensions, Unpublished working paper. Service Management Institute, Helsinki. Long, M. and McMellon, C. 2004. Exploring the determinants of retail service quality on the Internet. Journal of Services Marketing 18(1): 78-90. Maltz, A. and Maltz, E. 1998. Customer service in the distributor channel empirical findings.
Journal of Business Logistics 19(2): 103-128.

Mehta, S.C. and Durvasula, S. 1998. Relationships between SERVQUAL dimensions and organizational performance in the case of a business-to-business service. Journal of
Business & Industrial Marketing 13(1): 40-53.

Mentzer, J.T., Gomes, R. and Krapfel, R.E.Jr. 1989. Physical distribution service: a fundamental marketing concept? Journal of the Academy of Marketing Science 17(1): 53-62. Mentzer, J.T., Flint, D.J. and Kent, J.L. 1999. Developing a logistics service quality scale.
Journal of Business Logistics 20(1): 9-32.

64

Mentzer, J.T., Flint, D.J. and Hult, G.T.M. 2001. Logistics service quality as a segmentcustomized process. Journal of Marketing 65(4): 82-104. Min, H. and Min, H. 1996. Competitive benchmarking of Korean luxury hotels using the analytic hierarchy process and competitive gap analysis. Journal of Services
Marketing 10(3): 58-72.

Min, H. and Min, H. 1997. Benchmarking the quality of hotel services: managerial perspectives. International Journal of Quality & Reliability Management 14(6): 582597. Min, H., Min, H. and Chung, K. 2002. Dynamic benchmarking of hotel service quality.
Journal of Services Marketing 16(4): 302-321.

Novack, R.A., Rinehart, L.M. and Langley Jr., C.J. An internal assessment of logistics value.
Journal of Business Logistics 15(1): 113-152.

Parasuraman, A., Zeithaml, V.A. and Berry, L.L. 1985. A conceptual model of service quality and its implications for future research. Journal of Marketing 49(Fall): 41-50. Parasuraman, A., Zeithaml, V.A. and Berry, L.L. 1988. SERVQUAL: A multiple-item scale for measuring consumer perceptions of service quality. Journal of Retailing 64(Spring): 12-40. Parasuraman, A., Zeithaml, V.A. and Berry, L.L. 1991. Refinement and reassessment of the SERVQUAL scale. Journal of Retailing 67(Winter): 420-450. Pariseau, S.E. and McDaniel, J.R. 1997. Assessing service quality in schools of business.
International Journal of Quality & Reliability Management 14(3): 204-218.

Rafele, C. 2004. Logistic service measurement: a reference framework. Journal of
Manufacturing Technology Management 15(3): 280-290.

Rahman, S. 2006. Quality management in logistics: an examination of industry practices.
Supply Chain Management: An International Journal 11(3): 233-240.

Ramanathan, R. and Ganesh, L.S. 1995. Energy resource allocation incorporating quantitative and qualitative criteria: An integrated model using goal programming and AHP. SocioEconomic Planning Sciences 29(3): 197-218.

Richard, M.D. and Allaway, A.W. 1993. Service quality attributes and choice behavior.
Journal of Services Marketing 7(1): 59-62.

65

Saaty, T.L. 1980. The Analytic Hierarchy Process. McGraw-Hill, New York, NY. Saaty, T.L. 1990. How to make a decision: The Analytic Hierarchy Process. European
Journal of Operational Research 48: 9-26.

Saaty, T.L., Vargas, L.G. and Dellmann, K. 2003. The allocation of intangible resources: the analytic hierarchy process and linear programming. Socio-Economic Planning
Sciences 37(3): 169–184.

Sahney, S., Banwet, D.K. and Karunes, S. 2004. A SERVQUAL and QFD approach to total quality education: A student perspective. International Journal of Productivity and
Performance Management 53(2): 143-166.

Seth, N., Deshmukh, S.G. and Vrat, P. 2006. A framework for measurement of quality of service in supply chains. Supply Chain Management: An International Journal 11(1): 82-94. Shapiro, R.D. and Heskett, J.L. 1985. Logistics Strategy: Cases and Concepts. West Publishing, St. Paul: MN. Stank, T.P., Goldsby, T.J. and Vickery, S.K. 1999. Effect of service supplier performance on satisfaction and loyalty of store managers in the fast food industry. Journal of
Operations Management 17(4): 429-447.

Stank, T.P., Goldsby, T.J., Vickery, S.K. and Savitskie, K. 2003. Logistics service performance: estimating its influence on market share. Journal of Business Logistics 24(1): 27-55. Suresh, N.C. and Kaparthi, S. 1992. Flexible automation investments: A synthesis of two multi-objective modeling approaches. Computers and Industrial Engineering 22(3): 257–272. Tam, M.C.Y. and Tummala, V.M.R. 2001. An Application of the AHP in vendor selection of a telecommunications system. Omega 29(2): 171–182. Teas, R.K. 1993. Expectations, performance evaluation, and consumers’ perceptions of quality. Journal of Marketing 57(4): 18-34. Tucker, F.G. 1980. Customer service in a channel of distribution: the case of the manufacturer-wholesaler-chain drug retailer channel in the prescription drug industry, Unpublished PhD thesis. Ohio State University, Columbus, OH.

66

Vaidya, O.S. and Kumar, S. 2006. Analytic hierarchy process: an overview of applications.
European Journal of Operational Research 169(1): 1-29.

Vandamme, R. and Leunis, J. 1993. Development of a multiple-item scale for measuring hospital service quality. International Journal of Service Industry Management 4(3): 30-49. Wall, T.D. and Payne, R. 1973. Are deficiency scores deficient? Journal of Applied
Psychology 58: 322-326.

Winston, W. L. 2004. Operations Research: Applications and Algorithms, 4th ed. Thomson Learning, Inc., London. Xing, Y. and Grant, D.B. 2006. Developing a framework for measuring physical distribution service quality of multi-channel and “pure player” internet retailers. International
Journal of Retail & Distribution Management 34(4/5): 278-289.

Youssef, F., Nel, D. and Bovaird, T. 1995. Service quality in NHS hospitals. Journal of
Management in Medicine 9(1): 66-74.

Zahedi, F. 1986. The analytic hierarchy process- A survey of the method and its applications.
Interfaces 16(4): 96-108.

Zeithaml, V.A., Parasuraman, A. and Berry, L.L. 1990. Delivering Quality Service:
Balancing Customer Perceptions and Expectations. Free Press, New York, NY.

Zeithaml, V.A., Bitner, M.J. and Gremler, D.D. 2006. Services Marketing: Integrating
Customer Focus Across the Firm, 4th edition. McGraw-Hill/Irwin, New York, NY.

67

Appendices Appendix A The PWSQ Questionnaire
Expectations Section DIRECTIONS: Please think about the kind of pharmaceutical wholesaler company that would deliver excellent quality of distribution service and with which you would be pleased to do business. Please show the extent to which you think such a pharmaceutical wholesaler company would possess the feature described by each statement. If you feel a feature is not at all essential for excellent pharmaceutical wholesalers such as the one you have in mind, circle the number “1”. If you feel a feature is absolutely essential for excellent pharmaceutical wholesalers, circle “7”. If your feelings are less strong, circle one of the numbers in between. There are no right or wrong answers- all we are interested in is a number that truly reflects your feelings regarding pharmaceutical wholesaler companies that would deliver excellent quality of service. Strongly Disagree Reliability E1. Excellent pharmaceutical wholesalers will deliver orders within the time frame they promise to do so. E2. The time between placing and receiving orders will be consistent when working with excellent pharmaceutical wholesalers. E3. Excellent pharmaceutical wholesalers will perform a service right the first time. E4. Excellent pharmaceutical wholesalers will deliver accurate orders, without any items missing or redundant. E5. Excellent pharmaceutical wholesalers will deliver orders in good condition (undamaged). E6. Invoices will match purchase orders when working with excellent pharmaceutical wholesalers. Responsiveness E7. Excellent pharmaceutical wholesalers will deliver my orders regardless of the amount. (no minimum order size constraints) E8. Excellent pharmaceutical wholesalers will be able to respond to unexpected/urgent orders. E9. Excellent pharmaceutical wholesalers will always be willing to help their customers. Flexibility E10. Excellent pharmaceutical wholesalers will have flexible and easy order placement procedures. 1 2 3 4 5 Strongly Agree 6 7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

68

Appendix A – cont’d
Strongly Disagree E11. Excellent pharmaceutical wholesalers will have flexible payment methods. E12. Excellent pharmaceutical wholesalers will offer flexible payment conditions. Availability E13. Excellent pharmaceutical wholesalers will always have my ordered items available in stock. E14. Excellent pharmaceutical wholesalers will track new products and will hold a wide variety of products in stock. Assurance E15. Return procedures of excellent pharmaceutical wholesalers will be convenient and easily applicable. E16. Excellent pharmaceutical wholesalers will offer sales conditions (surpluses, discounts, due dates etc.) that are convenient with the market and advantageous. E17. Customers of excellent pharmaceutical wholesalers will feel safe in their transactions with the company. Personnel Contact Quality E18. Employees of excellent pharmaceutical wholesalers will be consistently courteous with customers. E19. Employees of excellent pharmaceutical wholesalers will have the knowledge to answer customer questions. E20. Employees of excellent pharmaceutical wholesalers will tell customers exactly when services will be performed. E21. Communication officers of excellent pharmaceutical wholesalers will be clear, comprehensible and facilitating communication. Tangibles E22. Employees of excellent pharmaceutical wholesalers will be neat-appearing. E23. Excellent pharmaceutical wholesalers will have modernlooking equipment. 1 2 3 4 5 Strongly Agree 6 7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

69

Appendix A – cont’d
Perceptions Section DIRECTIONS: The following set of statements relate to your feelings about Nevzat Pharmaceutical Wholesaler’s distribution service. For each statement, please show the extent to which you believe Nevzat Pharmaceutical Wholesaler (NPW) has the feature described by the statement. Once again, circling a “1” means you strongly disagree that NPW has that feature, and circling a “7” means you strongly agree. You may circle any of the numbers in between that show how strong your feelings are. There are no right or wrong answers- all we are interested in is a number that best shows your perceptions about NPW’s service quality. Strongly Disagree Reliability P1. NPW deliver orders within the time frame they promise to do so. P2. The time between placing and receiving orders is consistent when working with NPW. P3. NPW perform a service right the first time. P4. NPW delivers accurate orders, without any items missing or redundant. P5. NPW deliver orders in good condition (undamaged). P6. Invoices match purchase orders when working with NPW. Responsiveness P7. NPW delivers my orders regardless of the amount. (no minimum order size constraints) P8. NPW is able to respond to your unexpected/urgent orders. P9. NPW is always willing to help you. Flexibility P10. NPW has flexible and easy order placement procedures. P11. NPW has flexible payment methods. P12. NPW offers flexible payment conditions. Availability P13. NPW always has my ordered items available in stock. P14. NPW tracks new products and holds a wide variety of products in stock. 1 2 3 4 5 Strongly Agree 6 7

1

2

3

4

5

6

7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1

2

3

4

5

6

7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1 1

2 2 2

3 3 3

4 4 4

5 5 5

6 6 6

7 7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

70

Appendix A – cont’d
Strongly Disagree Assurance P15. Return procedures of NPW are convenient and easily applicable. P16. NPW offers sales conditions (surpluses, discounts, due dates etc.) that are convenient with the market and advantageous. P17. You feel safe in your transactions with NPW. Personnel Contact Quality P18. Employees of NPW are consistently courteous with you. P19. Employees of NPW have the knowledge to answer your questions. P20. Employees of NPW tell me exactly when services will be performed. P21. Communication officers of NP are clear, comprehensible and facilitating communication. Tangibles P22. Employees of NPW are neat-appearing. P23. NPW has modern-looking equipment. Strongly Agree

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

71

Appendix B The AHP Questionnaire An AHP Questionnaire for comparing a customer's opinions of PSQ with ESQ
Nevzat Pharmaceutical Wholesaler (NPW) values your opinion to improve the quality of its service. We request you to fill this feedback form. Please tick the appropriate box for each of the questions below. The first part is concerned with obtaining your views on the significance of some features expected from NPW. In the second part, your views on the actual levels of NPW’s service are compared with your expectations from NPW. Part I In assessing the services rendered by NPW, the following are significant features for your assessment. 1. Reliability 3. Flexibility4. Availability 5. Assurance The ability to perform the promised service both dependably and accurately The flexibility to allow for different transaction options and methods. The availability of products in stock 2. Responsiveness - The willingness to help customers and to provide prompt service

fff

The ability to convey trust and confidence into customers and make them feel that they are receiving good service

6. Personnel contact The knowledge and courtesy of employees as well as their ability to ease iqualitycommunication with customers 7. Tangibles The appearance of physical facilities, equipment, personnel, and communication materials.

How will you compare the importance of "Reliability" with that of other features? Please use the table below to pick the number that best defines your opinions. If you feel that Reliability has very strongly preference over the other feature being compared Reliability has strong preference over the other feature being compared Reliability has definite preference over the other feature being compared Reliability has weak preference over the other feature being compared Reliability has the same level of preference as the other feature being compared The other feature being compared has weak preference over Reliability The other feature being compared has definite preference over Reliability The other feature being compared has strong preference over Reliability The other feature being compared has very strong preference over Reliability Select 9 7 5 3 1 1/3 1/5 1/7 1/9

72

Appendix B- cont’d
Please write the appropriate number in the table below that shows your comparisons.
Reliability Reliability Responsiveness Flexibility Availability Assurance Pers.Quality Tangibles

Using a similar procedure, please compare the right hand entities with other features in the table below. Please note that you need not fill the shaded cells.
Responsiveness Flexibility Availability Assurance Pers. Quality Tangibles

Part II We would like to record the performance of NPW as compared to your initial expectations in the following questions below. Very much worse than your expectations 7

Very much better than your expectations

Slightly worse than your expectations

Slightly better than your expectations

Much worse than your expectations

Much better than your expectations

They just met your expectations

1/9 1/7 1/5 1/3 1. In terms of delivery of orders within the promised time frame, how would you rate the performance of NPW as compared to your initial expectations? In terms of the consistency of the length of time between placing and receiving orders, how would you rate the performance of NPW as compared to your initial expectations? In terms of their ability to perform a service right the first time, how would you rate the performance of NPW as compared to your initial expectations?

1

3

5

2.

3.

Worst performance 9

Excellent

73

Appendix B – cont’d

Very much worse than your expectations 7

Very much better than your expectations

Slightly worse than your expectations

Slightly better than your expectations

Much worse than your expectations

Much better than your expectations

They just met your expectations

1/9 1/7 1/5 1/3 4. In terms of their ability to deliver accurate orders (without any items missing or redundant), how would you rate the performance of NPW as compared to your initial expectations? In terms of their ability to deliver orders in good condition (undamaged), how would you rate the performance of NPW as compared to your initial expectations? In terms of the level of match between invoices and your purchase orders, how would you rate the performance of NPW as compared to your initial expectations? In terms of delivering all orders regardless of the amount (minimum order constraints), how would you rate the performance of NPW as compared to your initial expectations? In terms of its ability to respond to your unexpected/urgent orders, how would you rate the performance of NPW as compared to your initial expectations? How would you rate the willingness of NPW to help you, as compared to your initial expectations? In terms of flexibility and ease in its order placement procedures, how would you rate NPW as compared to your initial expectations?

1

3

5

5.

6.

7.

8.

9.

10.

Worst performance 9

Excellent

74

Appendix B – cont’d

Very much worse than your expectations 7

Very much better than your expectations

Slightly worse than your expectations

Slightly better than your expectations

Much worse than your expectations

Much better than your expectations

They just met your expectations

1/9 1/7 1/5 1/3 11. In terms of its flexibility in payment methods, how would you rate NPW as compared to your initial expectations? In terms of its flexibility in payment conditions, how would you NPW as compared to your initial expectations? In terms of the availability of your orders in the warehouse, how would you rate NPW as compared to your initial expectations? In terms of its tracking of the new products and the variety of products it holds, how would you rate NPW as compared to your initial expectations? How would you rate the convenience and applicability of NPW’s return procedures as compared to your initial expectations? In terms of its sales conditions (surpluses, discounts, due dates etc.) being convenient with the market and advantageous, how would you rate NPW as compared to your initial expectations? In terms of feeling safe and assured in your transactions with the company, how would you rate NPW’s performance as compared to your initial expectations? How would you rate the courtesy of NPW’s employees, as compared to your initial expectations?

1

3

5

12.

13.

14.

15.

16.

17.

18.

Worst performance 9

Excellent

75

Appendix B – cont’d

Very much worse than your expectations 7

Very much better than your expectations

Slightly worse than your expectations

Slightly better than your expectations

Much worse than your expectations

Much better than your expectations

They just met your expectations

1/9 1/7 1/5 1/3 19. In terms of the adequacy of knowledge in answering your questions, how would you rate NPW’s employees as compared to your initial expectations? In terms of providing you exact information on the timing of a service, how would you rate the NPW’s employees as compared to your initial expectations? In terms of being clear, comprehensible and facilitating communication, how would you rate the communication officers of NPW, as compared to your initial expectations? In terms of neatness of appearance of its employees, how would you rate NPW as compared to your initial expectations? How would you rate the modern-looking nature of NPW’s equipment as compared to your initial expectations?

1

3

5

20.

21.

22.

23.

Worst performance 9

Excellent

76

Appendix C The modified section of the AHP questionnaire
An AHP Questionnaire for comparing a customer's opinions of PSQ with ESQ Nevzat Pharmaceutical Wholesaler (NPW) values your opinion to improve the quality of its service. We request you to fill this feedback form. Please tick the appropriate box for each of the questions below. The first part is concerned with obtaining your views on the significance of some features expected from NPW. In the second part, your views on the actual levels of NPW’s service are compared with your expectations from NPW. Part I In assessing the services rendered by NPW, the following are significant features for your assessment. 1. Reliability The ability to perform the promised service both dependably and accurately

2. Responsiveness - The willingness to help customers and to provide prompt service 3. Flexibility4. Availability 5. Assurance The flexibility to allow for different transaction options and methods. The availability of products in stock The ability to convey trust and confidence into customers and make them feel that they are receiving good service

6. Personnel contact The knowledge and courtesy of employees as well as their ability to ease qualitycommunication with customers 7. Tangibles The appearance of physical facilities, equipment, personnel, and communication materials.

How important are these factors to you in shaping your evaluations about the service quality of a pharmaceutical wholesaler? Please rate the seven factors according to their importance to you by giving each of them points between a minimum of 0 and a maximum of 100. Factor 1- Reliability 2- Responsiveness 3- Flexibility 4- Availability 5- Assurance 6- Personnel Contact Quality 7- Tangibles Points

77

Appendix D Sample AHP Questionnaire

78

Appendix D – cont’d

79

Appendix D – cont’d

80

Appendix E Sample PWSQ Questionnaire

81

Appendix E – cont’d

82

Appendix E – cont’d

83

Appendix E – cont’d

84