The Effect of Brand Image, E-Service Quality, and Customer Value on Decisions to Use Traveloka in Semarang

: Traveloka's existence in Indonesia makes it one of the major online travel agents that are in demand by the Indonesian people to fulfill their travel needs. However, in the last few years there has been a tendency to decrease in sales as indicated by data on the decrease in the percentage of sales in 2016-2020. This study aims to determine the effect of brand image, e-service quality, and customer value on purchasing decisions at Traveloka online travel agents. This type of research is explanatory research with sampling using non-probability sampling technique. Data collection is done by using google form. The sample used is 100 respondents who have used purchases at Traveloka online travel agents in Semarang City. This research uses validity test, reliability test, relation coefficient, coefficient of determination, simple regression, multiple regression, t-test significance, and f-test supported by SPSS. The results obtained in this study are brand image, e-service quality, and customer value, which individually or collectively affect purchasing decisions at Traveloka online travel agents. Based on the results of the analysis that has been carried out, the researchers suggest Traveloka to improve the platform design, optimize services for Traveloka users' complaints and provide a fast and personalized response. Companies are advised to provide training, direction, and attention to customer service in order to provide the best service. In addition, it is recommended to improve service quality intensively so that the benefits (results) obtained by buyers/users are in accordance with the effort and costs incurred, evaluate customer experience, and optimize the recommendation feature.


INTRODUCTION
The tourism industry in Indonesia today has a fairly rapid growth.This situation can be seen from the increasing number of online travel agents.Online Travel Agent (OTA) is a site that facilitates users in completing travel needs, including airline tickets, hotels, trains, buses, car rentals, and other tourist activities so that consumers do not need to go to ticket counter to order tickets or hotel reservations.It is now more effective and efficient.More and more consumers are now turning to OTAs to buy tickets or hotel reservations.This is because it allows buyers/users to easily search for travel products by reading reviews of other buyers/users, comparing prices and benefits.The existence of online travel agents makes it easier for consumers who want to travel so that all their traveling needs become easier, more practical, and users can find out the information needed about their trip correctly.This situation certainly requires service providers to continuously innovate to offer superior travel products according to the needs of the Indonesian people.Every service provider must be able to learn how buyers/users make purchasing decisions and what factors can influence consumer behavior when making purchasing decisions for travel products.Lu, Zeng, & Fan (in Hidayanto et al., 2007) show that if individuals are involved in e-commerce activities, individuals must take into account the high level of risk and uncertainty associated with opportunistic behavior from other parties, which in this context is aimed at on Traveloka as an online travel agent.Furthermore, the tight competition in online travel agents has caused the percentage of Traveloka's Top Brand Index (TBI) in 2016-2020 to tend to decrease (topbrand-award.comaccessed on June 1, 2021).The index, which tends to decrease, indicates that purchasing decisions at Traveloka have also decreased.This is based on TBI measurement based on three parameters, namely mind share, market share, and commitment share.The size of the index can also identify the image of Traveloka itself, then it can influence potential buyers/users in purchasing travel products at Traveloka.In addition to the percentage of TBI which tends to decrease, researchers also found a number of negative reviews related to Traveloka, both on the Google Play Store and on Traveloka's own social media.Most Traveloka users complain about the quality

RESEARCH METHODS
This study uses the type of explanatory research with a quantitative approach.Explanatory research is research to find an explanation of the correlation between variables in research and to know the existence of the influence between variables in addition to testing the hypothesis that was formulated initially (Sugiyono, 2006).The population that will be used by the researcher is Traveloka users in Semarang City, whose number is not identified.Based on this population, the sampling refers to the theory of Cooper and Emory (1996), which appoints 100 Traveloka users in the city of Semarang.The sampling technique used was non-probability sampling with purposive sampling.The research instrument used for data collection was a questionnaire.  1, it is known that the value of the correlation coefficient of the two variables is 0.664.The correlation coefficient value of 0.664 is in the range of 0.60 -0.799 so it can be interpreted that between the brand image variables and purchasing decisions there is a strong correlation level or level of relationship.So, if there is an increase or decrease in the brand image variable on the respondent's perception, it can have an impact on the purchasing decision variable.2, it is known that the value of the correlation coefficient of the two variables is 0.777.The correlation coefficient value of 0.777 is in the range of 0.60 -0.799, so it means that between the variables of e-service quality and purchasing decisions there is a strong correlation.If there is an increase or decrease in the e-service quality variable on respondents' perceptions, it can have an impact on the purchasing decision variable..000N 100 100 **correlation is significant at the 0.01 level (1-tailed) Source: primary data, 2022 Based on Table 3, it is known that the value of the correlation coefficient of the two variables is 0.719.The correlation coefficient value of 0.719 is in the range of 0.60 -0.799, so it means that the customer value variable and purchasing decisions have a strong correlation.If there is an increase or decrease in the customer value variable on the respondent's perception, it can have an impact on the purchasing decision variable.Based on Table 8, the significance value is 0.00 < 0.05, then Ho is rejected and Ha is accepted, so that the brand image variable (X1) has a positive influence on the purchasing decision variable (Y).The regression coefficient for the brand image variable (X1) is 0.465 with a constant value of 5.835.Based on Table 9, the significance value is 0.00 < 0.05, then Ho is rejected and Ha is accepted, so that the e-service quality variable (X2) has a positive influence on the purchasing decision variable (Y).The regression coefficient for the variable e-service quality (X2) is 0.418 with a constant value of 5.381.Based on Table 10, the significance value is 0.00 < 0.05, then Ho is rejected and Ha is accepted, so the customer value variable (X3) has a positive influence on the purchasing decision variable (Y).The regression coefficient for the brand image variable (X1) is 0.466 with a constant value of 7.702.11, a constant value of 2.808 is obtained where if there are no brand image, e-service quality, and customer value variables, the magnitude of the perception of the purchase decision is 2.808.The regression coefficient for the brand image variable (X1) is 0.135, e-service quality (X2) is 0.234, and customer value (X3) is 0.176 so it is said that brand image (X1) is 0.135, e-service quality (X2) is 0.234, and customer value (X3) can have a positive influence on purchasing decisions (Y).

HYPOTHESES TESTING t-test
The t-count of the brand image variable is 8.800.Furthermore, the value of the t-table is searched through the previous df (degree of freedom), so that df = n -2 = 100 -2 = 98 is obtained.The df is obtained as much as 98, then if observed through t-one tailed using 5% significance, the value will be obtained t table with a magnitude of 1.660 so it can be concluded that the value of t-count (8.800) > t-table (1.660), then Ho is rejected and Ha is accepted.This means that hypothesis 1 "brand image has a positive effect on purchasing decisions (Y)" is accepted.Based on the calculations from Table 5, it is found that the t-count of the e-service quality variable is 12.201.Furthermore, the value of the t-table is searched through the previous df (degree of freedom), so that df = n -2 = 100 -2 = 98 is obtained.The df is obtained as much as 98, then if observed through t-one tailed using 5% significance, the value will be obtained t table with a magnitude of 1.660 so it can be concluded that the value of t-count (12.201) > t-table (1.660), then Ho is rejected and Ha is accepted.This means that hypothesis 2 "e-service quality has a positive effect on purchasing decisions (Y)" is accepted.Based on the calculations from Table 6, it is found that the t-count of the customer value variable is 10.236.Furthermore, the value of the t-table is searched through the previous df (degree of freedom), so that df = n -2 = 100 -2 = 98 is obtained.The df is obtained as much as 98, then if observed through t-one tailed using 5% significance, the value will be obtained t table with a magnitude of 1.660 so it can be concluded that the value of t-count (10.236) > t-table (1.660), then Ho is rejected and Ha is accepted.This means that hypothesis 3 "customer value has a positive effect on purchasing decisions (Y)" is accepted.Based on the calculations in Table 11, it is known that the calculation results from the F value are 62.588 and the f-table value is searched through the previous df (degree of freedom), so that we get df = n -2 = 100 -2 = 98.Obtaining df as much as 98, then if observed through the t-one tail using a significance of 5%, the t-table value will be obtained with a magnitude of 2.697.From the results of these calculations, it is shown that the calculated F value (62.588) > F table value (2.697), then Ho is rejected and Ha is accepted.This means that hypothesis 4 "there is a positive influence between brand image, e-service quality, and customer value on purchasing decisions" is accepted.

DISCUSSION
Based on the tests that have been carried out, the results show that the brand image variable has a positive influence on the decision to use services at Traveloka in Semarang City.This situation can be seen from the results of the study which explains that the relationship between the brand image variable and purchasing decisions is stated to be strong because it has a correlation coefficient value of 0.687 where this value is in the interval range of 0.60 -0.799.Furthermore, the results of testing the coefficient of determination between brand image and purchasing decisions produce an R2 value of 0.441 or 44.1%, meaning that there is 44.1% of the purchasing decision variables contributing to the influence of the brand image variable, while the remaining 55.9% is explained by the variable others that were not studied in this study.Based on the calculation of the t test, the results obtained t count is 8,800 > t table which is 1,660 then Ho is rejected and Ha is accepted, so that hypothesis 1, namely "there is a positive influence between brand image on purchasing decisions" is accepted.
In addition, it is known that the e-service quality variable has a positive effect on the decision to use services at Traveloka in Semarang City.This situation can be seen from the results of the study which show that the relationship between the e-service quality variable and purchasing decisions is included in the strong category because it has a correlation coefficient value of 0.777 where this value is in the interval range of 0.60 -0.799.Furthermore, the results of the calculation of the coefficient of determination between e-service quality and purchasing decisions that produce an R2 value of 0.603 or 60.3%, meaning that there are 60.3% of the purchasing decision variables contributed to the influence of the e-service quality variable, while the remaining 39, 7% is explained by other variables not examined in this study.As for the results of the calculations on the t test, the results obtained where the value of t count (12.201) > t table value (1.660) then Ho is rejected and Ha is accepted, so that hypothesis 2, namely "there is a positive influence between e-service quality on purchasing decisions" is accepted.
In addition to brand image and e-service quality, it is found that the customer value variable has a positive influence on the decision to use services at Traveloka in Semarang City.This situation can be seen from the results of the study which explains that the relationship between the variable customer value and purchasing decisions is declared strong because it has a correlation coefficient value of 0.687 where this value is in the interval range of 0.60 -0.799.Furthermore, the results of testing the coefficient of determination between brand image and purchasing decisions resulted in an R2 value of 0.517 or 51.7%, meaning that there was 51.7% of the purchasing decision variables contributed to the influence of the customer value variable, while the remaining 48.3% was explained by the variable others that were not studied in this study.Based on the t-test calculation, the results obtained by tcount are 10.236 > t-table which is 1.660, then Ho is rejected and Ha is accepted, so hypothesis 3, namely "there is a positive influence between customer value on purchasing decisions" is accepted.This study also found that the variables of brand image, e-service quality, and customer value had a positive effect on purchasing decisions.This is evidenced by the results of the analysis which shows that the relationship between brand image, e-service quality, and customer value with purchasing decisions is included in the very strong category because it has a correlation coefficient value of 0.813 where this value is in the interval 0.80 -1,000.Furthermore, the coefficient of determination of brand image, e-service quality, and customer value on purchasing decisions is 0.662 or 66.2%, meaning that there are as many as 66.2% of the purchasing decision variables contributing to the influence of the brand image variable, e-service quality, and customer value, while the remaining 33.8% is explained by other variables not examined in this study.As for the results of the calculations on the t test, the results obtained where the calculated F value is 62.588 > the F table value is 2.697, so that it can be concluded that hypothesis 4, namely "brand image, e-service quality, and customer value have a positive effect on purchasing decisions" is accepted.

Table 1 .
The scoring in the research questionnaire uses a Likert scale.The analytical methods used are validity test, reliability test, correlation coefficient test, multiple correlation test, coefficient of determination test, determination test (R2), simple regression analysis, multiple regression analysis, t test, and F test. Brand Image Correlation Test Results on Purchase Decisions

Table 3 .
Customer Value Correlation Test Results on Purchase Decisions

Table 4 .
Test Results of Brand Image Determination Coefficient on Purchase Decisions

Table 4 ,
it is known that the coefficient of determination of brand image on purchasing decisions is 0.441 or 44.1%.

Table 5 .
Test Results of E-Service Quality Determination Coefficient on Purchase Decisions

Table 5 ,
it is known that the coefficient of determination of e-service quality on purchasing decision is 0.603 or 60.3%.

Table 6 .
Test Results of Customer Value Determination Coefficient on Purchase Decisions

Table 6 ,
it is known that the coefficient of determination of e-service quality on purchasing decisions is 0.517 or 51.7%. ISSN:

Table 7 .
Test Results of Brand Image, eService Quality, Customer Value Determination Coefficient on Purchase Decisions Constant), Brand Image, E-Service Quality, Customer Value Source: primary data, 2022Based on Table7, it is known that the coefficient of determination of brand image, e-service quality, and customer value on purchasing decisions is 0.662 or 66.2%.

Table 8 .
The Simple Regression Test Result of Brand Image on Purchase Decisions Coefficients a

Table 9 .
The Simple Regression Test Result of eService Quality on Purchase Decisions Coefficients a

Table 10 .
The Simple Regression Test Result of Customer Value on Purchase Decisions Coefficients a ISSN:

Table 11 .
Multiple Linear Regression Test Results of Brand Image, E-Service Quality, and Customer Value on Purchase Decisions

Table 11 .
Calculated F Test Results between Brand Image, E-Service Quality, and Customer Value on Purchase Decisions