Designing The Proposed Improvement of TIX ID Online Cinema Ticket Purchase Service Quality

TIX ID is an application in Indonesia that provides new experiences in purchasing cinema tickets. The success of a company can be seen from service quality in fulfilling the satisfaction of customers. Therefore, this study aims to design the proposed improvement in TIX ID online cinema ticket purchase service quality. This study was conducted in three stages, namely, testing the hypotheses of the relationship between variables using PLS-SEM, measuring service quality using the servqual method, and designing the proposed improvement. The result of this study is e-service quality has a significant effect on e-satisfaction and e-loyalty. In addition, the dimension of eservice quality that is a dominant effect on the low TIX ID service quality is compensation, with a gap of 27%. Recommendations for improvement are providing customer access to contact customers service directly and SOP of the cooperation TIX ID and DANA.


I. INTRODUCTION 1
According to the survey results conducted by the APJII (2019), it was agreed that the penetration of internet users in Indonesia in 2018 reached 64.8% or 171,17 million internet users, up to 10.12% from the previous year. The increase reached 27 million users from Indonesia's total population based on BPS data, which amounted to 246.16 million people. The internet's existence makes daily activities more manageable and provides convenience for various aspects of business activities. The nature of service and the process of providing service has changed along with the development of information technology (Paluch & Blut, 2013).
Along with the development of technology in Indonesia, the growth of the film industry in Indonesia is also increasing with the increasing number of domestic film production and the number of viewers. Based on Mediarta (2020), it is known that in 2017 there were 42.248.908 viewers with the number of films circulating as many as 112 films. In 2018 there were 51,192,832 viewers with the number of films distributing as many as 128 films, and in 2019 there were 51,901,745 viewers with the number of films circulating as many as 129 films.
With the development of the film industry in Indonesia, many people take advantage of it to create many business opportunities. TIX ID is one of Indonesia's applications that provides a new experience in purchasing cinema tickets. The presence of TIX ID makes a positive impact for its users, which is increasing the efficiency of wasted time because queuing to buy tickets at the counter that sometimes already queued long. Still, the ticket has been sold out for a certain schedule or even the day (Andrea, 2019).
However, a company's success can be seen from the quality of service in fulfilling people or customers (Swaraswati, 2019). In this case, there are still many customers who report complaints or negative reviews of services provided when purchasing online tickets using TIX ID. Based on the review's rating of the Playstore, in 2018, users who gave a rating of 1 and negative reviews on TIX ID as many as 1250 users. In 2019 who gave ratings 1 and negative reviews on TIX ID for as many as 910 users, and in 2020 who gave ratings 1 and negative reviews on TIX ID for 273 users. Negative reviews such as there is a unilateral cancellation, but a refund from the TIX ID has not been received (Fei, 2019), ticket booking has been successful, and the balance has been cut, but the cinema ticket is not upcoming (Widyatmoko, 2018;Henry, 2018).
Creating a good quality of service is one of the most important things and need to get serious attention from the company because it will impact customer satisfaction (Hadi, 2011). According to Fornell (1992), Customer satisfaction is key in creating customer loyalty. The company's decision to make service improvements is a decisive umbrella in following up on customer complaints in a problem that later can bind customer loyalty (Elu,2015).
Measurement of service quality is needed for companies to assess customer opinions about their service performance. Voice of the customer serves as a guide for service providers by giving their suggestions and complaints (Howey, 2001). Parasuraman et al. (1985) Recommend one of the approaches used to measure quality, namely the Service Quality method. Zeithaml et al. (2010) mentioned that the customer gap is the heart of the model.
The Servqual model is originated from the gap model, whose main purpose is to investigate whether customer's expectations are met or not accurately. Calculating the expectation score and perception score is not enough to know the actual causes of service declination. The gap score is the effective and benevolent implication for understanding the real situation and how the over-promises or lower performance destroy service quality. The service gap can be demonstrated through the subtraction between expectation and perception (Sultana & Rana, 2010). After measuring the quality of the TIX ID service with servqual, determine the dimension priority of e-service quality given the improvement proposal's design by using a Pareto diagram. The Pareto principle is also known as the 80/20 rule; doing 20% of the work can produce 80% benefit of the work (Bong et al., 2019).
Many previous studies examine the effect of service quality on customer satisfaction and loyalty. Research conducted by Santika & Pramudana (2018) aims to find out the role of e-satisfaction in mediating the relationship of eservice quality with e-loyalty on e-commerce sites for ticket reservations and hotel rooms (online travel sites) using linear regression analysis. Research conducted by Lorena (2018) aims to determine the effect of electronic service quality on online consumer satisfaction that impacts consumer loyalty on Bukalapak online shopping sites using path analysis with the program. Research conducted by Widjaya (2016) aims to find the effect of e-service quality, customer satisfaction, and loyalty on traveloka customers using SEM analysis with AMOS 23 application.
The differences in research conducted by Tunjungsari directly and Suryajaya (2012) (2018) with this research is in this research not only to know the influence of e-service quality on e-satisfaction and e-loyalty but also measure the quality of service and designing the proposed improvement TIX ID online cinema ticket purchase.

II. RESEARCH METHOD
The first method used in this study is SEM. Structural Equation Modeling (SEM) is a secondgeneration multivariate data analysis method often used in marketing research because it can test linear causal models and theoretically supported additives (Chin & Marcolin, 2003;Haenlein & Kaplan, 2004;StatSoft, 2013). There are two approaches in estimating SEM, namely the covariant-based approach known as Covariance-Based SEM (CB-SEM) and the variance-based approach known as Partial Least Squares SEM (PLS-SEM). CB-SEM is mainly used for research aimed at proving the truth of a theory. In comparison, PLS-SEM is primarily used to develop theories in exploratory research. The main purpose of using PLS-SEM in structural equations is to predict and explain latent variables (Widarjono, 2015). This study using PLS-SEM. Evaluation of PLS-SEM model results can be seen in Table 1. PLS-SEM is used to know the influence between variables used in this research, namely E-Service Quality, E-Satisfaction, and E-Loyalty. Therefore, the research was adapted from previous research conducted by Santika & Pramudana (2018), Lorena (2018), and Widjaya (2016). The following is the hypothesis model used ini figure 1. Based on Figure 1., the hypothesis models used are as follows.
H1. e-service quality (X) has a significant effect on e-satisfaction (Y) H2. e-service quality (X) has a significant effect on e-loyalty (Z) H3. e-satisfaction (Y) has a significant effect on eloyalty (Z) H4. e-service quality (X) has a significant effect on e-loyalty (Z) through e-satisfaction (Y) After the influence of the variables is known, the service quality is measured using the servqual method. Servqual is a service quality method for measuring the scale of differences between expectations and perceptions of consumers. The servqual method comes from a gap whose main purpose is to investigate whether customer expectations are met or not accurately (Parasuraman et al., 1985;Parasuraman et al., 1988).
To facilitate the filling and distribution of questionnaires, the questionnaire was created using Google form to assess respondents' expectations and perception regarding the purchase of TIX ID online cinema tickets with a Likert scale of 1-5. The population in this study are users of the TIX ID application.
According to Sugiyono (2011), the sample is part of the number and characteristics that the population are belonged to. Sampling in this study was conducted using nonprobability selection with the purposive sampling technique. Purposive sampling is a technique to determine the sample with certain criteria so that the data obtained will be more representative. The sample criteria used are respondents who use the TIX ID application to purchase online cinema tickets with a minimum usage frequency of 1-2 times/month. The minimum sample size recommended for the use of PLS-SEM is 30-100 cases (Ghozali, 2006).  Chin in Ghozali (2006), The outer loading value between 0.5-0.6 is considered sufficient.

Discriminant Validity
Average Variance Extracted (AVE) must be more than 0.5. Inner Loading

Path Coefficient
If t statistic > t table, then the effect is significant and the otherwise.

III. RESULT AND DISCUSSION
Questionnaires were created using Google form and distributed from May 1 to May 5, 2020, through various social media, which earned as many as 70 respondents.

PLS-SEM
The first evaluation is the outer model. This evaluation consists of indicator reliability, discriminant validity, internal consistency, and convergent validity. The path diagram of an outer model can be seen in Figure 2.  Indicator reliability is known by looking at outer loading in Table 2. This outer loading represents how much variation in items/ indicators is explained by latent variables and is referred to as extracted variants (Hair et al., 2016). Based on Table 2, it can be known that there are no indicators that have values below 0.5, meaning that it can be said that outer loading is valid.
Discriminant validity is the extent to which a latent variable is completely different from other latent variables by empirical standards. Thus, discriminant validity implies that a latent variable is unique and captures phenomena that are not represented by other variables in the model (Hair et al., 2016). Discriminant validity can be seen on The Heterotrait-Monotrait Ratio of Correlations (HTMT). HTMT is the average of heterotroit-heteromethod (i.e., correlation indicators of throughout the constructs that measure different phenomena), relative to the average motraitheteromethod correlation (i.e., correlation indicators in the same construct) (Henseler, Ringle, dan Sarstedt, 2015). Based on Table 3, it can be known that the HTMT value is < 0.9. Therefore, it can be said that HTMT is valid.
Internal consistency is a form of reliability used to assess results' consistency across items on the same test. Internal consistency can be seen in composite reliability. Composite reliability is used to provide reliability estimates based on correlations between observed indicator variables (Hair et al., 2016). Based on Table 4, it can be seen that the value of the composite reliability of each latent variable ≥ 0.7. So, it can be said that composite reliability is valid.
Convergent validity means that a set of indicators represents one latent variable and which underlies the latent variable. These representatives were expressed using the mean values of extracted variants / AVE. Average Variance Extracted (AVE) is defined as the extent to which a latent variable explains the indicator's variance (Hair et al., 2016). Based on Table 5, it can be seen that all values of AVE> 0.5. So, it can be said that the Average Variance Extracted (AVE) is valid.
Based on the four criteria, the outer model is valid, so there is no need to eliminate the indicator.
Next is evaluating the inner and structural equation models for the R2 test and the significance test. The path diagram of an inner model can be seen in Figure 2. For more details, the results of the path analysis inner model can be seen in the following Table 6.
With significance level (α) 5% or 0.05 and df = 67, t-table value of 2.00 is obtained. Based on the t-statistic value of the E-Service Quality variable against E-Satisfaction obtained a value of 16,342. If the t-statistic value is 16.342 compared to the ttable value of 2.00, then it can be known that the t-statistic value is greater than the t-table with the value P 0.000 < 0.05 so that H1 is accepted.
Next, the t-statistic value of E-Service Quality against E-Loyalty obtained a value of 2.633. If the   t-statistic value of 2.633 is compared to the ttable value of 2.00, it can be known that the tstatistic value is greater than the t-table with a pvalue of 0.009 < 0.05 so that H2 is accepted. Furthermore, the t-statistic value of E-Satisfaction against E-Loyalty obtained a value of 3.219. If the t-statistic value of 3.219 is compared with the t-table value of 2.00, it can be known that the t-value is greater than the t-table with a p-value of 0.001 <0.05 so that H3 is accepted.
Then, the t-statistic value of E-Service Quality towards E-Loyalty indirectly through E-Satisfaction obtained a value of 3.016. If the tstatistic value of 3.016 is compared with the ttable value of 2.00, it can be known that the value of t-count is more significant than the t-table with a value of p 0.003 <0.05 so that H4 is accepted.
The coefficient of determination (R2) is used to measure the value of variation in the independent variable changes to the dependent. In this study, the coefficient of determination of E-Satisfaction was 0.715. This value indicates that the influence of E-Service Quality on E-Satisfaction is 71.5%, and other variables outside this study influence the remaining 28.5%. At the same time, the coefficient of determination of E-Loyalty is 0.639. This value indicates that the impact of E-Service Quality and E-Satisfaction on E-Loyalty is 63.9%, and other variables outside this study influence the remaining 36.1%.

Servqual Method
Based on PLS-SEM, it is produced that E-Service Quality has a significant effect on E-Satisfaction and E-Loyalty. Therefore, it can be seen that if the quality of service is improved. will have an impact on increasing customer satisfaction and loyalty.
Therefore, this research uses the servqual method to measure the service quality of the TIX Figure 3. Path Diagram of Inner Model ID application. Using the Servqual method will generate gap value, which is the difference between perceived perception when using the TIX ID service and the TIX ID service's customer expectation. The gap value can be seen in Table 8.
After calculating the average gap, it can then be processed into a Pareto diagram to prioritize improvements. Figure 4 shows that an enormous average gap occurs in the E-Service Quality variable with the compensation dimension. The compensation dimension has an effect of 27% on the low quality of service from TIX ID.

Proposed Improvement
Complaints/negative reviews from consumers regarding the compensation dimension are when the ticket purchase transaction occurs, and the application says that the ticket purchase transaction was successful. However, a few moments later, it was said that the ticket was canceled and the refund was successful, but in fact, the money deducted for the purchase of the ticket did not return to the consumer DANA account. When consumers want to report complaints, access to talking directly to customer service is difficult because it does not include contact from the company. Consumers must find these contacts through Google. When consumers report the complaint, the consumer service of TIX ID directs consumers to inquire about the problem to DANA's consumer service and vice versa. Due to the consumer complaint, the dimension of compensation was improved, focusing on consumer access in channeling complaints and Standard Operational Procedure (SOP) of the cooperation TIX ID and DANA, so that consumers are not confused when going to report the complaint.

IV. CONCLUSION
Based on the results of hypothesis testing relationships between variables, it can be seen that e-service quality has a significant effect on esatisfaction, e-service quality has a significant effect on e-loyalty, e-satisfaction affects e-loyalty, and e-service quality has a significant effect on eloyalty indirectly through e-satisfaction. Therefore, it can be concluded that if e-service quality improved, then e-satisfaction and eloyalty will also be improved.
Based on the results of service quality measurements using the servqual method, it can be seen that the compensation dimension provides the highest impact on the low service quality of TIX ID online movie ticket purchase with a percentage of 27%. Therefore, the compensation dimension is a priority in the design of proposed improvements.