An interval type-2 fuzzy Kano-prospect-TOPSIS based QFD model: Application to Chinese e-commerce service design

Highlights

• A quality function deployment model is proposed under linguistic environment.

• Identify the attribute category of CRs using the interval type-2 fuzzy Kano model.

• Determine the weights of CRs based on the development stage of enterprises.

• Rank DRs by an extended TOPSIS method considering the behavioral factors.

Abstract

Owing to the rapidly changing customer preferences and demands, the e-commerce industry encounters various uncertainties and risks to enhance competitiveness. Quality function deployment (QFD) is a commonly used model, which can translate customer requirements (CRs) into products or service design requirements (DRs), to improve competitiveness by launching a new business. To measure the uncertainties and behavioral risk factors in e-commerce service design, we propose a new QFD model based on the Kano model and TOPSIS method by considering the behavior of experts with prospect theory under interval type-2 fuzzy linguistic environment. The categories of CRs are identified using the Kano model and the weights of CRs are determined dynamically according to the development stages of enterprises. The priorities of DRs are ranked using the extended TOPSIS method with prospect theory. A case study of China’s e-commerce service design is used to show the application of the proposed QFD model. The prioritizing results show the flexibility of the proposed model on determining the weights of CRs and the priorities of DRs for enterprises at different development stages.

Introduction

As a large population as well as a big demand country, China has become the biggest e-commerce country in the world [1]. In recent years, remarkable advances in technologies coupled with dynamically changing customer desires have urged e-commerce enterprises to reconsider their service strategies [2]. Quality function deployment (QFD) is a useful tool to transform customer requirements (CRs) into products or service desire requirements (DRs) [3], [4]. However, the critical parts of QFD, such as determining the weights of CRs and ranking the priorities of DRs, are facing the challenge of uncertainty caused by the subjective behavioral cognition of customers and stakeholders [5]. This paper aims to investigate Chinese e-commerce service design using an extended QFD model under an uncertain environment.

The importance of CRs is critical for the ranking of DRs in QFD. It is popular to compute the weights of CRs using multi-criteria decision-making (MCDM) methods based on experts’ evaluation [6], [7], [8]. Onar et al.  [6] proposed a new QFD model to determine the importance of CRs via the analytic hierarchy process (AHP) with hesitant fuzzy linguistic terms. Wu et al. [7] utilized the hesitant fuzzy decision-making trial and evaluation laboratory (DEMATEL) to analyze the interrelationships among CRs and determine their weights. Mistarihi et al. [8] integrated a QFD framework with a fuzzy analytic network process (FANP) to determine the degree of importance of the engineering characteristics. The above studies mainly used MCDM algorithms to improve the calculation process of expert ratings to determine the importance of CRs. However, the dependence of CRs’ importance on the decision-making environment, such as the classification of CRs and their impact on enterprises in different development stages, is ignored.

The classification of CRs is mainly related to user experience. Kano model is a user demand classification tool that can help us to understand and capture CRs from the perspective of customers [9]. Analyze the importance of CRs based on customers’ direct opinions using the Kano model can enhance customer satisfaction and loyalty, and minimize dissatisfaction [10], [11]. Due to the subjectivity of expert evaluation or the uncertainty of customer perception, the Kano model is also improved by fuzzy techniques [12]. Wang and Fong  [13] employed the fuzzy Kano model to capture customer perceptions of service attributes and convert them into quantitative degrees of customer satisfaction. Dou et al. [14] proposed an intelligent type-decreased mass-customized product configuration method with an improved fuzzy clustering algorithm that combines Kano’s model and social inertia. However, the existing fuzzy Kano models are deficient in processing human vagueness and certainties.

Ranking DRs is another critical process in QFD and it is also often solved using MCDM methods under fuzzy environments  [15], [16]. Among many MCDM methods, TOPSIS is a commonly used one which evaluates alternatives according to their proximity to an ideal target [17]. Bilgen [18] proposed an integrated fuzzy QFD and TOPSIS method for energy saving. In practice, the decision-making process is easily influenced by human behavior [19]. Prospect theory is a useful tool to analyze the behavioral factors of decision-makers [20]. Wang et al. [21] designed product concepts based on the combination of QFD and cumulative prospect theory. Liu et al. [4] developed a QFD model with extended prospect theory under a hesitant linguistic environment. However, the consideration of prospect theory in existing studies faced the risk of information loss, especially in fuzzy environments.

The uncertainty in the above process of CRs’ weights determination and DRs’ priorities mainly refers to the vagueness and subjectivity expressed by customers and stakeholders. Language is the main way in which human beings intuitively express emotions and feelings [22]. The interval type-2 fuzzy sets (IT2 FSs) is a possible representation to support the representation and calculation of linguistic terms, providing more freedom with more parameters than other fuzzy techniques to handle uncertainties [23]. The specific interval type-2 fuzzy linguistic variables can be constructed using the questionnaire survey method according to application context [24], [25]. IT2 FSs has been widely employed in the field of MCDM  [26], [27], [28] and logic control [29], [30], [31]. However, the scene applicability of fuzzy linguistic variables and the flexibility of IT2 FSs in dealing with uncertainties are rarely considered in current QFD studies.

According to the above analysis, the decision behavior and uncertainty of QFD have been studied partly. But there are still some deficiencies in current QFD models that need to be solved:

(1) In CRs weights calculation, the influence of attribute classification is rarely considered under a linguistic uncertain environment.

(2) The weights of CRs are almost considered equally important for ranking DRs without considering the behavioral factors of stakeholders when enterprises at different stages.

(3) The uncertain decision behavior of stakeholders is not sufficiently considered in current QFD models.

To overcome the above limitations, some assumptions are given as follows:

Assumption 1

The e-commerce enterprises do market research to collect CRs before launching new businesses.

Assumption 2

The e-commerce enterprises at different development stages have different strategies for implementing new businesses.

Assumption 3

Neither customers nor stakeholders are completely rational when making decisions.

According to these assumptions, we solve the above limitations by identifying the categories of CRs using the Kano model, determining the weights of CRs according to the development stages of enterprises, and ranking DRs considering experts’ uncertain behavior. The main contributions of this study are introduced as follows:

(1) Identify the attribute category of CRs using the Kano model. Kano questionnaires of CRs are designed with linguistic terms and sent out to target customers. The effective questionnaires are gathered and processed with interval type-2 fuzzy sets and CRs are classified using the interval type-2 fuzzy Kano model.

(2) Determine the weights of CRs according to the development stages of enterprises. The objective weights of CRs are firstly computed based on the statistical results of Kano questionnaires and the normalized weights are obtained considering stakeholders’ risk attitudes on the development stages of the target enterprise.

(3) Rank the priorities of DRs using an extended TOPSIS method considering the behavior of experts. The correlation between CRs and DRs is evaluated by domain experts with interval type-2 fuzzy linguistic terms. The priorities of DRs are ranked considering experts’ behavioral factors using the extended TOSIS method with prospect theory.

A case study of service design in Chinese e-commerce is given to analyze the effectiveness of the proposed QFD model. Five CRs and ten DRs are selected for the case study according to the user experience and complaint monitoring report of the China E-Commerce Research Center (CECRC) during 2013 and 2020. 320 Kano questionnaires are sent out to target customers and 267 effective respondents are recycled. According to the statistical results, two CRs belong to the Must-be attribute, two CRs belong to the Attractive attribute, and one CR belongs to the One-dimensional attribute. The normalized weights of CRs and priorities of DRs are analyzed for enterprises at different development stages. The comparison analysis shows that the proposed QFD model is available for enterprises at different development stages.

The remainder of this paper is organized as follows. The preliminaries of QFD, IT2 FSs, Kano model, and prospect theory are introduced in Section 2. A QFD model is proposed based on the Kano model and TOPSIS method considering experts’ behavior under a linguistic environment in Section 3. A case study in Chinese e-commerce is used to show the effectiveness of the proposed QFD model in Section 4. A conclusion is given in Section 5.