The Moderating Role of Electronic Word of Mouth (eWOM) in the Influence of Perceived Value on Repurchase Intention

The purpose of this research is to determine whether word of mouth which has developed into electronic (eWOM) in the digitalization era can strengthen or weaken the influence of the perceived value received by customers from previous purchases to repurchase or not. This study uses a quantitative approach, whereas the method of data collection was purposive sampling and has determined as many as 116 samples. The population in this research are followers of Shopee Indonesia in Instagram, with the various characteristics that have been determined by researchers to be taken as a sample for this research. The method of data analysis using equations from Cohen et al. (1975) in the journal of Sharma et al. (1981). The regression analysis is based on the researcher's comprehension of the negligibility correlation between moderation and predictor variables. Based on the results, Perceived Value has a significant and positive effect on Repurchase Intention, and the regression model analysis was found that the inclusion of the moderation variable, strengthened the influence of Perceived Value on Repurchase Intention by 61,5% from the previous results (55,8%).


INTRODUCTION
Business is an option for the general public to get profits. All people can do business, ranging from opening businesses individually to building a company by an organization. Business is an organized activity to generate funds, sell goods or services to benefit in meeting the needs of the community (Suwarso, 2018).
The existence of technological advances in this day also contributes to changing business strategies to rely on technology from the internet, so that the term "e-commerce" is now familiar to the public. According to Cashman (2007), e-commerce, or short for electronic commerce, is a business transaction that occurs in an electronic network such as the internet. British research institute Merchant Machine has released a list of the 10 fastest growing e-commerce countries in the world.

Figure 1. The 10 Fastest Growing E-commerce Countries in the World
Source: m.databoks.katadata.co.id (2019) Based on figure 1, countries located in the Asian continent dominate the top 10 countries with the fastest e-commerce growth in the world, with Indonesia as the country with the highest growth compared to other countries, so it can be said that Indonesia becomes a "fertile field" for e-commerce companies or owners from abroad to expand its market in Indonesia.
Shopee is a Singapore-based electronic trading platform founded in 2009 by Forrest Li, Shopee first launched in Singapore in 2015 and has expanded to Malaysia, Thailand, Taiwan, Indonesia, Vietnam, and the Philippines. (id.wikipedia.org/wiki/Shopee), Shopee is one of the e-commerce that is widely known by the people of Indonesia because Shopee has several times used famous figures as its brand ambassadors. For instance, famous footballer Cristiano Ronaldo, and Indonesian comedian Tukul Arwana. Even more, Shopee is the main sponsor in the highest league in football in Indonesia, Liga 1 2019 being the first sports tournament fully supported by Shopee. The change of sponsors also make the 2019 league logo transformed into a typical Shopee color, orange (Harahap, 2019). Shopee Indonesia has its famous tagline which is "Free ongkir se-Indonesia" (Free delivery around Indonesia). However, to get free delivery, buyers must meet the requirements that are quite difficult to enjoy the free delivery, and also every month the free delivery policy is always changing, so this can make the expectations obtained by consumers different from what expected, and can also give rise to a negative form of electronic word of mouth (eWOM) and spread through social media. Not only that, the form of fraud that occurs can also affect consumer perception, the Indonesia Financial Services Authority (OJK) recorded the number of public complaints related to online shopping services or online shops reached tens of thousands. Complaints that come in mostly regarding shopping items are not sent after payment (Liputan6.com, 2019).
Social media or social networking becomes one of the platforms for the public or consumers to find the information that they need. According to Evans (2010), social media is a form of democracy from information, social media can change the society that was previously a reader into a publisher. So that social media can be a tool for someone to give their opinion on something through social media until finally, it can influence others (persuasive).

THEORETICAL REVIEW Perceived Value
According to Hutt & Speh (2007), perceived value is defined as the perception and evaluation of customers about how beneficial the relationship with suppliers is and the benefits or profits and sacrifices provided. Consumer perception of value can be defined in several senses, namely, value is low price, value is whatever consumers want from the product, value is the quality obtained for price and value is what is obtained for what has been given (Rahab et al., 2015). The dimensions of perceived value itself consist of four main aspects (Perwira et al., 2016). 1. Emotional Value, a situation where customers are satisfied with the overall benefits of a product or service. In this study, it can be seen from 2 aspects, consisting of: a) Personal desire to use Shopee (X1.1). b) Convenience in shopping by using Shopee (X1.2). 2. Social Value, a situation in which consumers feel the social level increases when using products or services, consists of: a) The impression is good at using Shopee (X2.1). b) Pride in using Shopee (X2.2). 3. Performance, which is the assessment of consumers when expected expectations are in accordance with what has been obtained, consists of: a) Good Shopee services quality (X3.1). b) Design and interesting Shopee features (X3.2). c) Shopee can be used to buy various kinds of consumer needs (X3.3). 4. Price or Value of Money, which is the customer's assessment of the level of cost efficiency that can be met, consists of: a) Shopee often holds seasonal discounts for the products they sell (X4.1).
b) The price of the product purchased through Shopee in accordance with the quality of the product (X4.2).

Word of Mouth Marketing
According to Hasan (2010), word of mouth marketing (WOM) is a conversation designed online and offline that has multiple effects, non-hierarchy, horizontal and mutational. Such a strong reason in WOM is a reciprocal conversation that can't be found with hundreds of other messages in the company's conventional folder. The basic philosophy of word of mouth marketing according to Hasan (2010) is: 1. Sustainability of customer's voice, not the voice of the company, owner, or marketer. 2. The natural, the original, honest process is not artificial, and also manipulation 3. Consumers look for sources of information, not companies, owners, or marketers. 4. Consumers talk about a product, service, or brand, and they have had experience.

Electronic Word of Mouth (eWOM)
The definition of word of mouth marketing is an action that can provide a reason for everyone to be easier and prefer to talk about your product (Andy, 2006). According to Chatterjee (2001), the increasing use of the internet and social networks is also an important thing where today WoM is not only done by individuals but can be in any form including the internet called electronic word of mouth (eWOM). According to Bataineh (2015), there are several indicators that can be used to measure eWOM: 1. eWOM credibility, refers to the credibility of the source that is one of the factors of the influence of eWOM.Others as recipients will easily trust the information if the source of the information has credibility, which can affect the direction of decision making.EWOM credibility indicators are: a) Social trust (X2.1.1). b) Information source (X2.1.2). c) The number of reviews (X2.1.3). 2. eWOM quality, referring to the persuasive power in the purchase of quality and generally acceptable information, includes: a) Contents of information according to consumer needs (X2.2.1). b) Contents of information are very helpful and easy to understand. (X2.2.2). 3. eWOM quantity, referring to the number of reviews or comments about an object can measure the popularity and market performance of the object.

Repurchase Intention
Repurchase intention is the individual's judgment about buying again a designated service from the same company, taking into account his or her current situation and likely circumstance (Hellier et al. 2003), it can be interpreted briefly that repurchase intention is a person's decision to buy back service through the same company. According to Riptiono (2013), repurchase intention is the possibility of buyers with the intensity to repurchase or buy back a product. Indicators of repurchase intention according to Unjaya & Lilyana Santoso (2015) are: 1. Resistance against the better alternative, which is an attitude from customers not to switch to other products. a) Customers will still choose Shopee even though there are other e-commerce services. (Y1.1). b) Customers are interested in reusing services from Shopee (Y1.2). 2. Willingness to pay a premium price is the willingness of a customer of a product or service to pay more expensive to get better service or facilities without feeling dissatisfied. Consists of: a) Customers still choose Shopee even though the price of products in other ecommerce is cheaper (Y2.1). b) Customers are willing to create a personal account in Shopee (Y2.2). 3. The intention of word of mouth, which is a consistency of customers of a product or service to keep dissatisfaction with a product or service when experiencing a disappointment and notify the management of a company providing products or services when experiencing dissatisfaction in a product or service. a) Customers do not need to inform others when experiencing service problems at Shopee (Y3.1). b) Customers need to notify Shopee when experiencing service problems (Y3.2).
Researchers look at previous studies that have research variables similar to this study to help researchers to form conceptual frameworks and research hypotheses. Some previous studies that were used as literature reviews can be seen from the table below.

METHODOLOGY
This research is explanatory research, with a quantitative approach. Determination of the number of samples using the formula from Machin and Campbell (Machin & Campbell, 1987), with the results of 116 people to be sampled. Data collection technique by spreading questionnaire online to respondents. The location of the study is in Indonesia and the respondent is a follower of the Shopee Indonesia Instagram account and which is included in the characteristics that have been determined by researchers to be taken as a sample, some of the respondent characteristics are minimum age 17 years, make purchases at Shopee at least 2 times, and get the information about Shopee from electronic media. Researchers use online media to collect data and information needed, namely through social media using Google Forms.

Descriptive Statistical Analysis
This analysis is used to find out the picture of data characteristics, compile and present research data in the form of frequency distribution tables obtained from the authors of the questionnaire through Google Forms. After getting data from 116 respondents, it is further to know the average score. Each item and indicator will be grouped in the interval magnitude of the Likert Scale score which is hung by the following formula:   (2008) Based on table 2 it will be known the average results of all answers from 116 respondents to each item on the indicator for each variable so that the score results will later be used to provide suggestions on the conclusion section to be interpreted referring to in table 2. C = − 1

t-Test and Interaction Test
The t-test analysis is done to determine the effect of X variables against Y variable before the presence of the moderator variable (X2). This analysis is based on the researcher's comprehension of the negligibility correlation between moderation and predictor. The variables Interaction regression analysis is carried out with the multiplication between moderator variable (X2) and independent variable (X), this aims to find out how the effect of the moderator variable in the results of the relationship between X variable with Y variable with multiplication through two formula equations as follows:

RESULTS AND DISCUSSIONS Results of Statistical Analysis & Distribution of Respondent Answers
The results of the average questionnaire answer of 116 respondents who can be seen from appendix 1 will be interpreted through table 2. The result shows that in variable X (Perceived Value), "personal desire to use Shopee (X1.1)" gets the highest score (4.24), and "pride in using Shopee (X2.2)" get the lowest value (3.93).
The highest item (4.17) on variable Y (repurchase intention) are on the "customers are interested in reusing services from Shopee (Y1.2)", and "customers need to notify Shopee when experience service problems (Y3.2) whereas for the lowest item score (3.66) is in "customers do not need to inform others when experiencing service problems at Shopee (Y3.1)".

Residual Normality Test & One-Sample Kolmogorov-Smirnov Results
Normality testing is done using a P-P graphic testing plot.

Multicollinearity Test Results
This study conducts multicollinearity testing using the variance inflation factor (VIF) method. If the value of VIF> 10, the results of the request indicate multicollinearity. Conversely, if VIF<10, it shows the absence of multicollinearity.  (2021) Based on Table 4, it is found that in model 2, VIF value <10 with tolerance value >0.1, meaning between independent variables there is no strong correlation or there is no multicollinearity (fulfilled assumption).

Heteroskedasticity Test Results
The method for testing the heteroskedasticity in this study is by looking at the plot graph between the value of Y variable predictions with the residual. If the points have formed a particular pattern that is regular, then indicating a heteroskedasticity has occurred. Conversely, if there is no clear pattern, or the points spread above and below the number 0 on the Y axis, then there is no heteroskedasticity.  figure 6, it is known that the points are scattered randomly (not patterned), both the above and those below the number 0 on the Y axis, which means that the assumption of heteroskedasticity has been fulfilled in the model 1 and 2 (variance residual homogeneous).

t-Test (Equation 1)
The t-test is a regression equation that calculates and explains how the effect power of the X (perceived value) variable on the Y (repurchase intention) variable.

Y = 4.509 + 0.527 X + εi
From the regression results, it was found that the X (perceived value) variable has a positive and significant effect on the Y (repurchase intention) variable, which can be seen from the value | tcount | is greater than ttable (11,988> 1,981), it can also be seen from the p-value that is smaller than α (0.000 <0.050). Positive regression coefficients explain that the increase in X (perceived value) variable will be significantly able to increase the Y (repurchase intention) variable.

Interaction Test (Equation 2)
Interaction test is a regression equation that explains the effect of X (perceived value) variable, and the interaction of X*X2 on Y (repurchase intention) variable. The results of the regression are shown in table 6 below: The regression model based on the results of the analysis of the table 6 is: Y = 9.322 + 0.164 X + 0.008 X*X2 + εi From the results of the interaction test, it can be known that the X*X2 interaction variable has a positive and significant effect on the Y variable (repurchase intention), which can be seen from the value |tcount| which is greater than the ttable (4,112 <1,981), or p-value values smaller than α (0,000> 0,050). This means that the entry of the X2 variable (electronic word of mouth) as pure moderation, can strengthen the effect between the X variable on the Y variable.

Coefficient of Determination t Test (Equation 1)
Based on table 5, it is known that the value of R Square (R 2 ) t-test or equation 1 is 0.558 (55.8%), which means that the Y (repurchase intention) variable is explained by 55.8% by the X (perceived value) variable. Whereas the remainder of 44.2% is explained by other variables or factors outside of the regression equation.

Interaction Test (Equation 2)
Based on table 6, an R Square (R 2 ) value was found in the interaction test of 0.615 (61.5%), which means that the Y (repurchase intention) variable is explained by 61.5% by independent variables plus moderator variable or independent variable interactions with moderator variable (X and X*X2). Whereas the remainder of 38.5% is explained by other variables outside the regression equation.

CONCLUSION
1. The X (perceived value) independent variable has a positive and significant effect on the Y (repurchase intention) dependent variable, seen in equation 1 with a p-value t-test of 0.000 smaller than 0.05. 2. The effect of the X2 (electronic word of mouth) moderator variable on the results of the X (perceived value) independent variable on the Y (repurchase intention) variable is positive. The regression coefficient that shows a positive value explains that an increase in the variable interaction X*X2 will increase the Y (repurchase intention) variable. Increased results on the previous R Square of 55.8% to 61.5% showed an increase in the effect of the X (perceived value) variable on the Y (repurchase intention) variable which was moderated by the X2 (electronic word of mouth) variable.

Suggestion
1. Theoretical Aspect, researchers suggest if want to do further research, use the latest theories or maybe theories that will be found in the future to follow the development of electronic marketing technology, some theories such as the electronic lifestyle, electronic service quality, and electronic trust can be used to find out Its influence on the electronic word of mouth and electronic repurchase intention. Further research that uses moderator variables should use Moderated Regression Analysis (MRA) and is done by using a source of formula that has a strong basis, namely using the MRA analysis test recommended by Sharma et al. (1981). Like the basis that refers to the research that has been done before, determining whether or not the relationship of the moderator variable to the bound variable at the beginning of the study is very important before entering into the collection of samples and data analysis.

Practical Aspects a. Perceived Value (X)
Most of Shopee consumers in this study get or have a good assessment of Shopee, especially for Shopee in Indonesia, so they use Shopee based on their own personal desires without being influenced by others, and this must be maintained by the shopee to continue to give what It is expected that consumers when consumers use services from Shopee, and Shopee in Indonesia must look for the latest breakthroughs that are not owned by ecommerce competitors so that there is a characteristic of the shopee itself which can be proud of by Shopee users. b. Repurchase Intention (Y).
Shopee should immediately be able to overcome any complaints or requests for assistance entering from consumers due to customers consider the service issues they meet can be finished more quickly if contacting the shopee itself, without the help of outside parties, and this must be maintained by Shopee. Shopee customers continue to prioritize the price factor for the purchase decision in Shopee, if consumers find cheaper prices for the same product in other e-commerce, most likely consumers will switch to other e-commerce, so that Shopee must be careful in terms of competition With its competitors, especially in terms of competitive prices. c. Electronic Word of Mouth (eWOM) (X2).
Shopee Indonesia should prioritize more often to convey the information through more credible platforms or sites, for example through official media accounts, notifications through applications or websites, and through official e-mail.