اکتشاف مدل رضایت مشتری از دیدگاه جامع

This study provides a model of customer satisfaction from a comprehensive perspective and tries to use the nonlinear fuzzy neutral network model to verify the assumptions of the study. Samples are taken from the information and tourism industries at a proportion of 2:1 based on the population in Taipei and Kaohsiung cities. A total of 207 questionnaires are returned. As the result of the empirical research shows, the interpersonal-based service encounter is better than the technology-based service encounter in functional quality, while the technology-based service encounter is better than the interpersonal-based service encounter in technical quality. The functional quality has a positive and significant effect on customer satisfaction; the service quality has a positive and significant effect on service value; the service value has a positive and significant effect on customer satisfaction. The service encounter has a positive and significant effect on relationship involvement and the relationship involvement has a positive and significant effect on customer satisfaction.

The industrial structure in Taiwan has transformed from traditional agricultural and manufacturing industries to the tertiary service industry that is booming and dominating the market with its high employment rate and production value. In the circumstances, it is very important for the service industry to improve their service quality. The core value provided by the service industry to consumers includes not only the uniqueness of tangible and intangible products, but also various factors involved in the process of service delivery to customers, such as physical facilities, company image, and quality of the service delivery. Emphasizing customers’ perception of service quality is important to enhance the extended perception and impression of consumers on products, and to increase the added value of products. With the rise of consumer awareness, requirements of customers changing from mass production to customization and quality. Therefore, how to understand the requirements of customers and provide them with products and services they need, reduce customer costs and improve customer value, and continually trace the satisfaction of consumers are essential for an enterprise to march toward success. Much research on the cause of service quality and customer satisfaction has been published in recent years. This research focuses on the following dimensions: (1) discussing the antecedent variable that affects customer satisfaction. Zeithaml and Bitner (2000), for example, found that customer satisfaction was affected by service quality, product quality, price, personal and situational factors. Oliver and DeSarbo (1988) found that service quality was the antecedent variable of customer satisfaction, while Anderson, Fornell, and Lehmann (1994) found that customer satisfaction was dependent on service value; (2) presenting theoretical models to explain the importance and necessity of customer satisfaction. Anderson and Suillivan (1993) pointed out that the satisfaction or dissatisfaction of customers could be measured based on the gap between the expectations of customers before purchasing and the performance of the purchased products or services. Woodruff, Cadotte, and Jenkins (1983) raised the norms in models of consumer satisfaction. They found that consumers evaluated products of different brands based on “norms” and their satisfaction was determined by the conformity among these norms. Meuter, Ostrom, Roundtree, and Bitner (2000) introduced the Technology Infusion Matrix to analyze how to enhance the provision of customized service by helping staffs or consumers use hi-tech tools and ensure the service failure compensation to improve customers satisfaction; (3) bringing up research dimensions and approaches for measuring customer satisfaction and service quality. For example, Perkins (1993) pointed out that industries were different in target market segmentation and, thus, measurement of customer satisfaction varied depending on the characteristics of individual industries. He raised three dimensions to be taken into consideration for empirical research on measurement of customer satisfaction: operation dimension (availability, delivery on time, price, credibility), service dimension (sales service, technical support, product line), product dimension (technical value, reliability, design). Parasuraman, Zeithaml, and Berry (1985) referred to “encounter” and “feasibility” as key dimensions. After interviewing senior management and consumers of four service industries and conducting comprehensive exploratory research, they raised the five-gap theory and brought up an important measurement approach. With this approach, excess of expected service (ES) over perceived service (PS) was represented by ES > PS, indicating that the overall service quality was not acceptable, while satisfactory service quality was represented by ES = PS and excess of perceived service over expected service was represented by ES < PS, indicating an inclination toward ideal quality. Previous research on customer satisfaction had the following characteristics: (1) emphasizing the discussion of performance. Oliver (1980), for example, found that satisfaction was a value consumers give to a deal. The higher the service level that consumers expect to have, the more difficult for the service provider to provide, the higher the customer satisfaction. According to the conformity theory of Mittal and Tsiros (1999), when the expectation of consumers was not in conformity with the actual performance of a product, they might change their perception of the product and incline to eliminate the gap between the expectation and performance, and thus improve the satisfaction; (2) emphasizing the discussion of cause and result of satisfaction. Cronin and Taylor (1992), for example, evaluated service quality based on performance and argued that service quality was the cause of customer satisfaction. Parasuraman, Zeithaml, and Berry (1988) researched the gap between service quality and customer satisfaction; (3) explaining the cause of customer satisfaction from a single perspective, such as the attribution theory of Bitner (1990) and the equity theory of Oliver and DeSarbo (1988). Kotler (2000) found that customer satisfaction was closely linked to customer value and the measurement of customer satisfaction should be carried out from the perspective of value. As some previous literature shows, the topics such as “service quality”, “service value”, and “customer satisfaction” were researched separately and only a limited number of literature focused on integration research in the relationship between the behavior intention of consumers and the service quality, service value, and customer satisfaction as a whole (Cronin, Brady, Tomas, & Hult, 2000). This study provides integrated research of these variables and focuses on their relationship as the first research motive; (4) most of the satisfaction models used in previous research emphasized the perceptive (Kotler, 2000) and emotional perspectives (Price, Arnould, & Tierney, 1995) of consumers. However, what a provider of the service industry emphasizes is not only the interaction with consumers to ensure their satisfaction, but also the method of changing the trade exchange to the value increment exchange (the relationship continuum theory of Day, 2000). Therefore, this study proceeds to the research from the perspective of relationship involvement as the second research motive. Besides, the multivariate statistic analysis approach or case study was commonly used for empirical research (Bolton and Drew, 1991, Cronin et al., 1997 and Zeithaml, 1988) and the nonlinear research model (such as the nonlinear fuzzy neural network model) was not used often for theory verification. The nonlinear fuzzy neural network model is maturely developed and has a wide range of applications. It is not only useful for prediction and assortment, but also for uncertain behavior systems. Therefore, this study uses the nonlinear fuzzy neural network model to verify the initial data of the research as the third research motive. Based on the above, the purpose of this study is listed as follows: (1) To provide a customer satisfaction model from a comprehensive perspective of relationship involvement, service encounter, service quality, and service value. (2) To verify collected data with the nonlinear fuzzy neural network model and discuss its meaning for management.

We start with the discussion of the relationship between the service encounter and service quality based on Assumption 1. Each input variable has two membership functions Low and High. The zero-order Sugeno fuzzy model is used for the rule base. The average testing error after 105 learning epochs is 0.2459. The test result is shown in Fig. 3 (• stands for the testing data of the input variable, while ♦ stands for the data inferred from the fuzzy neural network model). The relationship between the “service encounter” with the “service quality” is significant as shown in the figure and, thus, Assumption 1 is supported. Full-size image (9 K) Fig. 3. Input variables and membership functions of the “service encounter” vs. “service quality”. Figure options It is found in the verification of Assumptions 1-1 and 1-2, which discuss the effect of different types of service encounters on service quality, that the “interpersonal-based” service encounter and the “functional quality” has a positive relationship as shown in the distribution of the empirical and inferential values in the function of the “interpersonal-based service encounter” and “functional quality” (Fig. 4(a)). Therefore, Assumption 1-1 is supported. In addition, it is found that no steady relationship exists in the distribution of the empirical and inferential values as indicated in the function of the “technology-based service encounter” and “technical quality” (Fig. 5(b)). Therefore, Assumption 1-2 is not supported. This indicates that enterprises should enhance externally delivered service quality to improve their competitiveness in a fiercer competition environment to meet the requirements of demanding consumers who increasingly emphasize the functional value of products. Full-size image (23 K) Fig. 4. Effect of the “interpersonal-based service encounter” on “service quality”: (a) input variables and membership functions of the “interpersonal-based service encounter” vs. “functional quality”; (b) input variables and membership functions of the “interpersonal-based service encounter” vs. “technological quality”. Figure options Full-size image (23 K) Fig. 5. Effect of the “technology-based service encounter” on “service quality”: (a) input variables and membership functions of the “technology-based service encounter” vs. “functional quality”; (b) input variables and membership functions of the “technology-based service encounter” vs. “technical quality”. Figure options Membership functions and rule bases are also used to verify Assumption 2 for the effect of the “service quality” on “customer satisfaction”. The average testing error after 162 learning epochs is 0.2517. The test result is shown in Fig. 6. The figure shows that the “service quality” has a consistent and significant effect on the “customer satisfaction” and, thus, Assumption 2 is supported. It is found in the verification of Assumptions 2-1 and 2-2 that the ’technical quality” has no significant effect on the “customer satisfaction” (Fig. 7(a)) and, thus Assumption 2-1 is not supported. However, the “functional quality” has a positive and significant effect on the “customer satisfaction”. Therefore, Assumption 2-2 is supported. Full-size image (9 K) Fig. 6. Effect of the “service quality” on “customer satisfaction”. Figure options Full-size image (23 K) Fig. 7. Effect of different service quality types on customer satisfaction: (a) effect of the “technical quality” on “customer satisfaction”; (b) effect of the “functional quality” on “customer satisfaction”. Figure options Assumptions 3 and 4 are verified by means of membership functions and rule bases. The average test errors after about 150 and 142 learning epochs are 0.2964 and 0.2371, respectively. The result of the test is shown in Fig. 8 and Fig. 9. As the figures show, the “service quality” has a positive and significant effect on the “service value” and the “service value” also has a positive and significant effect on the “customer satisfaction”. Therefore, both Assumptions 3 and 4 are supported. Full-size image (9 K) Fig. 8. Effect of the “service quality” on “service value”. Figure options Full-size image (10 K) Fig. 9. Effect of the “service value” on “customers satisfaction”. Figure options Assumptions 5 and 6 are verified also by means of membership functions and rule bases. The average test errors after about 162 and 155 learning epochs are 0.2436 and 0.2508, respectively. The result of the test is shown in Fig. 10 and Fig. 11. As the figures show, the “service encounter” has a positive and significant effect on the “relationship involvement” and the “relationship involvement” also has a positive and significant effect on the “customer satisfaction”. Therefore, both Assumptions 5 and 6 are supported. Full-size image (10 K) Fig. 10. Effect of the “service encounter” on “relationship involvement”. Figure options Full-size image (10 K) Fig. 11. Effect of the “relationship involvement” on “customers satisfaction”. Figure options The empirical research of this study has the following meanings: (1) the service industry provides intangible products with low-end technology and technical quality is less important than functional quality for consumers of the service industry. This conclusion matches with the emphasis on functional quality in the study of Sweeney et al. (1997); (2) previous research mostly discussed the relationship among the perceived value, satisfaction and behavior intention of consumers (Brady et al., 2001 and Cronin et al., 2000). However, the perspective of this study is closer to the perspective of Zeithaml, 1988 and Parasuraman and Grewal, 2000 and has the same argument that the service quality is the antecedent variable of the perceived value of consumers; (3) like previous research, the empirical research of this study proves the positive effect of the service quality on customer satisfaction. However, the functional quality is more important than the technical function for consumers of the service industry; (4) most previous research focused on the linkage of the service quality and customer satisfaction. This study emphasizes that enterprises of the service industry should make efforts to establish and maintain customer relationship. Besides, most previous research focusing on relationship involvement discussed the interaction among enterprises, such as the supply of material and cooperation with supplier or satellite factories. This study applies it to the buy-and-sale relationship between individuals involved in the service industry; (5) service providers can enhance their relationship with customers by various means, such as promotion or enhancement of the characteristics for peripheral services. Professional capability and knowledge are also important for establishing and maintaining customer relationship. This study uses the fuzzy neural network model to verify assumptions and numerous questionnaires are distributed to collect original data for the verification. Unlike previous research that discusses the topic based on a single perspective or theory, this study uses a multi-dimension approach to explain the factors that affect the customer satisfaction. This study is based on cross-section data. Follow-up research may acquire different results based on long-term longitudinal section observation. Researchers may also add other environmental or cultural variables in their research, for example, the attitude of customers or staffs toward perceived risks and service quality gaps in different environments. Researchers may verify the result using different approaches; for example, comparing the data verification results of the multi-variable statistic analysis and fuzzy neural network model and explaining the reason. This research may have more theoretical contributions to this topic in doing so. The limitations on this study are as follows: (1) 207 samples are collected for this study. However, since there are many variables on the questionnaire and the participants are busy for their business and have not much time or willingness to complete academic questionnaires, the number of returned questionnaires is not satisfactory. In the application of the fuzzy neural network model, at least 100 questionnaires should be returned to reduce errors. Though the returned sample of this study has reached this standard, learning epochs are used to avoid the problem of high test errors; (2) different operational approaches can be used for follow-up research. The willingness of the participants to complete the questionnaire should be enhanced to increase the number of returned samples and improve the precision of the analysis.