1
EvaluatingRetailing Delivery Service Quality
UsingAnalytical Hierarchy Process
Yu-Kai Huang, NanhuaUniversity
June 2008
Send correspondence to Yu-Kai Huang, Institute of Publishing & Culture Enterprise Management, Nanhua UniversityNo.32, Chong Keng Li, Dalin, Chia-Yi Shien, 622 (Taiwan) Tel: +886-988007614Fax:+886-5-2427139(Email: ).
ABSTRACT
In the logistics system of electronic commerce, the major difference between Taiwan and other countries is the retailing delivery (RD) system. The contributions of this paper are explores the structure of logistic for RD service for electronic commerce and develops a hierarchical framework and evaluation model to evaluate the logistics service quality of two RD providers in Taiwan. AHP and TOPSIS are respectively applied to determine the relative weights of evaluation criteria and the best compromise alternative. Four objectives and sixteen evaluating criteria are developed in this study. The dimension of objectives includes (1) barrier dimensions of the delivery system, (2) the delivery service, (3) the operations procedures and (4) logistics information. Secondly, we ranking the sixteen criteria for the RD provider of 7-11.com and CVS.com using the TOPSIS method and discuss the advantages and disadvantages of the 7-11.com and CVS.com services. Finally, this study used Importance-Performance Analysis (IPA) to discuss the relations of level of these sixteen criteria.Results and implications of these findings for retailing delivery providers are discussed.
Keywords: Retailing Delivery, AHP, Convenience Store, Electronic Commerce, TOPSIS
2. Literature review and research methodology
Electronic commerce (EC) is an increasingly popular shopping model for today’s businesses. In contrast to business-to-business (B2B)environment, the delivery policy of business-to-consumer (B2C) environment is different. The orders in an online B2C environment are small in size, instantaneous, ever changing, and placed by numerous consumers. Electronic commerce has been widely discussed in both academic and practical fields. In the past several years there have been many studies published concerning the difference between real space and cyber-space (e.g. RaghuandNormal, 2005; Rutner, Gibson and Williams, 2003; Khouja, 2001; Du, Li and Chou, 2005). Research has described the relative characteristic of Internet shopping and purchasing at off-line stores (Trunk, 2005; Ahn, Ryu and Han, 2004), the methods of payment and what customers purchase in cyber-shops (Josephand Elliot, 2005; Bremserand Chung,2005), and connections between characteristics of buyers and their purchasing behaviour (Peterson, 1997). However, there is a lack of research on the retailing delivery for the online B2C environment. Logistics can be considered as a service industry. In Taiwan, the retailing delivery (RD) system plays an important role in the logistics system for e-commerce. It is significant for e-retailing to evaluate the retailing delivery service quality. Different customers have a wide range of perceptions toward quality service, depending on their preference structures and roles in the process. Since the service industry contains is characterized by intangibility, inseparability and heterogeneity, it is more difficult to measure service quality. Service quality can be regarded as a composite of various attributes; including many intangible attributes that are difficult to measure. In order to overcome the issue, previous studies suggested using AHP to obtain criteria weights (e.g. Tsaur, Chang and Yen, 2002; Jukka, Anttiand Markku, 2001; Chen and Tzeng, 2004) and ranking through TOPSIS (Chiou, Tzeng and Cheng, 2005; Feng and Wang, 2006; Xu, 2001). The following section briefly reviews the conceptual approaches of AHP and TOPSIS.
2.1The analytical hierarchy process
The Analytical Hierarchy Process (AHP) is a decision-aiding method developed by Saaty in the 1970s (Saaty, 1980), which is a widely popular technique employed to model subjective decision-making processes based on the multiple. It aims at quantifying relative priorities for a given set of alternatives on a ratio scale, based on the judgment of the decision-maker, and stresses the importance of the intuitive judgments of a decision-maker as well as the consistency of the comparison of alternatives in the decision-making process. One of the main advantages of AHP is the relative ease with which it handles multiple criteria. In addition, the method is easier to understand and it can effectively handle both qualitative and quantitative data. In this paper, we used the AHP method to derive the weights of the evaluation criteria.The AHP has been successfully applied to resolve problems in business decisions like prioritizing corporate objectives, buying equipment, assigning management personnel, deciding on inventory levels, getting the best source for borrowing funds, finding markets and determining mergers and acquisitions (Saaty, 1980).
Basically, AHP has three underlying concepts: Structuring the complex decision problem as hierarchy of goal, criteria and alternatives, pair-wise comparison of elements at each level of the hierarchy with respect to each criterion on the preceding level, and finally vertically synthesizing the judgements over the different levels of the hierarchy. The steps of AHP process can be described briefly as the following steps (Chan, Kwok and Duffy, 2004; Triantaphyllou and Lin, 1995; Muller and Fairlie, 2001):
Step 1: Define the problem and determine its goal, then set up the hierarchical system by decomposing the problem into a hierarchy of interrelated elements.
Step 2: Structure the hierarchy from the top through to the lowest level which usually contains the list of alternatives.That is, from the overall objective to the intermediate level(s) factors or criteria to the lowest level.
Step 3: Assume that we have n different and independent alternative (A1, A2 …An) and they have the weights (W1, W2 …Wn). Construct a set of pair-wise comparison matrix A as the following:
(1)
The values assigned to aij according to Saaty scale are usually in the interval of 1-9 or their reciprocals as presents in Table 1.
______
Table 1 about here.
______
Step 4: Having recorded the numerical judgements aij in the matrix A, the matrix A is expressed as the following equation:
(2)
where aij=Wi/Wj, aij=1/aij, and the weight vector W = (W1, W2 …Wn). Saaty computes W as the principal right eigenvector of the matrix A, that is, AW=λmaxW, λmax is the principle eigenvalue of the matrix A.
Step 4: Hierarchical synthesis is now used to weight the eigenvectors by the weights of the criteria and the sum is taken over all weighted eigenvector entries corresponding to those in the next lower level of the hierarchy.
Step 5: Having made all the pair-wise comparisons, the eigenvector methods yields a natural measure of consistency. Saaty defined the consistency index (CI) as following: CI=(λmax -n)/(n-1). For each size of matrix n, random matrices were generated and their mean CI value, called the random index (RI), was computed and tabulated as shown in Table 2. Saaty defined the consistency ratio as the following: CR=CI/RI.
______
Table 2 about here.
______
Step 6: The consistency ratio CR is a measure of how a given matrix compares to a purely random matrix in terms of their consistency indices. A value of the consistency ratio CR≦0.1 is considered acceptable.
2.2The technique for order preference by similarity to ideal solution
Hwang and Yoon (1981) provide the technique for order preference by similarity to ideal solution (TOPSIS). The advantage of the method is simple and yields an indisputable preference order. The basic concept of this method is that ideal solution and the farthest distance from the negative-ideal solution in a geometrical sense. The optimal alternative is the one, which is closet to the ideal solution and farthest the negative ideal solution. Take the objective space of the two criteria as example which is indicated in Fig. 1, A+ and A- are, respectively, the ideal solution and negative ideal solution, and alternative A1 is shorter in distance ideal solution (A-) than alternatives A2. As a matter of fact, the ups and downs of these two alternatives are beyond comparison, only TOPSIS has defined such “relative closeness” so as to consider and correlated, as a whole, the distance to the ideal solution and the negative ideal solution. The TOPSIS procedure consists of the following steps:
______
Figure1 about here.
______
Step 1: Normalization of indicator values. Calculate the normalized decision matrix and the formula is as follows as in Eq. (3)
(3)
where i is the ith alternative, j is the jth evaluation indicator, rij is the indicator value after vector normalization for the ith alternative and jth evaluation indicator, xij is the original value of indicators for the ith alternative and jth evaluation indicator and m is the number of alternative.
Step 2: Calculate the weighted normalized decision matrix. The weighted normalized value vijis calculated as in the following:
vij = wirij, i = 1,…,n; j = 1,…,k (4)
where wi is the weight of the ith attribute or criterion
Step 3: To determine ideal (A+) and worst (A-) solution
(5)
(6)
J = { j = 1, 2, …, k|k belongs to benefit criteria}, benefit criteria implies a larger indicator value and a higher performance sore; J’ = { j = 1, 2, …, k|k belongs to cost criteria}, cost criteria implies a smaller indicator value and a higher performance score.
Step 4: To calculate the separation measure, using the n-dimensional Euclidean distance. The separation of each alternative from the ideal one (Si+) and the worst one (Si-) is given by
(7)
(8)
Step 5: Calculate the relative closeness to the ideal solution (Ci*) as in
(9)
where 0≦Ci*≦1 that is, an alternative i is closer to A+ as Ci* approaches to 1.
Step 6: To rank the preference order according to the descending order of Ci*.
4. Establishing the evaluation criteria
The typical multiple criteria evaluation problem focuses on a set of feasible alternatives and considers more than one criterion to determine a priority ranking for alternative implementation. The AHP allows group decision making, where group members can use their experience, values and knowledge to break down a problem into a hierarchy and solve it by the AHP steps. In a typical hierarchy, the top level reflects the overall goals of the decision maker. The elements affecting the decision are called objective, and they are represented at the intermediate levels. Objective can be further divided into criteria for additional refinement. AHP can be used to make relative measurements through paired comparisons of criteria and of alternatives, or to make ratings measurements of the alternatives with respect to the criteria.
According to the literature (e.g., Ahn, Ryu, and Han, 2004;Bremser and Chung, 2005; Davis and Mentzr, 2006; Khouja, 2001;Parasuraman, Zeithaml, and Berry, 1985; Popken, 2006;Tsaur, Chang and Yen, 2002) as well as practical considerations, the evaluation criteria we established include four dimensions of objectives (barrier dimension of the RD system, operation procedure, delivery service, and information system) and sixteen service quality evaluation criteria (system complexity, extra investment, delivery cost, fixed cost, user-friendly e-map, accurate map information, a limiting factor of the size of goods, convenient operating procedure, service quality of clerk, reliable delivery, return delivery, follow-up service when in damaged delivery, timely information, accurate logistics information, notice by cell phone to picks-up goods and e-tracking mechanism), Table 3 includes details. Once the hierarchy has been constructed, the decision maker begins the details of pair-wise comparisons to estimates on each level. The next step is the integration of these elements, using weights for an overall prioritization of decision alternatives.
______
Table3 about here.
______
5. Empirical study and discussions
5.1Survey
To investigate the issues presented above, an empirical study was undertaken for Taiwan e-retailing. Data collected via a mail-administered questionnaire. The questionnaire for service quality evaluation had two parts: questions to evaluate the relative importance of criteria and questions to evaluate retailing delivery providers’ performance corresponding to each criterion. AHP was used to obtain the relative weights of criteria. The two major retailing delivery providers examined in this empirical research were CVS.com and 7-11.com. In an effort to improve response rate and reduce non-response bias, several suggestions from the literature were adopted (Dillman, 2000). This included the enclosure of a stamped addressed envelope with the questionnaire, assurances of confidentiality and anonymity and a promise that a report of the results and managerial implications would be sent to the respondents after completion of the study.
A total of 42 questionnaires were sent out to e-retailing firms which use the retailing delivery service for their on-line customers. Two follow-up reminders with enclosed questionnaires were sent 2 and 4 weeks after the initial mailing. Of the 42 surveys, 32 were returned, for a return rate of 76%. The statistics for these firms were: 25% were large e-retailers (over than 5000 retail delivery orders per month), 34% were medium e-retailers (between 500 and 5000 retail delivery orders per month), 41% were small e-retailer (less than 500 retail delivery orders per month); 31% of the firms surveyed used logistics by 3PL; 41% were in the cosmetics industry and 16% were booksellers.
5.2Evaluation of criteriaweights
As we mentioned above, it can not be assumed that each evaluation criterion is of equal importance. The AHP was selected for this study for its suitability in evaluating multiple-criteria decision-making problems. Here, we employed the AHP method to derive the weights of the evaluation criteria. Fig.4 shows the relative weights of the four objectives of logistics service quality, which are obtained by applying AHP. Since the consistency ratio is below 0.1, the judgements in Fig.4 are considered consistent. The weights for each of the objective are: barrier dimensions of the RD system (0.1207), operation procedures (0.2985), delivery service (0.286) and information system (0.2948).
______
Figure4 about here.
______
5.3Performance measure of service quality and hypotheses testing
The performance measure of each respondent is calculated on five-point Likert Scale to obtain the overall performance measure of each retailing delivery provider. After obtaining the performance measure, we use TOPSIS as the main device in ranking the service quality of the 7-11.com and CVS.com. By sorting the values of Ci*, the ranking order of all alternatives is obtained.
In order to explore the difference in level of the satisfaction for each evaluation criteria for different RD providers in e-retailing at different scales, the ANOVA test is used to evaluate the following hypotheses:
H0i: there is no difference of satisfaction of evaluation criteria i for RD provider j in the e-retailing of different scales k
i=1, 2…16
j=1 (7-11.com), 2 (CVS.com)
k=1 (large scale e-retailing), 2 (medium scale e-retailing), 3 (small scale e-retailing).
The weights of all criteria by AHP, performance measures and ranking by TOPSIS and the ANOVA test, shown in Table 4. Ranked by the weights of all criteria as shown in Table 4, the top five evaluation criteria are: accurate map information (0.1317), accurate logistics information (0.1005), reliable delivery (0.10003), follow-up service when in damaged delivery (0.0963), convenient operating procedure (0.0805). Next discuss the view of all criteria of e-retailing of different scale as the following: as to large e-retailing, the top five evaluation criteria are: delivery cost (0.395), accurate map information (0.349), accurate logistics information (0.336), convenient operating procedure (0.317) and follow-up service when in damaged delivery (0.308); for medium e-retailing, the top five evaluation criteria are: accurate map information (0.438), accurate logistics (0.397), delivery cost (0.386), reliable delivery (0.324) and follow-up service when in damaged delivery (0.264); and form small e-retailing, the top five evaluation criteria are: accurate map information (0.325), follow-up service when in damaged delivery (0.291), notice by cell phone to picks-up goods (0.290), reliable delivery (0.289) and delivery cost (0.264).
______
Table 4 about here.
______
Clearly, the greatest e-retailing concern about is the e-map mechanism; this is probably because of the on-line shoppers are the real customers for e-retailing. On the other hands, e-retailing is less concerned about accurate information on the logistics information system.
According to Table 4, 7-11.com has the following eight criteria that are better than CVS.com: user friendliness of the e-map mechanism, accurate information on the e-map, service quality of clerk, reliable delivery, return delivery, timely information, accurate information, and e-tracking mechanism.In general, 7-11.com performs better in operational procedures (0.6978), delivery service (0.6630), and information system (0.6500), whereas CVS.com has better barrier dimensions of the RD system (0.6771). The ANOVA test indicates that different scales of e-retailing show significant differences in three of the evaluation criteria: delivery cost, accurate information and notification by cell phone to pick-up merchandise. The results show that large scale of e-retailers are more concerned with the criterion of “delivery cost” than other scale of e-retailers due to a large number of orders; as for medium scale of e-retailers are more concerned with the criterion of “accurate delivery information” than other scale of e-retailers because of the enough orders but poor information ability; and small scale e-retailers attach importance to the criterion “notice by cell phone to picks-up goods” due to the cost factor.
Table 1.Pair-wise comparison scale for AHP preferences (Saaty, 1980)
Numerical rating
/ Verbal judgments of preferences1
/ i is equally important toj3
/ i is slightly more important than j5
/ i is strongly more important than j7
/ i is very strongly more important than j9
/ i is extremely more important than j2, 4, 6, 8
/ Intermediate valuesTable 2.Average random index for corresponding matrix size (Saaty, 1980)
Matrix size (n)
/ 1 / 2 / 3 / 4 / 5 / 6 / 7 / 8 / 9 / 10Random index (RI)
/ 0 / 0 / 0.58 / 0.90 / 1.12 / 1.24 / 1.32 / 1.41 / 1.45 / 1.49Table 3.Definitions of evaluation criteria of retailing delivery
Evaluation criteria
/ DescriptionBarrier dimensions of the RD system
System complexity (C11)
/ The RD provider provides a user-friendly interface for the retailing delivery system on a web siteExtra investment (C12)
/ The e-retailing does not need extra investment when using retailing delivery systemDelivery cost (C13)
/ The delivery cost of e-retailing when provides retailing delivery serviceFixed cost (C14)