The Development of Customer Loyalty Model for Mobile Travel Agent Application

These factors


Introduction
The penetration rate of the handheld market in Indonesia was estimated to increase dramatically from 69.4 million in 2016 to 103 million in 2018 [1]. It encourages the development of m-commerce as a new way to carry out the transaction. M-commerce refers to the ability to conduct commercial transactions using mobile phones connected to a wireless network [2]. Indonesia has become one of the fastest-growing m-commerce markets in the world, with an increase of 155 percent in 2016 [3]. This growing market opens up opportunities for start-up companies to offer customers innovative products and services. An online travel agent (OTA) was one of the start-up categories widely used by people to trade [4]. The start-up companies included in the OTA category are Traveloka, The growth of smartphone users in Indonesia opens up business opportunities for service providers, in particular via m-commerce. However, along with these great opportunities and rivalry, service providers are required to make consumers faithful to the products and services they deliver. This research aims to develop a conceptual model for an application based on m-commerce, particularly in the context of the mobile travel agent application. The research model is based on the application of five dimensions of Mobile Service Quality (M-S-QUAL), customer engagement, and commitment as the predecessor factors of customer loyalty. The study was conducted on 760 MTA application users who participated in an online survey. The data were divided into two groups based on the level of income to be further analyzed using the Multigroup Structural Equation Modeling (SEM). The Multigroup SEM analysis shows that the concept of customer loyalty was explained differently by the two groups with different income levels.  [9]. Giovanis and Athanasopoulou [11] examined a model consists of e-trust and e-satisfaction, as well as six dimensions of E-S-QUAL consisting of reliability, responsiveness, privacy, quality of information, ease of use and web design on the e-commerce website. The results showed that E-S-QUAL has a positive impact on e-trust and e-satisfaction. Furthermore, e-trust and e-satisfaction had a positive effect on e-loyalty. Lee and Wong [10] used the E-S-QUAL factor by adding trust and commitment variables to test the effect on customer loyalty on the e-commerce website. The result showed that the E-S-QUAL factor has a significant impact on satisfaction and, together with trust and commitment, has a significant impact on loyalty. Moreover, Thakur [9] developed a model to examine the impact of customer engagement, customer satisfaction, usability, and convenience on customer loyalty in the online shopping application. The results indicated that customer engagement has a significant impact on customer loyalty.
Based on previous studies, no studies have been conducted on the issue of customer loyalty to Mobile Travel Agents (MTAs). This study developed a conceptual model for the e-commerce-based application, especially in applications for mobile travel agents. This study aimed to identify the factors affecting customer loyalty in the MTA application. There were two critical contributions to this research. First, this work integrated a customer loyalty model from Giovanis and Athanasopoulou [11] with commitment variables from Lee and Wong [10], customer engagement second-order variable from Thakur [9], and M-S-QUAL service quality dimension from Thakur [9]. M-S-QUAL was selected from E-S-QUAL [8,9] and SERVQUAL [16,17]. Second, this research used Multigroup Structural Equation Model to explore the characteristics of customer loyalty across groups with different income levels. The article is organized in the following: section 2 is a methodology that elaborates on the data and methods used to analyze the model; Section 3 describes the results and discussion of the study, and section 4 comprises conclusions and future work.

Model and Hypothesis
We developed a structural model to show the variable relationship. We also analyzed the model used the Structural Equation Models (SEM) and the Multigroup SEM. This section focused on structural equation model construction and variable-relationship hypothesis. The structural model can be seen in Fig. 1. The model used the Giovanis and Athanasopoulou [11] customer loyalty model as the basic model. The model was further expanded by adding a commitment variable from Lee and Wong [10], a second-order customer engagement variable from Thakur [9], and M-S-QUAL quality service variable from Huang, et al. [18]. The descriptions and indicators for each variable used in this analysis are shown in Table 1. Table 2 presents the research hypothesis.
The Customer Loyalty Model characteristics provided further information across groups, and a Multigroup SEM analysis was conducted. This study used the level of income as a grouping indicator because customer loyalty has a negative relationship with price sensitivity [8]. It means that income-sensitive customers tend to have low loyalty. Multigroup SEM analysis was performed by dividing the sample into two groups. Respondents with income below 9,000,000 IDR were categorized into the lower-middle group. Meanwhile, respondents with income over 9,000,000 IDR was grouped into the upper-middle group.

Reliability/ Fulfillment
The level of reliability of m-commerce service providers in the provision of applications that can provide full product and service information and meet user expectations precisely as promised.

FUL1-FUL3
[18], [20] Contact Service provider support to respond directly to the needs and concerns of the customer.

CON1-CON3 [18]
Responsiveness M-commerce service provider's capacity to offer applications that facilitate effective applicationsolving problems, including order details, cancellations, refunds.

Security/ privacy
The ability of m-commerce service providers to develop applications that maintain consumer protection in performing transactions, including safeguarding all issues related to user data confidentiality, and payment methods used from all forms of abuse.

Customer Satisfaction
An emotional/psychological condition in which the consumer is happy and satisfied with the products and services provided to meet individual standards when using the M-Commerce Program.

SAT1-SAT3 SAT4 SAT5
[23], [25], [24] Trust An accumulation of trust that the customer feels for the reliability of the knowledge, products, and services obtained from the experience of the application of m-commerce. TRU1-TRU5 [22], [26] Commitment The ability to maintain good relations and to use the future products and services of the same platform for m-commerce.

COM1-COM3 [27]
Customer Engagement Psychological problems are arising from ongoing encounters and two-way interactions between consumers and service providers to improve customers ' willingness to use products or services from the same m-commerce program.

Monetary Experience
Positive customer experience achieved exclusively through a series of attractive financial-related offers provided by m-commerce applications when customers use products and services.

Utilitarian Experience
Positive customer experience related to the availability of information that can help in the decision-making process to purchase products and services through the m-commerce application.

UTI1-UTI3 [9]
Customer Loyalty The attitude of each individual/customer to purchase and use products and services from the same application and to recommend the application to others.

Data collection method
This research used a quantitative approach. The primary data was obtained directly from the target sample. The population was users of mobile travel agent applications that had purchased products and services. The research instrument used open-ended and closed-ended questionnaires. The first part of the survey included an introduction and a short explanation of the research. The second part contained a few brief questions about the identity of the respondents. The last part was the central part of the questionnaire, which included statements representing the measurement indicators. The Likert six-point scale was used as a measurement scale to avoid neutral responses. The "1" scale means Strongly Disagree, and the "6" scale means Strongly Agree.
Respondents were asked to rate each question. Method for collecting data used convenience sampling. This method was chosen because of the ease and time efficiency. Questionnaires were distributed online via electronic media such as e-mail and sample target links. Seven hundred sixty samples were collected, 35 of which were used for initial data analysis.

Data analysis technique
Data analysis was conducted by SEM and Multigroup SEM using MPlus software. Furthermore, multigroup SEM consists of measurements of invariance, the structure of invariance, and the path of the coefficient of invariance. A multi-group analysis was conducted to prove the hypothesis of differences in factors affecting customer loyalty in the sample group with different income levels. The survey was divided into two groups based on the type of income referred to that in Model Development. Table 3 shows the results of the hypothesis test. Efficiency affected the satisfaction of the customer. This finding was compatible with Lee and Wong [10] and proved a decisive impact on customer satisfaction. Service providers must provide the MTA web interface with easy and quick access to information and transactions. Efficiency was the second significant influence of trust after privacy/security. It was consistent with Gustafsson, et al. [36]. Therefore, there was a significant influence in m-commerce between efficiency and trust. Application providers can consider these findings. The objective was to provide an easily and quickly accessible MTA application interface for information and transaction settlement.

Result and Analysis 3.1 SEM Analysis
Reliability/fulfillment had a positive impact on customer satisfaction. Giovanis and Athanasopoulou [11] found a positive relationship between reliability and customer satisfaction. Service providers must provide reliable services that can meet the expectations of the costumers. The quality of the service in question included the consistency of the order details (order name, order quantity date, order code), the pricebased information on the order page, and the speed of delivery of the e-voucher. Reliability was part of the quality of service aspect had a significant impact on customer satisfaction. Giovanis and Athanasopoulou [11] found a positive relationship between trust and reliability. Kim, et al. [33] also supported this finding.
The contact did not impact on customer satisfaction. Service provider assistance in solving complaints and helping customers solve problems did not affect customer satisfaction. It was because the other four dimensions of service quality were considered to be far more critical in terms of customer satisfaction. This finding was similar to the finding in Kumbhar [37], which found that contact had no significant effect on customer satisfaction. Al-nasser, et al. [21] found that contact had a significant impact on ecommerce trust. Access to service provider assistance in responding to concerns and helping customer problems were expected to increase consumer trust in online transactions. However, In this study, the relationship did not prove it.  Responsibility influenced customer satisfaction. It was confirmed by Giovanis and Athanasopoulou [11] and Lee [32]. Giovanis and Athanasopoulou [11] showed that customer satisfaction did not directly affect trust. However, it had an indirect effect. This research had different perspectives. Responsiveness and customer satisfaction were directly related. The availability of access to communication with service providers was able to make customers feel safe and convinced to transact using the MTA application. Privacy/security had a significant direct effect on customer satisfaction. The issue of personal data confidentiality, particularly credit card data used, was an essential aspect of MTA application [10]. Service providers must ensure data security to satisfy customers.
Privacy/security had a positive impact on trust. The finding was confirmed by Lee and Wong [10]. Customers who felt protected were prone to have more trust to use the MTA application. Service providers should implement security standards that ensure the protection and security of the costumers' data. Customer satisfaction had a significant impact on customer loyalty. Customers who were satisfied with the quality of the services provided were likely to re-purchase or re-use products and services from the same service provider [27] [11][10] [9] [22]. Customer satisfaction had a significant direct influence on trust [26] [11][10]. Satisfied customers should form a bond with the service provider [38].
Trust had a positive effect on customer loyalty. In m-commerce, customer trust was significant [33]. Trust was the basis for starting, establishing, and maintaining relationships between customers and online retailers [33]. Trust had a significant effect on commitment. The effect of trust on commitment was more substantial than the effect of trust on customer loyalty [10]. The commitment had a significant influence on customer loyalty [5] [35]. Customer engagement had a positive impact on customer loyalty. Service provider needed to develop a strategy to support customer relationship [9] [31] [32]. Customers with positive financial transactions and access to information were more engaged with the service provider [9].

Multigroup SEM
The results of the Multigroup SEM path analysis are shown in Table 4. Based on Table 4, six hypotheses were rejected. These were H0-1a, H0-1b, H0-2a, H0-2b, H0-7, and H0-9. It indicates another relationship between the constructs caused by the differences in the sample group. Efficiency in the lower-middle-income group had a significant direct effect on customer satisfaction. It was different from that in the upper-middle groups (H0-1a). Respondents of lower-middle-income feel that efficiency was one of the factors affecting the satisfaction of the MTA application. This result proved the previous research that found the role of income in moderating the customer satisfaction model relationship [39]. Efficiency influenced trust in the lower-middle-income group (H0-1b). However, It did not have a positive influence on trust in the upper-middle groups. To lower-middleincome, customer's efficiency as one of service quality aspects had an impact on their trust in using the MTA application. However, their trust in using the MTA application was not influenced by efficiency for the upper-middle-income group. This study supported Bakar and Ilkan [19]. Reliability/fulfillment did not influence customer satisfaction in lower-middleincome groups (H0-2a). However, The efficiency influenced customer satisfaction for upper middle income. Upper middle-income groups were concerned about the reliability of the application in assessing overall quality. The Reliability affected their satisfaction with the MTA application. Higher-income individuals considered more factors to be satisfied [39]. Therefore, it was concluded that the income difference significantly affected the relationship between reliability and customer satisfaction. The result showed that reliability/fulfillment did not impact trust in both groups (H0-2b). However, Lower middle-income respondents showed lower reliability/ fulfillment effects on trust compared to upper-middle-income respondents. This different level of trust can be attributed to the overall lower level of trust between low-income groups [24]. ISSN : 1978- Customer satisfaction influenced trust in the lower-middle-income group (H0-7). It indicated that only low-middle-income respondents were satisfied. The reason for this statement was that the lower-middle-income groups had more trust than the uppermiddle-income groups did [29]. However, for upper-middle-income respondents, regardless of whether they felt satisfied or not, they continued to trust the MTA application because their data were protected. Trust had a significant impact on commitment in the two groups tested (H0-9). It can be concluded that the strength of trust in commitment was explained differently by the two groups [10].

Managerial Implication
In order to improve customer loyalty, service providers must pay attention to customer satisfaction, trust, commitment, and customer engagement. The dimensions of service quality, which consist of efficiency, reliability, fulfillment, responsiveness, and privacy/ security, must be maximized, although their influence on customer loyalty is indirect. An excellent service must be provided to create customer satisfaction. Service providers may consider providing refunds, rescheduling online check-in, and two-step verification (for credit card transactions) features in their MTA applications. Differences in the relationship between the lower middle and upper-middle-income groups can be used as a factor in the planning of a business strategy. Strategies should be differentiated to represent the features of these two different groups.

Conclusion
Customer satisfaction, trust, commitment, and commitment were factors that had a significant impact on customer loyalty. Service providers must also pay attention to the quality of service measurements, which consist of efficiency reliability/fulfillment, responsiveness, and privacy/security in order to build customer satisfaction and trust. Another result was that variations in income levels might affect the relationship between variables in the study model. Grouping categories can further develop this study based on the level of using the MTA application. This analysis model can be developed by evaluating the interaction between customer satisfaction and responsibility for the same period. Another suggestion is to create a quality service assessment instrument that matches the characteristics of the Indonesian respondents and falls into the field of mobile commerce research.