Brand Image: As a Mediating Variable in the Relationship between E-WOM and Repurchase Intention of Teh Botol Sosro Products

research examines the influence of e-WOM and

non-probability sampling, data from 100 Tehbotol Sosro's consumers were analyzed with SmartPLS 3.0 software. The findings indicate that e-WOM positively and significantly affects repurchase intention and Brand Image. Brand Image also positively and significantly affects repurchase intention, acting as a mediator between e-WOM and repurchase intention.

INTRODUCTION
Based on the publication by the Ministry of Industry (2022) in the third quarter of 2022, the processed food and beverage industry recorded a growth of 3.57% (YoY). It became the subsector that contributed the most to the non-oil and gas manufacturing gross domestic product, amounting to 38.69%. In line with this, data from BPS(2022a) shows that the percentage of average monthly per capita consumption expenditure for processed food and beverages is the highest at 31.19% among the expenditure groups of food and beverages. This growth is supported by the increased consumption of packaged ready-to-drink tea, as the news portal tirto.id reported (2018), which collected data from the Food Consumption Statistics. Packaged tea is Indonesia's most consumed packaged beverage, with an average annual consumption percentage of 16.14%. As the capital of East Java Province and recognized as the second largest metropolitan city in Indonesia with a population of 3.157.126 people (Disnaker Surabaya, 2021), Surabaya City contributes the most to the processed food and beverage industry. According to BPS data (2022b), the average per capita expenditure in one month for the processed food and beverages group in Surabaya is 44.51%, the highest among other cities in East Java.  (2022) Although the industry has seen positive growth, this growth has yet to impact the Tehbotol Sosro brand positively. Data in Table 1. shows that the top brand index of Tehbotol Sosro has been declining since 2018 and has been replaced from the top position by Teh Pucuk Harum, a newcomer in this market. The measurement of the Top Brand Index consists of 3 parameters: brand awareness, last usage, and repurchase intention (Top Brand Award, 2022). Fluctuations in the Top Brand Index for a brand indicate significant changes in market dynamics, suggesting that the company needs to evaluate their strategies to compete more effectively.
According to the information from the Ministry of Communication and Information Technology's website (2019), there has been a circulating rumour among social media users regarding Teh Botol Sosro containing a harmful substance called Hydroxylic Acid. In reality, Hydroxylic Acid is another term for water or H2O. Rumours about harmful substances in products can harm Brand Image. Therefore, it is crucial for the company to promptly refute the false rumour and strengthen their communication about the safety and quality of their products, considering that Brand Image is one of the determining factors for consumer repurchase intention. With such rumours, users may lose trust in the brand and choose to seek alternative options, which can lead to negative valence from consumers, resulting in a decline in sales and market share for the brand brand (Prabowo et al., 2020).
Word of mouth (WOM) is one of the marketing communication mix elements, which involves oral, written, and electronic communication among customer communities to convey their experiences with consuming products and services (Kotler & Keller, 2018). The development of information technology has led to the evolution of Word of Mouth (WOM) into a new form of marketing communication known as Electronic Word of Mouth or e-WOM (Menendez et al., 2019). The current electronic communication activities are closely linked to the use of Google, one of the world's largest providers of information services. Google Trends is a service provided by Google to analyze the popularity level of the most popular search queries on Google Search in various regions and languages. This service utilizes graphs to compare the search query volume over time. Below is a graph for the search queries "Teh Pucuk Harum" and "Teh Botol Sosro" over the past 12 months: Tea" in 2022 in Indonesia Figure 1 shows that in Indonesia, during the year 2022, the search query "Teh Pucuk Harum" was more popular than the search query "Teh Botol Sosro." Similar results can also be seen on the official Twitter account of the Tehbotol Sosro brand; according to data from Social Blade (2023) the account created in September 2011 with the username @tehbotolsosroid currently has 81,878 followers and 12,954 tweets on Twitter. These numbers are lower compared to a competing brand that joined the Twitter platform in May 2011 with the username @pucukharumid, which currently has 91,275 followers and 29,096 tweets on Twitter.
The research conducted by Sari et al. (2021) indicates that Brand Image positively influences consumers' repurchase intention. Brand Image, or the perception of a brand, refers to how consumers view a product or company (Iswara & Santika, 2019), nd it is shaped by their previous experiences or information about the brand. According to the study by Santi & Suasana (2021) e-WOM refers to marketing communication through digital media, such as social media, which provides information to consumers about a product. The researchers also argue that e-WOM is one of the determining factors of consumers' repurchase intention. This argument is supported by other studies that also state that e-WOM positively influences repurchase intention (Arif, 2019). According to Arima et al. (2020) the more positive e-WOM a brand receives, the higher the consumer's repurchase intention for that product. Therefore, this disparity indicates the difference in opportunities for each brand to achieve the brand awareness necessary to build an image through such platforms.

THEORETICAL REVIEW Brand Equity
According to the American Marketing Association, a brand is a name, term, sign, symbol, design, or a combination of these, intended to identify the goods or services of one seller or a group of sellers and to differentiate them from competitors. Brand equity, on the other hand, refers to the value of a brand and can refer to two understandings of brand value: the strategic, subjective understanding of brand equity as a financial, objective expression of brand value (Heding et al., 2009, p. 9). According to Keller (1993), consumerbased brand equity can be measured in two parts: brand awareness and brand image.

Brand Image
According to Peter & Olson (2010, p. 45), Brand Image refers to consumers' perceptions and preferences towards a brand that are stored in their memory and manifest in various brand-related associations. Another viewpoint suggests that Brand Image is the public perception of a brand (Mirzai et al., 2016). Brand Image's goal is to shape brand differentiation, trust, security, exclusivity, and purchase decisions (Aaker, 1991, p. 144). Companies need to maintain their Brand Image to reduce consumer doubts and provide information that aids consumers in their purchase decisions (Erdem et al., 2002). Ambolau et al. (2015) proposed indicators that can be used to measure Brand Image, including: 1) Corporate Image, 2) User Image, and 3) Product Image. e-WOM (Electronic Word of Mouth) e-WOM, or electronic Word of Mouth, is the evaluation and recommendation of a product or company created by potential customers, current customers, and previous customers in text-based form, available to many consumers through written content on the internet (Weitzl, 2017, p. 2). Goyette et al. (2010) proposed indicators that can be used to measure e-WOM, which include: 1) e-WOM Intensity, 2) Valence of Opinion, and 3) Content.

Repurchase Intention
According to Solomon (2020, p. 340), purchasing decision results from a series of stages that lead to selecting one product over competing options. Solomon also adds that consumers will integrate as much information as possible about a product, weigh the pros and cons of each alternative, and make a decision that satisfies them. The process consists of problem recognition, information search, alternative evaluation, product selection, and post-purchase evaluation.
Repurchase intention is the consumer's attitude in the post-purchase evaluation phase. In his book, Solomon adds that marketers should study the steps of consumer decision-making. With that understanding, they can develop products and promotional strategies that fulfil the specific information needs of consumers in the most effective format (Solomon, 2020, p. 345). Another opinion states that repurchase intention is an individual's evaluation of repurchasing services from the same company, considering the current situation and the possibility of the circumstances (Balla et al., 2015). Furthermore, Balla et al. add that repurchase intention is an appropriate dependent variable for developing insights into strategic planning and service delivery. Ginting et al. (2023) propose indicators that can be used to measure repurchase intention: 1) Transactional Interest, 2) Referential Interest, 3) Preferential Interest, 4) Explorative Interest.
Effect of e-WOM on Repurchase Intention Goyette (2010, p. 6) argues that e-WOM is a post-purchase behaviour that involves spontaneous conversations between two individuals based on satisfying or dissatisfying purchase experiences. Measurement of e-WOM can be used to predict consumer repurchase intention and the tendency to praise a company. Based on the findings of Rahman et al. (2020), there is a significant positive relationship between e-WOM and repurchase intention. Other studies with similar results add that the better the e-WOM consumers obtain, the higher their repurchase intention in e-commerce. Negative e-WOM can decrease consumers' repurchase intention in e-commerce (Ginting et al., 2023). Based on several previous studies, it can be concluded that e-WOM can influence repurchase intention.
H1: e-WOM has a positive and significant influence on repurchase intention for Tehbotol Sosro products.
Effect of e-WOM on Brand Image Ismagilova et al. (2017, p. 21) argue that through e-WOM communication, consumers can receive information about products and services from a wide and geographically dispersed group of people. Additionally, e-WOM can serve as an alternative source for consumers to obtain information about products and services based on peer-to-peer recommendations, reducing the influence of companies compared to traditional marketing and advertising. Ismagilova et al. also adds that e-WOM can function as an effective tool for building Brand Image and acquiring and retaining customers. This statement is supported by research findings that show a positive and significant impact of e-WOM on Brand Image (Adriana et al., 2022;Lestari & Gunawan, 2021;Yohana et al., 2020). Therefore, it can be concluded that e-WOM influences Brand Image.
H2: e-WOM has a positive and significant impact on Brand Image for Tehbotol Sosro products.

Effect of Brand Image on Repurchase Intention
According to Kotler (2018, p. 32), companies address customer needs by emphasizing the value proposition, a set of benefits that meet those needs. A brand is an offer from a known source. The brand name carries various types of associations in people's minds, forming the Brand Image. Kotler also adds that to increase repeat purchase behaviour, companies must build a Brand Image with strong, beneficial, and unique brand associations as much as possible. Based on the research findings of Shihab et al. (2022) Brand Image has a positive and significant impact on repeat purchase behaviour. Shihab et al. also suggest that a brand with a positive perceived Brand Image can increase consumers' desire or intention to engage or transact, specifically to make repeat brand purchases. Similar findings were shown in the research conducted by Huang et al. (2019), indicating that Brand Image positively and significantly impacts repeat purchase behaviour. Therefore, it can be concluded that Brand Image influences repeat purchase behaviour.
H3: Brand Image has a positive and significant influence on repurchase intention for Tehbotol Sosro products.

Influence of Brand Image in Mediating the Relationship between e-WOM and Repurchase Intention.
According to Ismagilova et al. (2017, p. 13), even though e-WOM is more effective than traditional mass media communication, managers should still consider traditional marketing communication methods. Furthermore, Ismagilova et al. added that communication created by companies, such as advertising and Brand Image, can mediate the influence of e-WOM, including its impact on repurchase intention. This statement is supported by research findings that demonstrate the positive and significant mediating role of Brand Image in the relationship between e-WOM and repurchase intention (Gunawan & Wayan, 2022;Iswara & Santika, 2019;Sari et al., 2021). Therefore, it can be concluded that Brand Image mediates the relationship between e-WOM and repurchase intention.
H4: Brand Image has a positive and significant mediating effect on the relationship between e-WOM and Repurchase Intention of Tehbotol Sosro products

METHODOLOGY
This research employed a quantitative method, utilizing a non-probability sampling technique called purposive sampling. The population of this study consisted of consumers who had purchased Tehbotol Sosro products in Surabaya. The researcher did not know the exact size of the population, so the sample determination technique was based on Ferdinand's (2014), sample measurement method, which involves multiplying the number of indicators by 5-10. This study utilized 10 indicators with a measurement parameter of 10. The sample size for this research was 100 respondents who were at least 17 years old, had social media accounts, and had purchased Tehbotol Sosro products in Surabaya. Secondary data were collected from various sources such as articles, journals, websites, and social media related to the research problem. The data were analyzed using Structural Equation Modeling -Partial Least Squares (SEM-PLS) technique with the SmartPLS 3.0 software.  (2023) Based on the table above, the validity of the indicators is measured by looking at the Factor Loading values from the variables to their indicators. It is considered valid if the Factor Loading is greater than 0.5 or the T-Statistic value is greater than 1.96 (the Z value at α = 0.05). Factor Loading represents the correlation between an indicator and a variable. If it is greater than 0.5, it meets the validity criteria. Similarly, the significance is met if the T-Statistic value is greater than 1.96.

Outer Model Testing (Indicator Validity)
Based on the outer loading table above, all reflective indicators for the variables e-WOM, Brand Image, and Re-Purchase Intention show factor loadings (Original sample) greater than 0.50 or significance (T-Statistic value greater than the Z value α = 0.05 (5%) = 1.96). Therefore, the estimated results for all indicators meet Convergent Validity or have good validity. The table also shows that the most significant indicator for the e-WOM variable is Valence of Opinion (X1.2), the most significant indicator for the Brand Image variable is Product Image (Y1.3), and the most significant indicator for the Re-Purchase Intention variable is Preferential Interest (Z1.2).
Source: Processed Primary Data (2023) Figure 3. Outer Model (Factor Loading, Path Coefficient and R-Square)

Reliability Test
The results of the composite reliability testing show that the e-WOM variable (X) has a value of 0.874, the Brand Image variable (Y) has a value of 0.789, and the Purchase Intention variable (Z) has a value of 0.877. All three variables show composite reliability values above 0.70, indicating that all variables in this study are reliable.

R Square
The R-Square value for Brand Image (Y) is 0.259. This value can be interpreted as the model's ability to explain 25.9% of the variation in Brand Image, which is influenced by the independent variable, e-WOM (X). In comparison, the remaining 74.1% is influenced by other variables not included in this study. Additionally, the Repurchase Intention (Z) value is 0.608. This value indicates that the model can explain 60.8% of the variation in Purchase Intention, influenced by the independent and intervening variables (e-WOM and Brand Image). In comparison, the remaining 39.2% is influenced by other variables not included in this study.  Table 4, it can be concluded that the hypotheses stating: 1) Brand Image has a positive effect on Re-Purchase Intention: This hypothesis is supported with a path coefficient of 0.584 and a T-Statistic value of 8.572, which is greater than the critical value of Z α = 0.05 (5%) = 1.96. Therefore, it is considered significant (positive). 2) e-WOM has a positive effect on Brand Image: This hypothesis is supported with a path coefficient of 0.509 and a T-Statistic value of 5.868, which is greater than the critical value of Z α = 0.05 (5%) = 1.96. Therefore, it is considered significant (positive). 3) e-WOM has a positive effect on Re-Purchase Intention: This hypothesis is supported with a path coefficient of 0.299 and a T-Statistic value of 3.716, which is greater than the critical value of Z α = 0.05 (5%) = 1.96. Therefore, it is considered significant (positive).  Table 5, it can be concluded that the hypothesis stating: 4) e-WOM has a positive effect on Re-Purchase Intention through Brand Image as a mediating variable: This hypothesis is supported with a path coefficient of 0.297 and a T-Statistic value of 5.398, which is greater than the critical value of Z α = 0.05 (5%) = 1.96. Therefore, it is considered significant (positive).

Effect of e-WOM on Repurchase Intention
Based on the research findings, it is concluded that e-WOM has a positive and significant effect on the Repurchase Intention of Tehbotol Sosro products in Surabaya; therefore, the hypothesis is accepted. The results of the descriptive analysis of the e-WOM variable indicate that the most significant indicator of this variable is the Valence of Opinion. On the other hand, the most significant indicator for the Repurchase Intention variable is Preferential Interest, indicating that consumers consider Tehbotol Sosro as their preferred choice due to the product's quality assurance and positive discussions about its taste on social media.
This indicates that positive testimonials or reviews from other users have a significant influence on consumers' purchasing decisions regarding Tehbotol Sosro products. These findings are in line with previous studies (Nathalia & Indriyanti, 2022;Nuraeni & Manoarfa, 2022;Putra et al., 2020;Wandoko & Panggati, 2022) that concluded that e-WOM has a positive effect on repurchase intention.

Effect of e-WOM on Brand Image
Based on the research findings, it is concluded that e-WOM has a positive and significant impact on the Brand Image of Tehbotol Sosro products in Surabaya, thus supporting the hypothesis. The results of the descriptive analysis of the e-WOM variable indicate that the most significant indicator is the Valence of Opinion. In contrast, Product Image is the most significant indicator for the Brand Image variable. This means that Tehbotol Sosro can provide a positive experience to its consumers through a favourable portrayal of the product, such as its glass bottle packaging that allows consumers to see the contents directly and its unique tagline "apapun makanannya, minumnya Tehbotol Sosro" (whatever the food, drink Tehbotol Sosro). These factors have led to positive consumer reviews, resulting in an enhanced Brand Image of Tehbotol Sosro.
This indicates that the more consumers provide positive opinions about a product, the better the Brand Image of that product is perceived by consumers. These findings are consistent with research studies (Alrwashdeh et al., 2019;Mihardjo et al., 2019;Nuseir, 2019;Siddiqui et al., 2021) which concluded that e-WOM has a positive and significant impact on Brand Image.

Effect of Brand Image on Repurchase Intention
Based on the research findings, it was found that Brand Image has a positive and significant influence on the Repurchase Intention of Tehbotol Sosro products in Surabaya. Therefore, the hypothesis is accepted. The results of the descriptive analysis of the Brand Image variable indicate that the most significant indicator is Product Image. In contrast, the most significant indicator of the Repurchase Intention variable is Preferential Interest. This means that consumers prefer to consume Tehbotol Sosro compared to other tea brands due to the positive portrayal of Tehbotol Sosro products, such as its easily recognizable logo, distinctive tea aroma, and the use of natural ingredients.
This indicates that the better the Brand Image of Tehbotol Sosro products in the eyes of consumers, the higher their intention to repurchase the product in the future. These findings are consistent with previous research (Cuong, 2022;Putri & Yasa, 2022;Shihab et al., 2022;Tunjungsari et al., 2020) which concluded that Brand Image has a positive and significant impact on Repurchase Intention.

Role of Brand Image in Mediating The Effect of e-WOM Towards Repurchase Intention
Based on the research findings, it is concluded that Brand Image mediates the relationship between e-WOM and Repurchase Intention, therefore supporting the hypothesis. Brand Image in this study acts as a partial mediator (Partial Mediation). The descriptive analysis of the Brand Image variable indicates that the most influential indicator is Product Image. The most significant indicator for e-WOM is the Valence of Opinion, while the most influential indicator for Repurchase Intention is Preferential Interest. This suggests that the strengths of Tehbotol Sosro, such as its guaranteed tea quality, distinctive tea aroma, use of natural ingredients, attractive packaging, recognizable logo, and unique tagline, are being discussed by consumers on social media, leading to their interest in repurchasing Tehbotol Sosro over other tea brands.
This means that the more favorable the consumers' perception of the product's brand image, the more positive reviews from consumers, which can increase their intention to repurchase the product in the future. These findings are consistent with studies (Aldrin Akbar et al., 2019;Candra & Suparna, 2019;Hendro & Keni, 2020) that conclude that Brand Image has a positive and significant influence in mediating the relationship between e-WOM and Repurchase Intention.

CONCLUSIONS AND RECOMMENDATIONS
Based on the analysis and discussion in the previous chapter, the following conclusions can be drawn: 1) e-WOM has a positive and significant influence on Repurchase Intention in Surabaya. It can be concluded that the higher the positive reviews from consumers about Tehbotol Sosro, the higher the consumer's intention to repurchase the product in the future. 2) e-WOM positively and significantly influences Brand Image in Surabaya. It can be concluded that the higher the positive reviews from consumers about Tehbotol Sosro, the better the product's Brand Image. 3) Brand Image positively and significantly influences Repurchase Intention in Surabaya. It can be concluded that the better the Brand Image of Tehbotol Sosro in the eyes of consumers, the higher their intention to repurchase the product in the future. 4) Brand Image has a positive and significant mediating effect on the relationship between e-WOM and Repurchase Intention in Surabaya. This is because the respondents' intention to repurchase Tehbotol Sosro increases due to positive reviews on social media about the excellence and guaranteed quality of the tea. With the increase in the product's Brand Image in the eyes of consumers through positive reviews, the intention to repurchase the product in the future also increases.
Based on the above conclusions, there is recommendation can be considered or beneficial in making decisions Tehbotol Sosro should maintain product quality and develop innovative products to strengthen the Brand Image. Additionally, the company can optimize marketing through social media by monitoring e-WOM related to Tehbotol Sosro products. This will help obtain real-time feedback and customer responses, enabling quick and effective responses to strengthen Brand Image and enhance repurchase intention.

FURTHER STUDY
For future research, researchers can consider adding other variables influencing repurchase intention, such as customer satisfaction, price, and promotion. They can also expand the scope of the study to a wider geographical area, use a larger sample size to improve the generalizability of the results and employ more complex data analysis methods.