The Influence Of E-Service Quality, E-Satisfaction, And E-Word Of Mouth Toward Revisit Intention On Tokopedia Website In Padang City THE INFLUENCE OF E-SERVICE QUALITY, E-SATISFACTION, AND E-WORD OF MOUTH TOWARD REVISIT INTENTION ON TOKOPEDIA WEBSITE IN PADANG CITY

: Purpose - This research aims to analyze: (1) The Influence of E-Service Quality toward E-Satisfaction on Tokopedia Website in Padang city (2) The Influence of E-Service Quality toward E-Word of Mouth on Tokopedia Website in Padang city (3) The Influence of E-Satisfaction toward E-Word of Mouth on Tokopedia Website in Padang city (4) The Influence of E-Word of Mouth toward Revisit Intention on Tokopedia Website in Padang city. Methodology - The type of this research is Causative research. The population on this research is all the consumer of Tokopedia in Padang who revisit Tokopedia website with the total sample is 150 person which was taken by using purposive sampling due to unknown the number of population. Finding - Regarding the causal model proposed, the data confirm the relationship set out in the hypothesis. It can therefore be stated that (1) E-Service Quality has a positive effect and significant to E-Satisfaction (2) E- Service Quality has a positive effect and significant to E-Word of Mouth (3) E-Satisfaction has a positive effect and significant to E-Word of Mouth (4) E-Word of Mouth has a positive effect and significant to Revisit Intention. Conclusion - The analysis highlights the positive influence which taking part in the E-Service Quality, E- Satisfaction and E-Word of Mouth toward Revisit Intention. This is the clearly shows in this relationship from the empirical point-of-view. Limitation


INTRODUCTION
The industrial revolution is now entering its fourth phase. The rapid development of science and technology has a major impact on human life. According to Kasali (2017), disruption does not only mean the phenomenon of today change but also reflects the meaning of the phenomenon of future change. One form of disruption is e-commerce. Based on data quoted from CNN, the number of visitors in the first rank is Lazada, with 49,000,000. Lazada is a market leader in Indonesia, the company is based in Southeast Asia. Meanwhile, Tokopedia, an Indonesian original company became a market challenger, with 36,666,667 visitors. to increase the popularity of e-commerce is the interest in re-visit (revisit intention). Revisit intention is a form of behavior (behavioral intention) or the customer's desire to come back, give positive word of mouth, stay longer than expected, shop more than expected (Zeithaml, Valerie A. Mary Jo Bitner, Dwayne D. Gremler. 2009 ) Several factors influence the interest of revisit intention. Through the internet, consumers can find various information about a product. This activity allows word-of-mouth (WOM) to occur which not only becomes a form of interpersonal communication that addresses a product, service or brand, but becomes a form of word-of-mouth communication (WOM) which then propagating globally through the internet and this is often called Electronic Word of Mouth (E-WOM) (Jalilvand, 2012).
Besides e-wom, the factor that can influence revisit intention is that E-service quality is a development towards service quality. According to Parasuraman et al. (2005: 217), the quality of electronic services is widely seen as a buyer's assessment to the extent that sellers facilitate efficiency and effective purchases and sellers facilitate the delivery of high-quality responsive services. Another factor E-Satisfaction is the level of one's feelings after comparing performance or perceived results compared to expectations. According to Zeithaml and Bitner (2004: 86) Satisfaction is to provide a full response to consumers.

METHODOLOGY
The type of research that will be conducted is causative research. Thus research explains and illustrates the relationship between the independent variables of E-Service Quality, E-Satisfaction, and E-Word of Mouth of the dependent variable is Revisit Intention. The population in this study is the people of Padang city who use Tokopedia website whose numbers are unknown. The sample that will be used in this study considered are  c. The majority of respondent jobs is Students, as many as 115 (77%) of respondents

Descriptive Variable Statistics
Variable descriptive analysis in this study aims to provide a general description of the data collection that can be seen based on the mean (average) and standard deviation of each variable.   Hussein (2015), if there is a value of outer loading below 0.6 in an indicator, then the indicator can be eliminated because it does not represent the existing construct. So that in this study there are no indicators that will be removed or deleted.

Discriminant validity
Testing of discriminant validity relates to the principle that every indicator contained in a different construct should not correlate with high (Ghozali & Latan, 2012). Based on table 7, it can be concluded that the square root value of AVE is greater than other constructs. This shows that all constructs in the estimated research model meet the predetermined discriminant validity criteria Average variance extracted (AVE) The measurement model of convergent validity is also seen from the AVE value. AVE value describes the magnitude of the diversity of the manifest/indicator variables contained in the construct. Based on Ghozali and Latan (2012), the use of AVE values is needed in converging validity testing. Furthermore, the recommended AVE value is at least 0.5 to indicate a good convergent validity

Composite reliability
Composite reliability measures internal consistency and the value must be above 0.7 (Ghozali, 2012). The results of composite reliability can be seen in

Data Primer, 2018
Based on table 9 it can be seen that each variable used in this study has a good composite reliability value that is greater than 0.7. Therefore, each variable in this study can be said to have fulfilled the requirements that have a level of reliability in accordance with predetermined requirements.

Evaluation Inner Model Inner model test or Structural Model Test
This test is used to evaluate the relationship between latent conditions as hypothesized in the study, based on PLS output, obtained as follows:

Hypothesis testing
To answer the research hypothesis can be seen T-statistics in table 10 : 1. There is the influence of E-Service Quality toward E-Satisfaction on Tokopedia website in Padang city, this research hypothesis can be accepted or proven and the findings support and provide reinforcement for previous research.
4. There is influence of E-Word of Mouth on Revisit Intention on Tokopedia website in Padang city, this research hypothesis can be accepted or proven and the findings support and provide reinforcement in previous research.

2.. Structural Model Testing (Inner Model)
Inner model evaluation or structural model aims to see the direct or indirect influence between variables. Evaluation on structural models can be started by looking at the R-square value of the endogenous latent variables. According to Ghozali and Latan (2012), R-square values of 0.75, 0.5 and 0.25 define the model strongly, moderately and weakly in explaining the effect of exogenous latent variables on endogenous latent variables.