Building Brand-Fan Relationships in Social Commerce Contexts: Mediators of Online Brand Relationships

The effect of the Internet and social media on consumer behavior has recently begun to outweigh the effect of traditional marketing strategies. Therefore, numerous enterprises have begun to use social media to communicate with brand users, and social media has b"ecome the optimal tool for brand marketing. According to current brand-community relationships from the perspectives of brand authenticity and brand identification, we identified the mediating factors (brand trust and brand passion) of relationships between community members and brands and developed an integrated model to determine the effects of these factors on community member purchase intentions. The research targets in the present study were consumers who had purchased products from one of the top three cosmetic brands in Taiwan and had been a member of the brand’s official fan page for at least 1 year. Overall, 484 valid samples were collected in this study. We employed structural equation modeling to examine our model. Five hypotheses were supported, and excellent model fits to the data were obtained. Additionally, this study discovered that brand trust and brand passion played crucial mediating roles. Finally, we propose management implications and make suggestions that can serve as a reference for brand community managers.


Literature Review and Research Hypotheses
The literature section proposes research framework, the relevant theory and the different constructs adpoted in the study, and hypothesizing relationships among these variables.

Research Framework
In this study, brand authenticity and brand identification (as defined in social identity theory; SIT) served as the foundation for the development of a theoretical model. The main content of identity perspective is derived from selfcategorization theory [77] and SIT [75]. SIT is considered useful for understanding customer behavior because customers are not required to act openly with other customers (or even understand them) but consider themselves part of a social group [52]. Applying SIT to brand community topics can explain the key points of community members' community identification formation. In virtual community contexts, community member brand identity helps members identify with the brand community to which they belong [81]. In particular, when they believe that their selfimage closely resembles the appeal of the brand in question, consumers verify the appropriateness and necessity of their self-existence through brands to consider whether their attitudes, values, and behaviors are socially acceptable [29].
In addition, the mediation roles of brand trust and passion in brand relationships were explored to understand the brand purchase intentions of fans. Figure 1 illustrates the research framework.

Brand Authenticity
The word authenticity, which originated from the Latin word authenticus and the Greek word authentikos, is defined as the quality of being authoritative, valid, true, real, genuine, or trustworthy [16]. In the humanities and social sciences, the concept of authenticity has broad connotations and researchers have provided various explanations for it [15]. As Kim et al. [50] reported, if a corporate brand clearly expresses the characteristics of a product or provides an authentic experience to consumers, the consumers are aware of the brand's authenticity [20].
Studies have defined the concept and dimensions of brand authenticity differently. Bruhn et al. [15] investigated consumer understanding of brand authenticity and reported that brand authenticity has four dimensions: continuity, originality, reliability, and naturalness. Eggers et al. [27] proposed three operational factors of brand authenticity: brand consistency, customer orientation, and brand congruency. Coary [22] concluded that brand authenticity has three components: being a category pioneer, maintaining the original product formulation, and adhering to certain

Social Identity Theory and Brand Identification
Social identity involves clarifying self, group, and intergroup concepts and relationships [39]. Consumer brand identification can be understood through social identity theory and social psychology [78]. Personal brand identification originates in a consumer's affinity with a brand. When a consumer is a brand community member, their social brand identification is reinforced [54]. A rich brand image that matches and protects a customer's self-image leads to a favorable evaluation of the brand, which in turn enhances the brand's reputation [43]. Strong brand identification indicates that brand users believe themselves to belong to a specific brand [52]. According to some studies, brand identification means that brand users consider a brand as a reference point to distinguish the brand's users from its nonusers [30]. Kim et al. [48] concluded that a brand can reinforce consumer brand identification to some degree.
Brands have become crucial for establishing a relationship with consumers and ensuring successful business operation [77]. Consumer brand identification is crucial for brand management, increasingly receives attention from experts and researchers [12], and affects the relationship between consumer behavior and brands; therefore, consumer brand identification should be further understood. According to Escalas and Bettman [30], Kuenzel and Halliday [52], and Yeh and Choi [81], brand identification can be defined as consumers identifying with a certain brand and considering themselves part of a specific group.

Brand Trust and Passion within Online Brand Relationships
To address the relevant constrcuts definitions in online brand relationships, including brand trus and brand passion.

Brand Trust
The concept of trust has been explored in various fields (e.g., social psychology, sociology, and commerce and marketing). Brand trust is composed of cognitive beliefs [24] and emotional perceptions [28]. Cognitive beliefs include predictable expectations concerning the reliability, consistency, competitiveness, and performance of a product [24]. Emotional perceptions, however, include evaluations stemming from emotions, from which emotionrelated expectations, such as a brand's trustworthiness and benevolence, arise [11]. Therefore, consumers believe that the brand can meet their requirements, and consumer interests are the priority when consumer disputes occur [24].
Brand trust indicates that consumers believe the brand has authentic product features [36]. Therefore, trust is consumer confidence that a brand business operator is fulfilling commitments and selling their product at a reasonable price. Brand trust has two dimensions: brand credibility (i.e., a brand has the ability to keep its promises and meet customer requirements) and consumer protection (i.e., a brand has the ability to protect consumer interests) [69]. Accordingly, a trustworthy brand always keeps its promises to consumers during development, production, sales activity, service provision, marketing activity, and product crises. Therefore, a trustworthy brand image helps the brand maintain a long-term cooperative relationship with customers. Based on this description, the present study combined the views of Bruhn et al. [14] and Habibi et al. [37] and defined brand trust as consumer confidence in the credibility of a brand and in the owners of said brand to fulfill their commitments.

Brand Passion
Brand passion is a core element of the consumer-brand relationship [74]. According to the triangular theory of love proposed by Sternberg [73], love is composed of three components, one of which is passion [74]. This theory has been applied in studies investigating consumption behavior. According to Bauer et al. [9], brand passion is a strong positive feeling toward a specific brand that leads to emotional attachment and affects related behavior and also explains consumer passion within consumer-brand relationships. In summary, passion is a strong desire and feeling, and when a person is passionate about someone or something, they exhibit certain behaviors. This study combined the views of Bauer et al. [9] and Albert and Valette-Florence [2] to define brand passion as a strong positive attitude and infatuation toward a particular brand.

Relationships among brand authenticity, brand identification, brand trust, and brand passion
People's trust in other parties is primarily dependent on whether other parties can be understood, what other parties have accomplished, and whether these parties have favorable reputations [4]. Similarly, these characteristics are the elements used by consumers to establish trust in a brand. The antecedents of authenticity (e.g., the consistency of brand marketing strategies and the originality of brand products) can enhance brand credibility and reputation [22]. Brand trust has the following cognitive dimensions: brand credibility, consistency, success, and predictability [11], [24]. When consumers perceive a high level of brand authenticity, they present a high level of brand trust [62]. Brand trust was defined as consumer belief that a brand will keep its promises [18]. The brand commitment of a product with brand authenticity reflects the brand's characteristics and its willingness to fulfill its promises. One study reported that the two aspects of brand authenticity (i.e., brand consistency and continuity) significantly and positively affect brand trust [27]. Brand authenticity was also demonstrated to enhance brand trust [71]. Thus, brand authenticity substantially affects brand trust. Therefore, we proposed the following hypothesis (H1): H1: Brand authenticity positively affects brand trust.
Consumers who perceive high authenticity positively perceive their own characteristics strongly; accordingly, consumers' perception of brand authenticity reflects their personal intentions to be authentic [22]. Swimberghe et al. [74] revealed that being able to reinforce self-concepts or reflect a personal brand positively affects brand passion. Brand authenticity affects the brand identification of consumers who have a close relationship with the brand [22]. Brand authenticity is thus a crucial factor affecting successful brand identification [15]. In addition, brand passion can express consumer values and demonstrate brand identification. Therefore, brand identification is positively related to brand passion [3]. In summary, brand authenticity substantially affects brand passion. Thus, we proposed the following hypothesis (H2): H2: Brand authenticity positively affects brand passion.

Relationships between Brand Identification and Brand Trust
Brand identification is based on consistency between consumers and brands, it also reflects a brand's symbolism in society, and is shared among people in a community [30]. It can reflect consumers' personalities, improve their selfesteem and social status, and help them form a long-term relationship with a brand [79]. Mutual trust is crucial for successful social interactions. Therefore, building a trusting relationship with customers, employees, investors, and other organizations is crucial for companies. Customers often identify with trustworthy organizations as a way to express their self-definitions and enhance their self-esteem [57]. When organizations that have the trust of consumers are competent, benevolent, and honest, the consumers become aware of the relationship between their self-identification and the organization's specific brands. Therefore, we proposed in this study the following hypothesis (H3): H3: Brand identification positively affects brand trust.
Napoli, Dickinson, Beverland, and Farrelly [62] revealed that postmodern consumers use brands to create a true self and reconnect themselves to particular locations, time points, and cultures. When a consumer strongly identifies with a brand, the consumer and other consumers who love the same brand form a brand community [59]; these consumers' self-identification and social identification become brand identification [47]. Brand passion reflects consumer personality in addition to internal emotions. Recent studies have demonstrated that when consumers love a brand, their brand identification is central to their brand passion. Studies have indicated that brand identification enhances brand passion [3]. Thus, we proposed the following hypothesis (H4): H4: Brand identification positively affects brand passion.

The Effect of Brand Trust and Brand Passion on Brand Purchase Intentions
Brand trust can affect consumer attitude and behavior (including purchase behavior, attitudinal and behavioral loyalty, perceptions of brand value, and brand referrals) toward a brand [24], [28]. Becerra and Korgaonkar [11] investigated trust as it relates to the quality of the Internet environment and discovered that brand trust can enhance willingness to purchase online. Similarly, Reichheld [67] concluded that a person's recommendation of a brand to others may affect their own reputation; therefore, consumers tend to recommend only trustworthy brands that meet their expectations. Therefore, authenticity not only inspires trust in consumers but also makes them willing to pay a premium price [13]. Thus, we proposed the following hypothesis (H5):

H5: Brand trust positively affects brand purchase intentions.
Marketing researchers have achieved consensus that passion is the core element of emotional connection and leads to consumer loyalty [9]. Rozanski et al. [68] reported that brand fanatics are passionate and loyal to their favorite brands and therefore develop extreme behavior. Accordingly, consumers can be brand messengers who communicate their brand love by word of mouth and encourage others to purchase and use the brand [17]. Turri et al. [76] concluded that passionate consumers are highly likely to develop repetitive purchase behavior. Based on these arguments, we proposed the following hypothesis (H6): H6: Brand passion positively affects brand purchase intentions.

Sample and Data Collection
We selected the fan clubs of the top three cosmetic brands in Taiwan as the study population, chosen according to the Cosmetic Marketing Research Institution's [23] 2015 Q4 report on the number of views each brand's cosmetic products received on UrCosme.com. The three brands (Lancôme, Kiehl's, and Shiseido) are renowned international cosmetic brands. We sampled participants from the Facebook fan clubs of Lancôme, Kiehl's, and Shiseido. The participants must have purchased a product from one of the three brands and been members of the fan page for at least a year. Most respondents were notified of the survey through postings of online questionnaires on the brands' Facebook fan pages and the forums of beauty and cosmetics website Fashion Guide. Fan page and forum members were encouraged to forward the questionnaire to other members. They were invited to use the online MySurvey platform to complete a questionnaire. The first question was used to determine whether a participant qualified to participate in the survey.
The formal questionnaire survey was conducted from March 1 to June 30, 2016 (approximately 4 months). In total, 568 questionnaires were returned; after removing 84 invalid questionnaires, 484 valid questionnaires remained for data analysis. The valid return rate was 85.21%. The valid questionnaires were completed by 103 male (21.30%) and 381 female (78.70%) participants. Most users of beauty and cosmetic products are women, indicating great potential for product development to cater to men. Men often consult female friends before making purchases, and most men's beauty and cosmetic products are purchased by women.

Measures and Operational Definitions
In this study, various scales were designed based on scales proposed in other studies. A seven-point Likert scale was adopted for the brand authenticity and brand identification scales (1 = strongly disagree and 7 = strongly agree).
A five-point Likert scale was used in the brand trust, brand passion, and brand purchase intention scales (1 = strongly disagree and 5 = strongly agree; refer to Appendix 1 for the items in each scale). The adoption of preventive measures can effectively eliminate common method variance (CMV), which is particularly severe when respondents repeatedly encounter the same types of questions and scales. Therefore, to maintain the stability of the original author's scale design and to prevent single-source bias, this study followed the original questionnaire scale designed by its authors to prevent the respondents from producing errors of consistent responses when encountering items from a single scale. When the questionnaires were analyzed, the scale scores were first standardized before causality verification was conducted.

Brand Authenticity
In this study, following the measured dimensions proposed by Bruhn et al. [15], brand authenticity was operationally defined as the fans of a brand considering its continuity, originality, reliability, and naturalness. For the brand authenticity scale, we used 15 items originally proposed by Bruhn et al. [15].

Brand Identification
Similarly, following the definitions proposed by Escalas and Bettman [30] and Yeh and Choi [81], brand identification was operationally defined as the degree to which brand community members use the brand in question as a reference point to compare themselves with nonusers of the brand and consider the brand's image part of their selfimage. In the brand identification scale, we used three items originally used by Yeh and Choi [81].

Online Brand Relationship Factors: Brand Trust and Brand Passion
We divided brand relationship factors into three categories: brand trust and brand passion. Our notions of brand trust mainly followed from the views of Bruhn et al. [14] and Habibi et al. [37], who operationally defined brand trust as brand fans considering a brand trustworthy and reliable. Brand passion mainly followed the views of Hwang and Kandampully [45] and Mohammadian et al. [60], who operationally defined brand passion as brand fans being passionate about a brand. In total, six items were used to assess brand trust (three items originally used by Habibi et al. [37]) and brand passion (three items originally used by Kim et al. [49]).

Brand Purchase Intention
Brand purchase intention was defined primarily by following the definition proposed by Becerra and Badrinarayanan [10]; the definition was revised to be the likelihood of a brand community member purchasing a brand's product. For the scale, four items originally used by Becerra and Badrinarayanan [10] were used.

Results
Structural equation modeling and the statistical software package AMOS 18.0 were used to analyze data. Following Anderson and Gerbing [5], structural equation modeling was used to analyze data in two stages. In the first stage (measurement model analysis), we determined whether the latent variables could be correctly measured from the measured variables. To assess discriminant validity, we conducted three tests. First, the average variance extracted (AVE) criterion proposed by Fornell-Larcker [32] was employed to test whether the square root of the AVE for each construct was higher than all correlations of the considered construct. Second, the present study adopted the heterotrait-monotrait (HTMT) technique to ensure the level of adequacy of each construct varied between HTMT0.90 and HTMTinference. Third, we used the cross-loading method [19] to test whether each construct could be effectively distinguished. In the second stage (structural model analysis), we used structural pattern analysis to determine the causal relationships between various variables. Finally, the mediation effects of brand trust and brand passion were examined through a Sobel test, bootstrapping, and regression analysis.

Measurement Model
We adopted in this study the assessment criteria for convergent validity suggested by Anderson and Gerbing [5], the assessment standards for confirmatory factor analysis proposed by Bagozzi and Yi [6], and the fitness index developed by Gefen et al. [34]. The measurement model developed in the present study had an excellent fit to the data (χ 2 /df = 2.919, goodness-of-fit index (GFI) = 0.853, adjusted goodness-of-fit index (AGFI) = 0.823, comparative fit index (CFI) = 0.955, and root mean square error of approximation (RMSEA) = 0.063). According to Hu and Bentler [44], a GFI or AGFI of 0.8 or higher is acceptable. In the present study, composite reliability and AVE were assessed according to the suggestion of Gaski and Nevin [33] that they should be greater than 0.7 and 0.5, respectively. Therefore, the convergent validity of each subscale used in the present study was acceptable. Table 2 presents the results of the measurement model analysis. Podsakoff et al. [65] advised to use multiple methods and control for CMV. This study used the following three methods to test as follows. Following Fornell and Larcker [32], we first used AVE to assess discriminant validity. If the square root of the AVE of a latent variable was greater than the correlation coefficient between the two underlying dimensions, then the two underlying dimensions possessed excellent discriminant validity. The results verified that correlation coefficients between every pair of constructs were less than the square root of the AVE for each construct (Table 3).
Henseler et al. [41] stated that in variance-based structural equation modeling the HTMT technique provides superior assessments of discriminant validity to those of AVE. Therefore, we adopted the HTMT technique to assess discriminant validity to ensure the adequacy of the model constructs. If the HTMT confidence interval value is less than 0.9, it means that the model has discriminant validity. Table 4 shows that the HTMT value of each construct was less than 0.9; therefore, no discriminant validity problem existed in this study.   The cross-loadings of each construct shown in Table 5 were determined by calculating the loadings of each item on its own construct and the cross-loadings on other constructs. The factor loading of each item on its own construct was higher than the cross-loadings on other constructs; thus, no significant overlap was observed between any constructs in the overall model. Each construct was effectively distinguished. Tables 3-5 indicate good discriminate analysis based on the criteria of Fornell and Larcker [32], Hair et al. [38], and Henseler et al. [41].
The value γ denotes the path relationship between an exogenous variable and an endogenous variable, whereas β denotes the path relationship between two endogenous variables. Regarding the hypothetical model, brand authenticity was discovered to significantly and positively affect brand trust (γ11 = 0.853, p < 0.001) but did not significantly affect brand passion (γ21 = 0.104, p > 0.05), whereas brand identification significantly and positively affected both brand trust (γ12 = 0.245, p < 0.001) and brand passion (γ22 = 0.845, p < 0.001). Additionally, brand trust (β41 = 0.223, p < 0.001) and brand passion (β42 = 0.200, p < 0.001) significantly positively affected brand purchase intentions. In summary, all hypotheses other than H2 were supported (Fig. 2).
The R 2 values of brand trust, brand passion, and brand purchase intentions were 0.726, 0.550, and 0.569, respectively. Overall, our research model explained the data well.   Regarding consumer behaviors, women are generally expected to be more emotional than men. Therefore, this study investigated whether men and women exhibit differences in brand passion. The independent sample t test revealed that the F test for brand passion yielded a significant result (p = 0.491 > 0.05). The variances of the two groups exhibited no significant difference. The calculated t statistic was 2.821 with two-tailed significance of p > 0.05. Thus, the null hypothesis could not be refuted. Unexpectedly, men exhibited significantly higher brand passion toward beauty and cosmetic products than did women (Table 8).

Discussion
In this study, we constructed a brand model by extending brand authenticity and brand identification to discern how various factors affected the brand purchase intentions of Facebook members who used the fan page of three popular cosmetic brands. According to the results, brand authenticity and brand identification affected brand purchase intention through brand trust and brand passion. Therefore, we identified four crucial paths: brand authenticity → brand trust → brand purchase intention; brand authenticity → brand passion → brand purchase intention; brand identification → brand passion → brand purchase intention; and, brand identification → brand trust → brand purchase intention. A customer-brand relationship based on these four paths can be a successful long-term relationship. Moreover, this study revealed that brand trust and brand passion both had mediating effects. In addition, men exhibited greater brand passion toward beauty and cosmetic products than did women.
This study determined that brand trust and brand passion significantly and positively affect brand purchase intention, again in agreement with previous studies [1], [74]. Thus, people who trust and love a brand are highly willing to purchase the brand's products. Therefore, brand trust, and brand passion are critical factors that motivate brand fans to demonstrate strong support for a brand.

Managerial implications
Brand authenticity was discovered not to immediately arouse brand passion. This may have been because in the cosmetics industry, some product access channels are cosmetics stores that rely on outdoor advertising and multichannel marketing instead of media or Internet marketing; in addition, professional consultation services are often provided to consumers in cosmetic stores. Although some articles about brand products had been posted on the fan pages of the brands considered in this study and other activities had been undertaken to interact with fans or motivate fans to share on social networks, no active marketing strategies highlighting brand image and spirit had been developed, and the positioning of the brands was ambiguous. Accordingly, fans did not have a close connection with the brands, thereby decreasing their brand passion. Nevertheless, brand authenticity aroused brand trust, which enhanced brand purchase intention. Therefore, brand authenticity is crucial to brand management [71].
In the current consumer market, brand authenticity substantially affects consumer behavior and is the basis of brand trust. Therefore, business operators must be clear about their brand's values, be loyal to the brand's spirit, and honestly fulfill their brand commitments if they are to establish a genuine, natural, and honest brand image and demonstrate brand discriminability and long-term brand consistency.
Loofah water, produced by Kuang Yuan Liang, can be used as an example. This brand is a Taiwanese cosmetics brand that incorporates Taiwanese culture and possesses local characteristics; the product development and brand concept reflect the Taiwanese spirit. In addition, the product is inexpensive and made from local materials, and its advertisements prominently feature Loofah farmers [55]. The brand's management focused on brand spirit; therefore, the product was successfully accepted by consumers and won consumer trust. Bruhn et al. [15] reported that business operators should develop an integrated brand image and enhance brand authenticity. For example, businesses can use media platforms to continually present brand names, logos, and slogans and increase brand credibility. Articles posted on fan pages can thus be used to convey a unique brand spirit (including the original intention of the brand founder and their brand commitments), enhance fans' understanding of the brand's authenticity, increase fans' brand trust, and help fans form brand purchase intentions.
If a brand can reflect consumers' personalities and ideal selves and enhance social status, then the brand can help the consumers reinforce their self-identification. Additionally, brand trust and brand passion strengthen brand purchase intention, which can in turn help consumers develop emotional attachments and convince them to purchase the brand's products. Consumers can use their favorite brands to express their real self or reflect their ideal self. Therefore, business operators should strive to evoke sympathetic responses to brand characteristics and values to create a close connection between a brand and its consumers.
To enhance brand relationships, business operators can invite brand users to post product photos or videos to the brand or product fan pages, write articles to introduce products, and indicate that they are brand fans. Facebook friends and brand fans can click on the like icon if they like a product photo, video, or article, and brand users whose post obtains the most likes could be rewarded. In this manner, product exposure can be increased and brand passion can be aroused, ultimately leading to brand purchase.
Skin care for men is a contemporary societal phenomenon. As the market of women's skin care products becomes saturated, an effective understanding of men's demand for specific skin care products generates a level of brand passion among men that far exceeds that among women. For example, Lancôme has established itself as a brand that focuses on developing professional skin care products for older men; 37% of its customers are over the age of 40. Biotherm, another skin care company, noted an increasing proportion of men actively seeking skin care and aging prevention, and that men demonstrate greater loyalty in skin care product selection. This indicates that that men's brand passion toward beauty and skin care products has slowly overtaken that of women.
The Internet is powerful and penetrative; fan pages are an excellent tool for marketing brands and strengthening consumer-brand relationships [46]. Business operators should endeavor to create an atmosphere on a fan page in which brand fans actively interact with one another; if they succeed, consumers will establish brand trust [31] and emotional attachment [58]. Fan pages are thus crucial channels for business operators to strengthen the consumerbrand relationship.
From a management perspective, online community managers should concentrate on brand characteristics and identification. A unique brand spirit and characteristic can affect brand trust, and brand identification is crucial for strengthening the relationship between community members and brand manufacturers. Community managers must understand how to maintain these relationships. Brand trust and brand passion played a critical role in our research model, but these qualities cannot be formed overnight. Therefore, how brand manufacturers can strengthen the relationship between community members and their brands is an urgent issue.

Limitations and Future Directions
In this study, we endeavored to meet all the requirements of management science research; however, this study did have some limitations. The limitations and corresponding suggestions are as follows.
First, the present study conducted a cross-sectional survey; future studies could adopt a longitudinal method to observe how community members make purchase decisions about their favorite brands, understand the causal relationships between various variables, and develop an effective assessment tool for subsequent analysis and verification. Second, a nonrandom sampling method was used in this study, and the participants were required to complete a questionnaire regarding their favorite brands. In future research, other types of social groups (e.g., discussion forums) could be investigated. Third, the research targets in this study were cosmetics brand fans, and most participants were women; therefore, results may have been biased. We suggest that in future studies, equal numbers of men and women should be recruited.
Finally, in our research framework, some psychological factors were not considered. Brand romance is one such factor and describes the passion about a brand and the close relationship formed with it when a person is in love with a brand [64]. Another psychological factor is the cocreation experience [13], which is the sharing of purchase experiences among community members that can affect online community members' purchase intentions. If these factors are considered, brand loyalty may be reinforced, and brand fans may become enthusiasts who will defend and recommend the brand. Consequently, the resulting economic benefits will likely exceed those gained from more traditional advertising.