ANTECEDENTS OF E-LOYALTY AS RESEARCH FOR THE QUICK COMMERCE INDUSTRY

ABSTRACT


INTRODUCTION
Globally, sales growth in e-commerce increased significantly from 2014 to 2021. In 2021, based on emarketer.com retail sales in e-commerce reached 4.9 trillion US dollars worldwide. This figure is expected to continue to grow by 50% over the next four years, reaching around 7.4 trillion dollars by 2025. The COVID-19 pandemic has also encouraged new customers to make purchases online for the first time. Due to social distancing measures, more and more customers are becoming aware of the convenience of online shopping and reducing their concerns about security and privacy. For example, after the first case was announced, the download rate of the Getir mobile app increased to 60%, the number of users reached 2 million consumers and the age of users rose to 70 which was previously between 18 and 45 before the pandemic in Turkey (Milliyet, 2020). In addition, many traditional grocery retailers started door-to-door fast delivery during the pandemic. Delivery Hero also reported a 400% year-over-year growth in q-commerce orders in Q1 2021 with over 400,000 q-commerce orders processed on average per day in April 2021 (Delivery Hero, 2021). As Milena Lazarevska, Vice President Commercial Quick Commerce of Delivery Hero puts it, "These services will soon become an intrinsic part of people's lives, making them wonder what they did before these services existed" (Delivery Hero, 2021).
In its implementation, it was found that customer satisfaction is not always an intermediary. E-Loyalty is formed from the quality of service followed by the close relationship between customer satisfaction and consumer trust. That is, satisfaction cannot stand alone and be influenced by others. In 2020, a study on customer engagement revealed that the subject did not positively affect customer loyalty. Customer experience and engagement can combine to form customer loyalty through satisfaction. Supporting the previous statement, perceived trust and perceived risk also directly shape customer loyalty (Kasri et al., 2021).
Although customer loyalty can be influenced by satisfaction, attitudes towards brands can be an intermediary to form customer loyalty (Smith, 2020). The perception of value and service quality will increase customer loyalty in positive relationships without passing customer satisfaction (Rizan et al., 2020). Surprisingly, e-commerce drivers shaped by care, character, choice, convenience, and customization have a positive impact on customer loyalty compared to communities and contact interactions that have a negative relationship and have no impact by not supporting customer loyalty (Alahmad et al., 2021). A 2020 study by Jeon et al. (2021) showed that online trust is statistically positively affected by offline trust and offline trust has a positive and significant effect on online satisfaction.
The occurrence of inconsistent relationships from the discussion of the literature above provides opportunities for gap research that can be investigated. Further research is also needed on specific mechanisms, where the effects of the formation of customer E-Loyalty occur and the limits of conditions where the dimensions of E-Service quality, perceived value and customer satisfaction increase customer E-Loyalty (Rizan et al., 2020). In response to this challenge for further research, we further explored the role of the E-Service quality dimension. The quality of E-Service with significant dimensions of safety, reliability, comfort, and responsiveness has a major influence on the quality of electronic services (Olaleye et al., 2021). The purpose of this study is to analyze the impact of E-Service quality services on customer satisfaction and E-Loyalty of q-commerce customers in Indonesia.
The hypotheses used in this study are: 1) H1: Customer satisfaction has an influence on customer loyalty; 2) H2: Service quality (Safety, reliability, comfort, responsiveness) has an influence on customer satisfaction; 3) H3: Service quality (Safety, reliability, comfort, responsiveness) has an influence on customer loyalty through customer satisfaction as a mediator; 4) H4: The value received has an influence on customer satisfaction; and 5) H5: The value received has an influence on customer loyalty through customer satisfaction as a mediator.

METHOD
In this study researchers collected data using survey methods (Quantitative) through electronic or online technology without intervention, besides that the research is regulated naturally. The unit in the analysis of this study is individual because q-commerce uses a business to customer (B2C) model, where the users are individuals, so it is more suitable to be used as a unit of analysis in this study. This research designs were descriptive and causal types (Malhotra, 2010). The researchers tried to discover the condition and behavior of customers at the time of the research so that data is taken at one time.

Figure 3. Research Flow
The target population in this study is customers who shop for basic/daily necessities online in urban areas (DKI Jakarta) and suburbs (Jabodetabek area) launched q-commerce. The parameters investigated were the value received, quality of E-Service, satisfaction and E-Loyalty of q-commerce application users. The sampling frame used in this study is application user information on the three largest q-commerce sites in Jabodetabek. The sampling method was non-probability sampling through the selection of two regions from five urban areas in Indonesia. Based on calculations with a sample determination formula when the population is unknown using the Cochran formula (Sugiyono, 2019). Based on this formula, 165 samples were obtained.
The structured questionnaire was designed to collect information from respondents related to socioeconomic factors, awareness of online purchases with benefits offered, and E-Service quality parameters in meeting consumer satisfaction. Closed-ended and open-ended questions were also included in the survey, which resulted in short responses from respondents. The survey questions are based on previous research that has been proven to have reliability and validity; among the survey scales used include the E-Service quality dimension (Blut, 2016;Holloway & Beatty, 2008;Parasuraman et al., 2005;Srinivasan et al., 2002), Accepted Values (Gallarza et al., 2019;Konuk, 2019;Sweeney et al., 1999), E-Loyalty (Gremler, 1995;Zeithaml et al., 1996), and customer satisfaction (Adeyeye et al., 2019; L. C. Harris & Goode, 2004;M. A. Harris et al., 2016). The reserach used a seven-point Likert scale (Ordinal scale type) in the question list to obtain more precise data and avoid bias. Researchers assure respondents of a guaranteed level of anonymity and confidentiality of the information provided. 400 surveys were distributed in the form of electronic questionnaires to users of the Q-Commerce application.
The primary data in this study relied on the results of questionnaires that collected significant data that could show hypotheses in the research model. Research questionnaires are distributed online using google form media to collect and manage answers from respondents.
Secondary data in this study was obtained by collecting important information from relevant journals, textbooks, and articles collected, one of which was by accessing Google Scholar. The secondary data collection process is useful to support the information obtained from the primary data and support a better framing process in the current research than the research that has been done.
To check the goodness of the data, a validity test was carried out using the pearson product moment correlation technique and for a reliability test using the Cronbach alpha technique with a Smart Partial Least Score (PLS). Analysis of research data using Structural Equation Model (SEM) using Smart PLS. Variable dimension analysis using Smart PLS. Researchers use Smart PLS software where the software offers an analysis process that is easier to use. The purpose of using Smart PLS among others is to predict relationships between constructs, confirm theories and can be used to explain the presence or absence of relationships between latent variables.

RESULTS AND DISCUSSION Measurement Model
As a measure for the discriminant validity of constructs, correlation values between constructs are compared with the square root of the AVE for each construct using the Fornell-Larcker criterion (1981). With the exception of the second order construct, Table 1 illustrates the square root of the AVE construct in bold, as shown on the diagonal, which is greater than the inter-construct correlation for each construct in the measurement model (e-quality of service). The square root of AVE for this construct is 0.554, but the correlation between constructs between ESQ and SEC is greater (0.731).
With the exception of the second order construct (ESQ), the discriminant validity of all constructs in the model was validated using the Fornell-Larcker (1981) criterion. Meanwhile, Henseler et al. (2015), among other methodologists, have recently questioned the reliability of the Fornell-Lacker criterion in identifying and detecting the presence or lack of discriminant validity in very common empirical scenarios. In addition to giving criticism, they not only suggested the Heterotrait-Monotrait correlation ratio (HTMT) as a new alternative methodology for detecting discriminant validity, but they also performed Monte-Carlo simulations to determine the superiority of the HTMT method over the Fornell-Larcker approach. The HTMT approach was used in this investigation, with the ESQ construction not meeting Fornell-Larcker requirements.

Structural Model and Result
After completing the outer testing, proceed with testing the inner model to find out how much influence arises from each path, both direct and indirect influences based on models formed from exogenous latent variables on endogenous latent variables. The value of standardized direct effects (path coefficient) ranges from -1 to +1. Here is a table of direct effects Based on the table, it can be seen that the impact of perceived value on E-Loyalty and the effect of service quality on E-Loyalty are both significant with influence values of 0.091 and 0165 respectively. Based on these results, it can be seen that the direct influence of service quality on E-Loyalty is greater than the influence of perceived value on E-Loyalty.
In addition to direct effect, each variable has an indirect effect where this indirect influence is through an intermediate variable as shown in the following table. The intermediate variable in this study is E-Satisfaction. Based on the results of indirect effect as shown in table 8 where the p-value for both is below 0.05. This shows that E-Satisfaction mediates the effect of perceived value on E-Loyalty and E-Satisfaction also mediates the effect of service quality on E-Loyalty. After knowing the direct and indirect effects, the total effect can be known, where this total effect is a combination of direct and indirect influences of each variable. The total effect of all variables can be seen from the following table.  To determine the influence of each variable can be seen from the F-Square value where the F-Square value < 0.02 can be ignored or there is no effect while the F-Square value between 0.02 to 0.14 means that the effect is small, while the influence of F-Square between 0.15 to 0.35 has a moderate effect while the F-Square value above 0.35 has a big effect (Sarstedt et al., 2017). Based on the table below, it can be seen that the effect of Service Quality on E-Satisfaction is the highest at 0.738, while E-Satisfaction on E-Loyalty of 0.042 can be said to have no effect or can be ignored. Next to find out the magnitude of the influence of R-Square to find out the strength of the model. The R-Square describes how strongly an endogenous variable can be described by its exogenous variable. The R-Square seen is an adjusted R-Square where from the data in table 11 the highest R-Square value for the E-Satisfaction variable which has an influence of 0.912 which is very high means that the strength of this model is very good. The R-Square value for E-Loyalty also has a very high influence of 0.855. This means that the strength of this model is very good where 85.5% of this model can be explained by the variables in this study, where the remaining 14.5% is explained by other variables outside the variables specified in this model. Furthermore, from the fit model, it can be seen from the RMS Theta value or Root Mean Square Theta < 0.102 or the SRMR (Standardized Root Mean Square) value < 0.10 or the NFI value > 0.9. Of these three criteria, it is enough to meet only one. Based on the data obtained from the results of the study as shown in table 8, in the calculation of the Fit model, the saturated model value for SRMR is below 0.10 so that this shows that this measurement model is fit. After testing the fit model, proceed to bootstrapping analysis with basic methods to see the significance of each hypothesis. Here are the bootstrapping results for all hypotheses in this study. Based on these results, it can be seen that from 5 research hypotheses, it can be concluded that all hypotheses are supported by data. So it can be concluded that the whole hypothesis is proven. Based on the results of the analysis, it can be concluded that all of them have a positive effect, and the biggest influence is hypothesis 2, namely the relationship between Service Quality and E-Satisfaction, then the smallest influence is hypothesis 5, namely the influence of E-Satisfaction on E-Loyalty.

CONCLUSION
From the results of the analysis, it is proven that the 5 research hypotheses are accepted and the biggest influence is Service Quality on E-Satisfation. In addition, E-Satisfaction as an intermediary variable here is also proven to mediate the effect of service quality on E-Loyalty and perceived value on E-Loyalty. This means that in the q-commerce industry in Indonesia, consumers will become loyal if the quality service provided by the q-commerce platform is good, so that consumers will become satisfied and eventually become loyal. This proves that the positive and significant relationship engendered by the E-Service quality dimension in relation to E-Loyalty is essential for the sustainability of the company and its continued concern, regardless of its physical and virtual arrangements in the exchange of products (Goods and services); that is, to operate appropriately in retaining existing consumers. The results show that the four dimensions mentioned are good and positive predictors of E-Service quality. Based on the results of this questionnaire, it is also known that the majority of Quick Commerce users are women and work as private employees, so in the future it is necessary to make the right strategies aimed at Quick Commerce customers so that they can continue to shop loyally at Quick Commerce.
Reliability, security, responsiveness and convenience that have significantly high and positive values reflect the main impact of E-Service quality, implying that customers have higher satisfaction on the Quick Commerce platform in terms of protecting customers' personal information. As a result, data privacy guarantees are an important factor in building customer satisfaction. With the expectation of service quality through online purchases and its value to loyalty, received value, and electronic satisfaction, customer retention or repurchase remains an antecedent, especially among tier-1 city residents. In addition, the value received also has a positive impact on customer retention or antecedents of E-Loyalty. Values that are beneficial to customers such as economical prices, shopping comfort, ease of operation are able to provide customer satisfaction and become a catalyst for the creation of customer retention.
Advice that can be given to business people, the most important thing is the service quality provided by q-commerce for consumers, then followed by perceived value felt by consumers. Thus, the need for q-commerce continues to pay attention to the convenience and security of using these apps and ensure that all promises that have been given to consumers can run well. The results of it all will certainly provide satisfaction and loyalty for these q-commerce consumers.