Modeling Fashion Brand Authenticity Toward Brand Usage Intention: Evidence from Indonesia

Abstract

The growth of local fashion brands in Indonesia is interesting because of the speed with which new designs are made, many variants are created, and local brands are developed. However, when it comes to buying local fashion products, the brand isn't the sole consideration. As a result, the marketing effectiveness of local fashion brands needs to be examined more thoroughly. The purpose of this study was to understand more about brand authenticity and how it affects brand usage intentions. The unit of analysis in this study was Generation Z, who purchase from local fashion companies online. A quantitative research methodology was used. The data was analyzed using the PLS-SEM method to test the hypotheses with the dependent variables. The variables were the intention of using the brand, the quality of the brand relationship, and the authenticity of the brand. The results of this study indicated that brand authenticity had an effect on brand usage intent and brand relations quality mediates the positive effect of brand authenticity on brand usage intent.

1. Introduction

The global fashion industry is evolving along with the culture. The market is growing rapidly and is dominated by big brands which are categorized as global fashion trend setters. However, the growth of local fashion brands in Indonesia is something that attracts attention because of its speed in producing local designs, variants, and brands. The fashion sector contributes 18.01 percent to GDP, with a growth rate of roughly 4.05 percent every year, which is the second-highest growth rate in the Indonesian economy, according to the Central Bureau of Statistics. Furthermore, the data shows that the percentage contribution is around 166.1 trillion rupiahs, with the greatest export value of US$ 10.9 million. This expansion is beneficial because this sector can employ 4, 130, 000 people. Indeed, given that 77.56 percent of fashion entrepreneurs established their businesses between 1990 and 2014, the growth could have been influenced by existing developments. 865 entrepreneurs in the fashion sector recorded an income of IDR 50 billion per year, which is also the third-highest in Indonesian statistics (Statistik, 2018).

The Central Statistics Agency explained that in September 2020 Indonesia entered a demographic bonus, namely a situation where the population aged 15–64 years was greater than the population aged under 15 years or above 64 years (Bappenas, 2017). Practically, it is interpreted that the number of people of productive age is greater than the number of people of unproductive age. The circulation of global fashion brands in Indonesia is a signal that Indonesia is both a market and a source at the same time. The thing that needs to be considered then is the ability to manage the fashion brands so that Indonesia can not only become a leader of multinational brands but also can activate local fashion brands and manage local fashion brands as well.

Z Generation is a potential local fashion brand market. This generation was born between 1997 and 2012 and grew up with easy access to the Internet, technology, and social media which made this generation stereotyped as technology addicts and anti-social (Insider, 2019). Z Generation spent more than 8 hours connected to the Internet and continuously use various application platforms. It was discovered that fashion is the most purchased product through Shopee’s online shopping medium among the Z Generation (CX-GO, 2021). In contrast to its predecessors, the Z Generation today expresses its aspirations differently, preferring to seek out more real and original brands. As a result, the Z Generation is the target market for the local fashion business.

In Indonesia, there are local brands that regularly hold exhibitions to be able to get closer to the customers. The following research objects are the names of the 15 fashion local brands Niion, Chiel Shoes, Insurgent Club, Save My Monday, Jackhammer, Shoesayhelp, Pattent Goods, Footstep, Tailorwerk, Northy, Groot Watch, Amble, Terrell, Wallts, and Miracle Mates.

2. Literature Review

Aaker, Keller, Yoo, and Donthu, as well as Luming Wang and Adam Finn, created the notion of Consumer Based Brand Equity. The consumer-based brand equity model’s basic premise was that the strength of a brand was determined by what customers learned, felt, experienced, and heard about it as a result of the company’s efforts over time. Keller (2013) described customer-based brand equity as the difference in customer responses to a company’s marketing efforts as a result of brand knowledge. The brand structure, company, and people involved in the product were all considered in a complete approach to brand building (Keller, 2013; Ajzen & Fishbein, 1980).

A brand is a name, word, sign, symbol, design, or a combination of them, that identified the maker or seller of certain products and services. Consumers saw the brand as an important part of the product because a brand can add value to the product. The most prominent professional marketer skill was the ability to create, maintain, protect and enhance the image of a brand (Kotler, 2004; Ashley & Tuten, 2015). A brand is a company’s long-term investment which, if managed optimally, would provide great benefits for the company. Brand engagement would only occur if consumers experienced the direct benefits promised by the company (Kotler, 2004). Brand experience ranging from affective, cognitive, behavioral, and relational had a significant effect on brand loyalty. In the process, the level of brand involvement had a significant influence on the relationship between brand experience and brand loyalty. Practically, it was important to use branded engagement as part of a company’s marketing communication process (Kim et al., 2012).

H1: Brand Authenticity had a positive effect on Brand Relationship Quality.

Increasing customer satisfaction and confidence would encourage repurchase intentions (Hsu & Lin, 2015; Prawira & Sihombing, 2020; Goeltom et al., 2020; Nguyen, 2022). Previous research has found that the intention to use a brand is influenced by customer satisfaction. Customers as recipients of these stimuli were positively affected by communication. Hence brand communication was positively connected with brand equity as long as the message encouraged customer actions toward the branded product in question, as opposed to unbranded goods (Yoo & Lee, 2012). Effective brand communication can boost brand equity by conveying information about the products that buyers remember, influencing their decision-making process significantly (Keller, 1993).

H2: Brand Relations Quality has a positive effect on the Intention to Use a Brand.

H3: Brand Relations Quality mediates the positive effect of Brand Authenticity on Brand Usage Intent.

A conceptual framework illustrates what you expect to find through your research. It defines the relevant variables for your study and maps out how they might relate to each other. You should construct a conceptual framework before you begin collecting data. It is often represented in a visual format. The study hypothesis was described by three variables with three routes shown as arrows in this research model. The following was the image of the research model, along with the hypothesis.

Figure 1: Research Framework

3. Research Method

This study was an explanatory study on how brand authenticity impacted the quality of customer relationships, resulting in the consumers’ intention to utilize the brand. This study relied on a model from previous research (Sekaran & Bougie, 2016). This study was non-interventional in nature. This means that during the research time, no extra treatment or intervention was given to the research subjects.

In this study, only through observation of the subject and the use of a questionnaire created in accordance with the data for the research model was gathered. Variables were items that could be seen, measured, and had a range of values. The value of a variable’s data could be nominal, ordinal, scale, or ratio (Sekaran & Bougie, 2016). There were three types of variables in this study. The first was the dependent variable, which represented brand intention. The quality of the brand relationship was used as a mediating variable. Brand authenticity was used as an independent variable. In the research design, one of the important stages was determining the variables in the research model. A dependent variable also called an outcome variable, is the result of the action of one or more independent variables. It can also be defined as any outcome variable associated with some measure, such as a survey (Sekaran & Bougie, 2016).

The sample in this study was Generation Z who were consumers of local online fashion brands in Indonesia and had seen the marketing content of the brands from 2020 to 2021. As the size of the population of this study could not be certainly determined or estimated, a separate calculation was used to obtain the sample. In this study, the researcher used the proportional formula introduced by Lemesow et al. Thus, when using the proportional formula the minimum sample size required in this study is 384 respondents (Lemeshow, 2013). In this study, the sampling method was carried out using the purposive sampling method. It meant that the sample was taken from certain target groups of the population. The purposive sampling method limited the number of people who could offer the needed information, either because they already had it or because they satisfied the study’s criteria (Sekaran & Bougie, 2016). Generation Z, who bought local fashion labels directly online, was one of the respondents’ requirements. A link to the questionnaire was emailed if the responder satisfied these conditions and was willing to participate.

4. Results and Discussion

According to data collected from questionnaire distribution in 2021, respondents were Generation Z who did online purchasing for fashion products. All the data collected during the survey came from 703 participants, and 598 of them met the requirements. The compatibility of the launch of local fashion brands with the choice of local fashion brands that were used as a reference in filling out the questionnaire was then used to check respondents’ eligibility. 499 replies fit the criteria up to this point. Because the minimal number of sample criteria were reached, the number of responses met the standards.

The outer model was analyzed by testing the discriminant validity of the model (Hair et al., 2019). The next step in measuring the outer model is testing the reliability and validity of the construct (Table 1).

Table 1: Construct Reliability and Convergent Validity

From Table 1, it is found that all variables have a Cronbach alpha value above 0.7. Furthermore, in the composite reliability table, all variables have values between 0.7–0.95, hence no redundancy issue was found. According to the results of the consistency test, all indicators in this study model have been deemed reliable for measuring their respective constructs. The average variance extracted (AVE) value of the variables in the study model has a value of more than 0.50, as required, as shown in the table above (Hair et al., 2019). The indicator of the Brand Authenticity variable has the smallest AVE value of 0.543, although it is still above 0.5. As a result, the indicators in this research model are valid for measuring their respective constructs.

The HT/MT Ratio in Table 2 shows the discriminant validity test results where the heterotrait-monotrait ratio (HT/MT) of each variable is found to be below 0.9. Based on these results, it can be concluded that all indicators in the research model have been well discriminated against so that they can measure their respective constructs.

Table 2: Ratio HT/MT

Based on the four parameters of the reliability and validity test results on the outer model as above, namely the reliability indicator (with the outer loading value), construct reliability (with Cronbach’s alpha value and composite reliability), construct validity (with the average variance extracted or AVE value), and discriminant validity. A general conclusion can be drawn that in the measurement model of this study all indicators are reliable and valid to measure their respective constructs specifically. With these results, this research stage deserves to be continued in the next analysis stage, the inner model test or structural model.

The inner model analysis test aimed to assess the relationship between latent variables (constructs) in a research model. In the inner model analysis stage, a one tailed hypothesis test was carried out according to the proposed hypothesis (Hair, 2017). In an empirical test, the output of the inner model test had to be assessed and reported on the quality of the suggested research model before the hypothesis could be tested. The Variance Inflation Factor (VIF), R-square, f-square, Q-square, and Q-square predict were the model quality parameters used in the inner model (Hair et al., 2019). This was necessary to explain and predict the proposed research model’s ability.

The IPMA rescales the data to provide performance scores on a scale from 0 to 100. For the correct rescaling, the original scales of data are essential information. Here, the user can check and correct the possible ranges of the manifest variables. For example, a 7-point Likert sale must have a minimum value of 1 and a maximum value of 7 in the IPMA settings (Hair et al., 2019). The bootstrapping output in the inner model image above showed a research model with one dependent variable, one mediating variable, and one independent variable. The T-statistical values of the three paths in the research model are well displayed in the inner model. If the T-statistic value is more than the T-table value, this value indicated a significant relationship. The phases and extensive explanations of the outcomes of the inner model, as recommended by Hair et al. (2019), began with an evaluation of the model’s quality. The following are the findings of the research:

The first step in analyzing the inner model was to see whether there was a problem with collinearity among the variables. For the multicollinearity test, the value used was the inner variance inflation factor (VIF), where the optimum value was less than 3. If the VIF value is higher than 5, there can be multicollinearity concerns in the research model, which would affect the path coefficient value in the research model (Hair et al., 2019). The VIF values for all variables were found to be less than 3. So it could be interpreted that all the variables in the research model the inner VIF value were ideal. Based on this, it could be said that among the variables in this research model there was no multicollinearity problem found. These results indicated that the quality of the research model had been accepted in the event as there were no multicollinearity issues.

The second stage in the inner model analysis was to assess the quality of the research model from the R-square value (Table 3).

Table 3: R-square Value

Table 3 shows that the R-square value for the brand relation quality variable was 0.078, putting it in the weak category. The independent variable explained 7.8% of the dependent variable brand relation quality, while the remaining 92.2 percent may be explained by variables outside of this research model. An R² value of 0.477 was observed for the variable brand usage intention, putting it in the weak category. This shows that the independent variable explained 47.7% of the dependent variable brand authenticity, while the remaining 52.3 percent can be explained by variables outside of this research model.

The f-squared test, which is used to identify the effect size or magnitude of the influence of a construct if the value of the R-squared of a target construct changes when particular constructs as predictors are removed from the research model, is the next level of analysis in the inner model. The f-squared test calculates the size of the effect size, which is used to determine whether or not there is a significant influence. According to Cohen (1988), the size of the f-squared or effect size is 0.02 for a small effect size of a latent variable, 0.15 for a moderate effect size of a latent variable, and 0.35 for a large effect size of a latent variable. if f ² is found to be less than 0.15, it is said to have no effect size large enough to be significant. The independent variable’s path from brand usage intention to brand relation quality revealed a significant and big effect size of 0.35, precisely 0.914. (Table 4).

Table 4: f-square Value

The following stage in the structural model analysis is a significance test on the three paths that exist in this research model, which becomes the focus for addressing research questions. The purpose of this test is to determine the significance of the influence between variables in the research model so that it can be generalized at the population level. This test was carried out using the bootstrapping method using re-sampling and processed with SmartPLS™ 3.3 (Ringle et al., 2015). The results of hypothesis testing could be assessed by looking at the two values of the empirical test results - the significant value and the coefficient value. The direction of the coefficients had to be in accordance with the direction of the proposed hypothesis because the nature of this hypothesis was directional so a one-tailed test was carried out. If the T-Statistic value from bootstrapping was greater than the T-table value, assumed as 1.645 (with an alpha of 0.05), then the relationship between variables could be declared as significant (Ringle et al., 2015). The analysis of this research model was carried out using a one tailed hypothesis test with a significance level of 0.05. After seeing the significance, it was assessed to know how much standardized coefficient was in each path. If the test results had met these two conditions, it could be mentioned that the research hypothesis was supported.

The results of testing the H₁ hypothesis revealed that the T-statistical value for assessing the effect of brand authenticity on brand relation quality was 7.253. This value was higher than the T-table value of 1.645, indicating that it had a significant effect. With this significant value, it could be interpreted that this effect could be applied at the population level. The standardized coefficient value of the H₁ hypothesis was 0.279, which meant it had a positive direction and was in accordance with the direction of the hypothesis. Based on the two empirical data, it could be concluded that the H₁ hypothesis is supported. If the authenticity of local fashion brands increased, it would also be followed by an increase in the quality of relationships with local Indonesian fashion brands. This finding was in line with previous research (Bruhn, 2012; Bughin, 2007) that brand authenticity as a consumer’s perception of a brand had attributes that reflected that the brand was unique and original, and that affected the quality of the relationship between the brand and the customers.

The results of testing the H₂ hypothesis revealed that the T-statistical value for assessing the effect of brand relationship quality on the intention to use the brand was 21.373. This value was higher than the T-table value of 1.645, indicating that it had a significant effect. With this significant value, it could be interpreted that this effect could be applied at the population level. The standardized coefficient value Ho the H₂ hypothesis was 0.691 which meant it had a positive direction and was in accordance with the direction of the hypothesis. Based on the two empirical data, it could be concluded that hypothesis H₂ is supported. If the brand relationship quality of local fashion brands increased, it would also be followed by an increase in intention to use local Indonesian fashion brands. This finding was in line with the previous research (Zhang, 2013; Hollebeek et al., 2014; Brodie et al., 2013) that brand relationship quality had a significant effect on individuals for brand intention.

The T-statistical value for assessing the effect of brand authenticity on brand usage intent mediated by brand relation quality was 6.695. This value was higher than the T-table value of 1.645, indicating that it had a significant effect. With this significant value, it could be interpreted that this effect could be applied at the population level. The standardized coefficient on the mediation path with the brand relation quality was 0.192 which meant if the perception of brand authenticity increases, it will be followed by an increase in the brand relation quality and brand usage intent. Based on the two empirical data, it could be concluded that the H₃ hypothesis is supported. This finding was in line with the previous research (Surucu et al., 2020) that brand relationship quality mediates the positive effect of brand authenticity on brand usage intent.

The results of the empirical tests that had been analyzed by PLS-SEM, produced a model of research results as below (Figure 2).

Table 5: Hypothesis Test Results

Figure 2: Result Model

5. Conclusion

This study has tested three hypotheses with brand usage intent as the target construct. This study concludes that brand authenticity can influence brand usage intention through brand relation quality. The limitation of this study is that the data collection was carried out in a cross sectional manner that ignored fluctuations that might have occurred in the management of brand authenticity from time to time. Suggestions that can be given through this limitation are future research to examine the authenticity of local Indonesian fashion brands from time to time longitudinally.