**2. Literature Review**

## *2.1. Brand Attachment (BA)*

The definition of brand attachment is the emotional bond between consumers and brands [14,15], and the concept of brand attachment refers to the strength of the bond connecting the brand with the self, which is very powerful in leading to satisfied, trusting, and committed relationships. The strength of the attachment varies on the basis of differences in the link between consumers and brands, and brand attachment has a significant impact on consumer behavior [16].

Carroll and Ahuvia [17] considered brand attachment as a way for consumers to show their obsession with and loyalty to products, which is a kind of profound, heartfelt emotional attachment behavior. Mishra et al. [18] defined brand attachment as the strength of the bond linking brands with the self. Park et al. [19] believed that brand attachment reflects the emotional bond between people and brands. Moreover, it enriches consumers' memory network and the perception of the brand-self relationship. Bian and Haque [16] maintained that if consumers have a stronger brand attachment, they are more willing to engage in challenging activities that require time, money, energy, and reputation. Combining all these research results, we conclude that the external brand attachment behavior shown by consumers is actually an expression of satisfaction of their internal demands.

Hung and Lu [20] regarded brand attachment as an integration of the structure of emotions, which can be divided into positive and negative emotions. The results of their research demonstrate that the positive emotion of brand attachment is an effective predictive factor of brand repurchase intentions and word-of-mouth behavior. The dimensions of positive emotions are affection, passion, the brand–self connection, and brand prominence. The four dimensions of positive emotion proposed by Hung and Lu [20] are based on the pioneering empirical research on affection and brand attachment conducted by Thomson et al. [21], who elaborated that the bond of brand affection is characterized by deep connection, affection, and consumer passion. In addition, Park et al. [19] regarded the brand–self connection and brand prominence as two key factors of the brand attachment concept. On the basis of the research results above, this study adopts four key factors as brand attachment dimensions: affection, passion, the brand–self connection, and brand prominence.

Regarding the attributes of each dimension, the affection dimension is based on the study of Thomson et al. [21], who evaluated consumers' brand affection using a scale. This factor reflects consumers' warmth toward brands, which can be affectionate, loving, friendly, and peaceful. Among them, the peaceful attribute cannot be controlled by either consumers or businesses; therefore, it is excluded from this study. The passion dimension mainly reflects consumers' ardor, excitement, and positive emotions, which can include passion, delight, and captivation. The brand–self connection dimension is mainly based on the brand function evaluation dimension proposed by Rio et al. [22]. It includes the brand guarantee, self-identity, social identity, and status. Among these, the status function corresponds to the personal desire to gain recognition from others. However, it does not represent the groups recognized by brands; consequently, it is excluded from this study. The brand prominence dimension is based on the brand awareness concept proposed by Aaker [23]. Aaker conceptualized brands and divided brand awareness into six attributes on the basis of brand equity: brand cognition, brand memory, brand opinion, brand dominance, top-of-mind brand awareness, and brand knowledge. Brand cognition, brand memory, and brand knowledge were chosen as the evaluative attributes of brand prominence in the present study to match the questions regarding brand prominence. The content above is organized in Table 1.


**Table 1.** The dimensions and attributes of brand attachment.


**Table 1.** *Cont.*

Ko et al. [25] pointed out that once a business has begun to add value, the next step is trying to increase its competitive advantages. The best way to achieve this goal is to build consumers' emotional attachment to the business and its products. One way to build attachment is through experiential marketing. The purpose of experiential marketing is to fully understand consumers and think from the standpoint of the consumers to gain feedback from genuine experience, which can be converted into a precious resource to make brands grow.

#### *2.2. Strategic Management Plan for Brand Experience*

Brand experience is the sum of the product, shopping, service, and consumption experience gained by consumers when interacting and cooperating with brands with a clear intention to consume products and services [1,2,26]. Regarding consumers, compared with real products and intangible services, experience is the most unforgettable [27]. Bleier et al. [28] considered that there are two kinds of experiences: direct experiences, such as shopping, purchasing, and consuming products, and indirect experiences, such as consumers receiving commercials and marketing communication. Brakus et al. [2] believed that brand experience is formed by the brand-related stimuli (names, slogans, and pictures) received by consumers when searching for, purchasing, and consuming brands. These brand-related stimuli form the main source of consumers' subjective reactions. The foundation of brand experience lies within the brand itself. Many studies have mentioned that brand experience constitutes the best chance for consumers to have a sense of passion for and differentiate between brands and that the interaction between experience and brands is very close [29].

The meaning of a brand involves creating explicit consumer reactions; therefore, deeper relationships with consumers can be developed [27]. Experience can create brand image and offer opportunities for consumers' curiosity to be aroused and for consumers to explore, purchase, and become involved, thus creating a sense of favorability [30]. Jimenez-Barreto et al. [31] also proposed that brand experience can effectively arouse consumers' interest in brands and products and further arouse their desire to purchase. Mclean et al. [32] also conducted a related study on brand experience activities and developed a mobile application customer experience model in which consumers keep circulating and creating valuable things with brands. Molinillo et al. [33] observed the effect of brand experience on loyalty to retailers. They believed that the purpose of brand experience activities is not only to deliver brand messages but also to listen to the real voice of consumers through the activities. Only if a brand finds the core of what consumers care the most about through conversation can brand activities create more substantial benefits.

Schmitt [34] developed strategic experiential modules (SEMs), which divided experience into five forms: sensing, feeling, thinking, acting, and relating. With regard to sensing, feeling, and thinking, consumers' expectations of brands go far beyond the brand-related function and practicality. Maehle et al. [35] tried to satisfy consumer demands, considered that experience can achieve full customer satisfaction, and divided experience into five dimensions: definition, notification, imagination, immersion, and interest arousal. At the same time, Schmitt [34] pointed out the difference between products and experience and confirmed that the concept of brand experience genuinely depicts the spirit of a brand. He further explained that consumers' expectations of brands, which integrate consumers' feelings, sensations, and wisdom, go far beyond the brand-related function and practicality.

Many scholars continue to investigate experience issues, and they all propose their own views. On the basis of past research on consumers and marketing, Lam et al. [36] offered another perspective: when consumers search for products, go shopping, receive services, and consume products, experience automatically happens at the same time as such behaviors. They divided experience into sensing, feeling, cognition, and behavior. Nysveen et al. [37] believed that in addition to the four dimensions of sensing, feeling, cognition, and behavior, relating is essential and constitutes a fifth dimension. The measurement base of overall brand experience in this study adopted the brand experience factor dimensions proposed by Schmitt [9], which are recognized by major scholars and include five dimensions: sensing, feeling, thinking, acting, and relating. The five dimensions of brand experience proposed by Lam et al. [36] and Nysveen et al. [37] were also adopted, and the mobile application brand experience research conducted by Kim and Yu [38] was also included. Excluding the sensing dimension, which has little effect, this study adopted four dimensions as the structure of brand experience and strategic managemen<sup>t</sup> planning, i.e., feeling, thinking, acting, and relating, as shown in Table 2.


**Table 2.** Strategic managemen<sup>t</sup> plan of brand experience.

#### *2.3. Fuzzy Quality Function Deployment*

To truly understand consumers' reactions and voices with regard to brand attachment, the QFD procedure makes it possible to understand the brand attachment felt by consumers [46] and, in turn, provide a service experience. The QFD concept originated in Japan in the 1960s and had a widespread impact on Western countries in the 1980s. It is particularly important for the business industry when trying to understand consumers' versatile demands [47]. Through the planning and design of products or services and the execution of managemen<sup>t</sup> procedures, QFD can help business owners understand differences in the cognition of consumers, especially when business owners are providing new services and products [48]. This multistep procedure can be expressed by an internal relational structure matrix, and the easiest method is to use the HOQ framework [48]. The relationship between WHATs and HOWs and the weight and performance values of every element can be presented in an HOQ matrix. Figure 1 shows the six parts of the HOQ matrix: (1) customer requirements (WHATs), (2) technical requirements (HOWs), (3) the relationship matrix (correlation of WHATs and HOWs), (4) the customer needs assessment, (5) competence analysis, and (6) prioritized technical requirements. However, when the demands of consumers become complex and versatile, the message provided by the HOQ will subsequently become fuzzy. For example, the key demand attribute may be overestimated, or it may not be possible to precisely show its importance. Therefore, through the explanation of fuzzy sense, the process design can be more flexible.

**Figure 1.** The framework of HOQ.

Thus, this study attempts to use the QFD procedure to analyze the brand attachment attributes felt by consumers. Subsequently, it makes the analyzed data the basis for designing a brand experience strategy. Applying QFD involves collecting semantic data, which include the subjective judgment or emotional uncertainty of consumers or experts. Therefore, regarding the issue of brand attachment, consumers will be inconsistent in the judgment standard of favorability because of different cognition of their internal demands. If an evaluation is performed using precise numerical values, it is possible that the study team will be unable to know consumers' reactions and genuine feelings; thus, the evaluation may also be unreliable. In fact, using semantic data to show emotional characteristics, such as liking or attachment, is fundamentally fuzzy and uncertain. However, when the values and demands of consumers become complex and versatile, the message provided by the HOQ will subsequently become fuzzy [49]. To process these data properly, this study adopts fuzzy set theory to cope with the problem posed by the fact that traditional statistical evaluation cannot precisely measure the emotions of attachment, attitudes, and ideas.

Regarding FQFD research, Hsu and Tang [49] believed that expressing consumers' demands through natural semantics involves fuzziness and uncertainty. Therefore, using the concept of fuzzy logic can overcome some limitations of the demand side of the HOQ and make consumers' imprecise demands easier to understand. Yuen [50] found some limitations in the traditional QFD method: (1) the traditional QFD method cannot provide a reasonable explanation of how a consensus is reached in the process of member evaluation; (2) the presentation of traditional QFD information cannot explain the subject's cognitive conflict; and (3) the data used in traditional QFD are assumed to be clear values, which ignores the fuzziness and differences in subjects' attitudes and behaviors. Therefore, Yuen [50] maintained that integrating fuzzy set theory into the HOQ will make complex decision-making data easier to organize and represent in a more flexible way. In addition, Haber et al. [51] used both fuzzy values and clear values to make comparisons and see what uncertainty would cause in the demand attribute of the HoQ. Their results showed that using clear values enhanced the effect caused by uncertainty.

In recent years, studies that have combined fuzzy set theory and QFD include the following. Mehtap and Karsak [52] combined QFD and fuzzy multicriteria decision making (MCDM) to select suppliers. Hsu and Tang [53] combined QFD and the fuzzy analytic hierarchy process (FAHP) and considered the combination of customer relationship interest and the business relationship linking strategy. Ultimately, they developed a model of customer relationship managemen<sup>t</sup> strategy. Lam and Lai [54] used the fuzzy analytic network process (FANP) and QFD to develop and design new products. Dat et al. [55] combined

FQFD with the technique for order preference by similarity to ideal solution (TOPSIS) to maximize company interest and compare market segmentation, adding characteristics of different markets to companies' existing advantages. Lu et al. [56] combined the FAHP, QFD, and the brand knowledge concept proposed by Keller to develop a model of brand revitalization. Kayapınar and Erginel [57] combined a fuzzy multiobjective decision-making model and FQFD to design airport services. Xie et al. [58] proposed a multiobjective decision-making method to choose the technical index in the HOQ; the purpose of this method was to calculate and fix the fundamental importance of the technical index. They also used mobile phone product design, for example, to verify the utility and effectiveness of the method. Zaitsev and Dror [59] applied QFD to examine the relationship between multiple indexes of corporate social responsibility and their results. Lizarelli et al. [60] used fuzzy approaches and QFD to support improvement decisions in an entrepreneurial education service. Chen et al. [61] proposed a new integrated MCDM method for improving QFD, which integrated the hesitant fuzzy linguistic term set under an uncertainty environment. Haiyun et al. [62] proposed the innovation strategies in the energy industry using the QFD-based hybrid interval valued intuitionistic fuzzy decision approach combined with a novel hybrid methodology.

Analyzing the studies above, we learn that the messages that consumers send often contain ambiguity and multiplicity in the sensing dimension. Moreover, the data or attributes oriented toward qualitative data often show subjective uncertainty in evaluations. Hence, it seems that the analytical framework of traditional QFD cannot truly represent the importance of the evaluation of consumer demand attributes. Therefore, applying the method of fuzzy set theory to analyze the semantic data expressed by the subjects and using fuzzy numbers to represent consumers' feelings toward their favorite targets will make the analysis more objective and evidence-based.

#### **3. Brand Attachment FQFD Analysis Process**

The objective of developing the FQFD framework in this study is to solve the abovementioned problems encountered by traditional QFD [50]. One of the main advantages of using consistency fuzzy preference relation analysis [63] in this study is that rather than the traditional method of comparing n(n − 1)/2 attributes multiple times, only (n − 1) evaluations are required. This can reduce the number of importance comparisons between attributes. Another advantage is that the FQFD analysis process allows the multiple semantic data presented by the subjects to be properly measured and avoids the subjective judgment and uncertainty of consumers that may be implied by the semantic communication in the evaluation process of traditional HOQ [55]. The strategic process of strengthening brand experience was constructed by FQFD and based on brand attachment, as shown in Figure 2. Each step is explained below.

**Figure 2.** Experience strengthens the strategic importance evaluation process of brand attachment demands.

#### *3.1. Step 1: Building Brand Attachment Demand Attributes (WHATs) and Experience Strategic Management Plans (HOWs)*

In the FQFD analysis process, the brand attachment demand attributes were WHATs, and HOWs were experience strategic managemen<sup>t</sup> plans that could satisfy or make consumers form brand attachment. We used both a literature review and expert interviews to collect and confirm the content of each brand attachment attribute. *BRi* represents the brand attachment demand dimension/attribute, and *BEj* represents the experience strategic managemen<sup>t</sup> plan.

#### *3.2. Step 2: Analysis Process Platform of Brand Attachment*

In the FQFD analysis platform, this study used MATLAB software with Excel to analyze the data, and the calculation process applied was as follows: (1) the calculation of the relative importance value of WHATs ( *Wi*), (2) the calculation of the value of the relationship between WHATs and HOWs ( *Rij*), and (3) the calculation of the final relative importance value of HOWs ( *Rj*). This study did not include the HOWs value of *cij* in the calculation, and therefore, the final integrated calculation focused on the value of *Rj*.

#### *3.3. Step 3: Assessing the Consistency Fuzzy Preference Relations of Brand Attachment*

This study used the concept of fuzzy linguistic variables and consistency fuzzy preference relations, in addition to the nine linguistic scales proposed by Herrera-Viedma et al. [64], to build a consistent fuzzy preference relations matrix. On a practical level, consistent fuzzy preference relations can effectively rectify the inconsistency in the evaluation outcome as a result of increasing the number of constructs examined. This method can effectively solve the inconsistency problem that may occur when interviewed experts complete a questionnaire, and in turn, it increases the overall effectiveness and accuracy. This study used the characteristic of the fuzzy linguistic preference relation addition reciprocal to calculate and used defuzzification [65] to acquire the crisp relative weight value of the brand attachment attributes ( *w*1, *w*2,... , *wn*).

#### *3.4. Step 4: Developing the Technical Items of Brand Experience Strategic Management Plans*

This study used the brand experience dimensions proposed by Schmitt [9] as the basis of developing a whole brand experience strategy, including sensing, feeling, thinking, acting, and relating. This study also combined the five brand experience dimensions proposed by Nysveen et al. [37] and the mobile application brand attachment items developed by Kim and Yu [38] to build the four dimensions and strategic managemen<sup>t</sup> plan of this study. We also assumed that *T*1, *T*2, ... , *Tj* are the technical items of the brand experience strategic managemen<sup>t</sup> plan.

#### *3.5. Step 5: Evaluating the Incidence Matrix of Brand Attachment Attributes and Brand Experience Strategic Management Plans*

This study assumed that the number of subjects is *k*, they evaluate the *i*th brand attachment attribute and the *j*th brand experience strategic managemen<sup>t</sup> plan, and the fuzzy incidence is *Rijk* = (*Lijk*·*Mijk*·*Uijk*). A triangular fuzzy number is demoted simply as (*Lijk*·*Mijk*·*Uijk*). The parameters *Lijk*, *Mijk*, and *Uijk* indicate the smallest possible value, the most promising value, and the largest possible value, respectively. Therefore, the average formula *Rijk* below can be used to organize the views from *K* subjects and to convert fuzzy linguistics into crisp numbers through defuzzification. Finally, we can obtain the crisp incidence matrix value, *Rijk*.

$$\widetilde{R}\_{ijk} = \left(\frac{\sum\_{k=1}^{m} L\_{ijm}}{m}, \frac{\sum\_{k=1}^{m} \mathcal{M}\_{ijm}}{m}, \frac{\sum\_{k=1}^{m} \mathcal{U}\_{ijm}}{m}\right), k = 1, 2, \; m$$

#### *3.6. Step 6: Evaluating the Importance of Brand Experience Strategic Management Plan Items*

The target value of brand experience strategic managemen<sup>t</sup> plan items was calculated from the incidence matrix value, *Rijk*, multiplied by the relative weight value of the brand attachment attributes ( *w*1, *w*2, ... , *wn*), and we acquired the relative weight value of the brand experience strategic managemen<sup>t</sup> plan *Rj*. To evaluate all brand experience strategic managemen<sup>t</sup> plans on the same basis, this study used a formula to standardize each value. Then, this study obtained the technical importance evaluation value of the target brand experience strategic managemen<sup>t</sup> plan.
