Brand Loyalty of Millennials to Online Fashion Brands; E-commerce Sites in Kota Kinabalu, Sabah

As e-commerce has rapidly grown in past decades, the fashion industry in Malaysia has experienced positive growth in economic values. As millennials are the biggest spenders in Malaysia, their loyalty is often left behind. In fact, millennials’ brand loyal ty is difficult to attain by e-commerce sellers because of their preferences in buying online products and their tendency to be cautious of marketing techniques to avoid being manipulated. Despite immense spending power, millennials were found to be very disloyal to brands. This study aims at examining various literature on brand loyalty. There are three theories integrated into this study which consist Technology Acceptance Modal (TAM), Expectation Disconfirmation Theory (EDT), and Theory of Reasoned Action (TRA). It also aspired to identify the relative importance of income as a moderator to trust as one of the most critical factors influencing millennials’ brand loyalty to online fashion brands’ e -commerce sites. In addition, this paper seeks to determine the moderating effect of income on the relationship between trust and millennials’ brand loyalty. Based on empirical studies in marketing literature done by previous researchers, this study came out with a conceptual framework pertaining to the relationship between website efficiency, customer service, trust, and brand image on millennials’ brand loyalty ware explored empirically. The study involved three hundred and ten millennial respondents, and the data collected was analyzed using Statistical Package for Social Sciences (SPSS) version 25 and the Smart-PLS version 3 software. The findings in this study illustrate that customer service, trust, and brand image significantly impact brand loyalty, and income moderates trust and brand loyalty. In conclusion, besides contributing to enriching the literature, the study can assist policymakers in developing and formulating strategies to understand online customers, particularly millennials in Malaysia.


Introduction
The fashion industry contributes 24% of global sales in the global e-commerce market and is expected to reach $ 672,710 million in two years by 2023 (Alanadoly & Salem, 2022;Keegan et al., 2021). However, the fashion industry is struggling in the face of brand loyalty toward their products, especially among millennials (Cabigiosu, 2020;Ng et al., 2017), as millennials are difficult to be reached (Chatzopoulou & de Kiewiet, 2021;Soares et al., 2017). The advent of e-commerce has brought life-changing to business environments, and it seems to be a vital platform for the fashion industry because it is a medium of interplay and unlimited communication with customers and prospects (Soni et al., 2019). With the huge amount of advertisements every year, millennials tend to be more aware of marketing techniques and avoid being manipulated (Lewis et al., 2021). Millennials are obsessed with fashion and spend two-thirds of their money on clothing (Medalla et al., 2020;Lianto et al., 2021). Despite having large spending power (Husain, 2022), millennials were found to be very disloyal to a brand (Goranda et al., 2021;Munsch, 2021;Diniso & Duh, 2021).
In Malaysia, millennials' brand loyalty, fashion industry, and e-commerce are still inconclusive and have been a fascinating subject among scholars in the marketing literature (Shaari et al., 2022;Madan et al., 2022;Bostanudin & Rashid, 2021;Ruey & Ismail, 2021;Zhen & Hassan, 2021). Benedict (2021) reported that 61% of millennials spent on clothes during phases 1-3 of the Movement Control Order (MCO) in Malaysia. Although millennials are driven to spend more on clothing, this category has been classified as disloyal customers compared to other website efficiency, customer service, trust, and brand image have been identified among factors affecting brand loyalty (Kim, 2019;Cao et al., 2018;Bashir et al., 2018;Hassanzadeh & Namdar, 2018;Kim & Kim, 2017). In addition, there is little study regarding the level of income as a demographic factor, especially trust in millennials' brand loyalty. This study further investigates the moderating effect of income to brand loyalty which is still inconclusive because lower-income groups tend to be more at risk for their online purchasing (Akbarov, 2020;Tull et al., 2020;Kim, 2019;Kim & Kim, 2018). Thus, this study examines the factors affecting millennials' brand loyalty to online fashion brands' ecommerce sites in Kota Kinabalu, Sabah context. This study also contributes to body of knowledge regarding to millennials' brand loyalty which is in line with the Theory of Reasoned Action (TRA), Technology Acceptance Modal (TAM), and Expectation Disconfirmation Theory(EDT). Therefore, the objective of this study is to examines factors that contribute to millennials' brand loyalty to fill the gap in research, particularly in the context of e-commerce sites in the fashion industry. Independent variables such as website efficiency, customer service, trust, and brand image were used as predictors of the dependent variable (an outcome), and income exhibits a moderating effect on the relationship between trust and millennials' brand loyalty in Kota Kinabalu, Sabah. Table 1 shows the objectives of the study.

Literature Review
The underpinning theory of brand loyalty has been used by many researchers in their fields of study. Although many factors contribute to brand loyalty, in this study, four main independent variables will be highlighted, website efficiency, customer service, trust, and brand image in respect of the dependent variable, brand loyalty. In the present research, each independent variable in this study uses a specific theory to fit with the dependent variable as robust and has an antecedent connection between them. Although the marketing field lacks theories associated with advancing in marketing, the researcher has adopted the closest theories related to the research framework suggested by Varadarajan (2020). The underpinning theories used in this study (Table 2), which consist of Technology Acceptance Modal (TAM), Expectation Disconfirmation Theory (EDT), as well as Theory of Reasoned Action (TRA). Theory of Reasoned Action (TRA)

Brand Image
Theory of Reasoned Action (TRA) The first theory used in this study is the Technology Acceptance Model, or TAM Model (Davis, 1985). According to Ahmad et al (2020), several studies used TAM base to examine website efficiency (Ahmed et al., 2020;Vijayasarathy, 2004;Lederer et al., 2000). Meanwhile, according to a prior study conducted by Dadzie, Chelariu, & Winston (2005), the authors used EDT towards Logistic Chain Supply (LCS) on loyalty effect, and they found out that customer service such as action on complaints, on-time delivery, and completeness of order filled is crucial assessments in order to make customer loyalty and repurchase. Apart from that, A recent study employed TRA on brand image and trust to investigate customer loyalty (Singh et al., 2021).
The theoretical foundation of the relationship between variables such as the effect of website efficiency, customer service, trust, and brand image on brand loyalty is highlighted in this study. Moderating effects of income on the relationship between trust and millennials' brand loyalty is also addressed. To illustrate, website efficiency means how quickly a customer can search for information with minimal effort based on website accessibility and usability (Kim, 2019;Kim & Kim, 2017), customer service is provided to help customers solve their problems promptly through online representatives related to order instruction, shipping and handling, and return process information (Kim, 2019;Kim & Kim, 2017), trust is customers' assurance, and the privacy of customers' data is being protected through the site's safety (Kim, 2019), brand image refers to a brand's identity, which is an impression of a brand and becomes the top identity regarding customers' opinions of other brands (Martinez et al., 2009), brand loyalty refers to a customer's positive perception of the online sellers and encourages customers to repurchase (Srinivasan et al., 2002), and income refers to an individual financial (total gross expenditure and savings in a particular period), where it is easily identifiable, accessible and measurable (Akbarov, 2020).

Methodology
The quantitative study design emphasizes on factor affecting millennials' brand loyalty to online fashion brands' e-commerce sites in Kota Kinabalu, Sabah used in this study. The researcher used primary data collected through an online questionnaire (Google Form), and only the millennial category (23 to 45 years old in 2022) was allowed to answer the survey. Stratified random sampling involved 6 universities with 310 respondents around Kota Kinabalu namely Universiti Malaysia Sabah (UMS), Universiti Teknologi Mara (UiTM), Aseana College, INTI College, North Borneo University College (NBUC) and Open University. IBM© SPSS® Statistics Version 25 and SmartPLS 3.2.8 have been used for data analysis. This site offers a good satisfaction guarantee (e.g., money back guarantee) because this site desires its customers to be satisfied with their purchases.  Table 4 depicts the outcomes of this study's reliability test and validity. According to Hair et al (2010), Construct Reliability (CR) in Cronbach's Alpha should exceed the threshold value of 0.60 for a lower acceptable reliability level. The Construct Reliability (CR) scores for combinations reveal; Website Efficiency (0.928), Customer Service (0.873), Trust (0.917), Brand Image (0.918), and Brand Loyalty (0.946). Composite Reliability (CR) values for each formulation range from 0.873 to 0.946 and these scores are considered acceptable. Apart from that, Cronbach's alpha (α) values were observed for both structures, and these scores were found to be acceptable since the range value for Cronbach's alpha (α) was between 0.835 to 0.933. Hair et al. (2014) pointed out that convergent validity is a term that refers to average variance extracted (AVE), whereby an AVE percentage value of more than 0.50 is deemed acceptable (AVE ≥ 0.50). AVE scores in this study for website efficiency (0.647), customer service (0.579), trust (0.612), brand image (0.615) and brand loyalty (0.715). In this study, four direct hypotheses were formed to test whether it fits with a significant level by calculating data into numerical values in SmartPLS 3.0, employed with bootstrapping feature, with subsample 5,000, to generate t-satistical for all indicators. Formulated into numerical values of the path coefficient, shown in Table 5, only four of the relationships exhibit t-values ≥ 1.645, hence significant at 0.05 level of significant (Hair et al., 2014). This can be illustrated by the predictors of millennials' brand loyalty attributes, i.e., website efficiency (β= 0.076, P < 0.146) was found to be negatively correlated with millennials' brand loyalty, and customer service (β= -0.097, P < 0.020), trust (β= 0.360, P < 0.000), and brand image (β= 0.4444, P < 0.000) were found positively related to millennials brand loyalty. Therefore, H2, H3, and H4 are supported based on the P-value. The direct effect of income exhibits a moderating effect on the relationship between trust and millennials' brand loyalty to online fashion brands' e-commerce sites is also tested. It was found that income*trust* (β= -0.117, P < 0.001) is moderate on brand loyalty. Hence, H5 and H3 are significant. Table 5 shows hypotheses testing results where the predictor R2 for brand loyalty has 49.7% (0.497), which indicates that the variances in brand loyalty fulfil the suggested R2 value, which must be higher than 0.25 to demonstrate a substantial model (Cohen, 1988). As suggested by Cohen (1988), the level of coefficient of determination R2 values for endogenous latent variables is assessed as 0.26 (substantial), 0.13 (moderate), and 0.02 (weak). In this study, the CMB results of the Variance Inflation Factor (VIF) revealed website efficiency (1.032), customer service (1.231), trust (1.117), brand image (1.752), and brand loyalty (1.586). This proved that CMB was not a major concern in this research.

Results and Discussions
The fashion industry, particularly in Malaysia, showed a limited focus on millennials; most previous studies focused on other generations. Nonetheless, empirical studies were developing extended theories of Technology Acceptance Modal (TAM), Expectation Disconfirmation Theory (EDT), as well as Theory of Reasoned Action (TRA). This study also highlights that income moderates the relationship between trust and millennials' brand loyalty.The present study findings declared that the four hypotheses in the conceptual research model were supported. The findings of this study revealed that website efficiency had a negative impact on millennials' brand loyalty in Malaysia. The early empirical studies confirmed this study's results (Crutzen et al., 2011;Marimon et al., 2010). The research findings also showed that customer service had a significant impact on millennials' brand loyalty. The results supported by previous studies stated that customers would be satisfied when customer service is excellent and generates brand loyalty provided by e-commerce companies (Jenneboer et al., 2022;Brilliant & Achyar, 2021). Besides, the results also stated that trust positively influenced milennials' brand loyalty as online businesses should provide sufficient information, precise terms, information about store locations, dispute contact options, accurate security and policy terms, and display security certification. The prior empirical studies approved this research's findings (Cyr;2008;Flavian et al., 2006;Yoon, 2002). Furthermore, the research consequences also unveiled that brand image had a positive influence on millennials' brand loyalty. The prior empirical examinations verified these outcomes (Oetama, 2022;Anwar et al., 2019). The findings in this study also illustrate that income moderates trust and brand loyalty. Previous studies confirmed that income as a moderator is a crucial criterion related to satisfaction and loyalty (Akbarov, 2020;Henrique, & De Matos, 2015;Walsh et al., 2008).

Conclusion
This research exhibits evidence in compliance with research questions with the hypotheses testing on factors affecting millennials' brand loyalty to online fashion brands' e-commerce sites in Kota Kinabalu, Sabah. Independent variables such as website efficiency, customer service, trust, and brand image have been used, and a dependent variable in this study is brand loyalty. Income has been addressed as a moderator effect on the relationship between trust and millennials' brand loyalty. The findings in this study revealed that independent variables such as customer service, trust, and brand image significantly impact millennials' brand loyalty to online fashion brands' e-commerce sites in Kota Kinabalu, Sabah. This study showed that income moderates the relationship between trust and millennials' brand loyalty.