Consumer Behavior in Adopting Application-Based Transportation Services

Consumer Behavior in Adopting Application-Based


Introduction
The digital revolution in transportation has altered people's lives and market power dynamics, resulting in increased rivalry. Competition exists between conventional and online transportation and online transportation business actors (Kurniawati & Khoirina, 2020). Online transportation has become one of the important needs for people in Pematangsiantar City. These services' existence is related to disruptive innovations in the transportation sector (Alamsyah & Rachmadiansyah, 2018). One of the signs of the beginning of the era of fast transfer of information is unlimited and flexible internet access. The increasing use of internet services is considered to have the potential to be juxtaposed with transportation which also requires innovation in service access (Lie et al., 2019). Judging from technology's rapid and sustainable development, developers must create an innovation movement by providing convenient aspects to customers with the Technology acceptance model (TAM) and theory of reasoned action (TRA) approaches, states that performance expectancy has a significant effect on the intention to use information system services. The same thing was also conveyed by Purnamasari et al. (2020) by using the same theoretical approach in their research, which stated that the intention to use technology in the financial service system for the micro, and small business sector was strongly influenced by performance expectancy. Therefore, the hypothesis proposed in this study: Hypothesis 1 (H1): Performance expectancy significantly affects behavioral intention.

Effort expectancy
The optimistic expectation is the level of ease associated with using the system (Venkatesh et al., 2003). This condition has captured three models such as perceived ease of use (TAM), complexity (MPCU), and ease of use (IDT). Most of the previous studies have discussed investigating consumer intentions to use. It has been found in the acceptance and use of information technology consumers (Khatimah & Halim, 2014). Sung et al. (2015) states that business expectations positively affect behavioral intentions of mobile learning services. The same thing was conveyed by Ghalandari (2012) on the acceptance model of E-Banking service technology; with the research results, the discussion concludes Effort expectancy has a dominant influence on behavioral intention. Wang et al. (2020) shows that performance expectations, effort expectations, facilitation conditions, and social impact positively and significantly influence consumers' behavioral intention to use and together accounted for 68.0% of the variance. Therefore, the hypothesis proposed in this study: Hypothesis 2 (H2): Effort expectancy significantly affects behavioral intention.

Social influence
In the case of application-based transportation, social impact is characterized as external influences promoting or affecting digital technologies, in this case, online transportation. The implications of social factors will help individuals adapt to the environment, including accepting new technology as an individual effort to survive in the existing environment (Venkatesh & Davis, 2000). The social atmosphere has various influences, including those of relatives or family, friend recommendations, the environment, commercials, user testimonials, and so on (Putri, 2018). Research by Sudarsono et al. (2021) using the theory of innovation diffusion approach states that social factors strongly influence the adoption of Islamic banking services. The same thing was expressed by Santoso & Nelloh (2017). It is heavily motivated by the advantages of social influences, with an orientation to the social sharing principle, which conveys the user's plan to use peer-topeer online transportation. Therefore, the hypothesis proposed in this study: Hypothesis 3 (H3): Social influence significantly affects behavioral intention.

Facilitating conditions
Facilitating conditions describe supporting facilities' availability from technology-based applications (Rathore, 2016). One of the important considerations in implementing application-based transportation services is the state of the supporting facilities. Without supporting facilities, it won't be easy to adopt new technologies in online transportation (Putri, 2018). Service units established by service providers for supporting facilities, such as operational equipment, awareness of usage, and customer assistance programs, are embodied in the manifestation of the facility's situation. Results of the discussion from the study Suzianti et al. (2018) said the intention to use the Gojek online application from the point of view of the Technology Acceptance Model (TAM) and Theory of Planned Behavior (TPB) is very dominantly influenced by technology facilitating conditions. In line with Aggelidis & Chatzoglou's (2009) research, they discovered that facilitating structures have a strong effect on behavioral intention by developing and testing the modified technology in the acceptance model. Therefore, the hypothesis proposed in this study: Hypothesis 4 (H4): Facilitating conditions significantly affect behavioral intention.

Hedonic motivation
The hedonism dimension may apply to the aesthetics and experience-based happiness extracted from the buying decision process, from identifying a need to post-purchase actions involving product or service use (Mort & Rose, 2004). The experiential view of hedonism broadly approaches purchasing and consumption processes (Rezaei & Ghodsi, 2014). Khatimah et al. (2019), using the UTAUT 2 approach to International Journal of Global Optimization and Its Application Vol. 1, No. 3, September 2022, pp.202-214. 205 using E-Money, states that hedonic motivation significantly affects behavioral intentions. Yeo et al. (2017) conveyed the same thing, which states that hedonic motivation significantly impacts behavioral intentions in online-based food delivery services. Salimon et al. (2017) stated the same about E-Banking adoption, which was heavily influenced by hedonic motivation. Therefore, the hypothesis proposed in this study: Hypothesis 5 (H5): Hedonic motivation has a significant effect on behavioral intention

Habit
Habit is one of the predictors in predicting behavioral intention to use a technology-based application (Putri, 2018). Verplanken et al. (1998) compared the Theory of Reasoned Action (TRA) and associated habit theory as predictors of behavioral intention, found that the positive effects of habits on behavioral intention outweighed the impact of attitudes and social norms. The same thing was conveyed by Rafique et al. (2020), which stated that the habit factor in the adoption of technology in library applications is expected to increase the continuation of user behavior in the future. Furthermore, Gefen et al. (2003) found that habit was a major factor in explaining the variance of continued website use, indicating the positive effect of habit on continued use of the same technology. Therefore, the hypothesis proposed in this study: Hypothesis 6 (H6): Habit has a significant effect on behavioral intention

Perceived Risk
Due to the high-risk level of uncertainties in using new technology, the trust factor is vital for users (Ha et al., 2021). In turn, the user's concern about payment security can influence their usage intentions and behavior (Zhou 2012). As a result, in addition to the non-cash payment mechanism, the JAKET application also includes an on-site payment system to help customers reduce their perceived risk. , from the Theory of Reasoned Action (TRA) approach, states that the perceived risk significantly affects the behavior of using mobile banking services in Vietnam. Lee said the same thing and that the intention to use online banking is subject to security/privacy risks. Therefore, the hypothesis proposed in this study: Hypothesis 7 (H7): Perceived risk significantly affects behavioral intention.

Materials and Methods
This study was conducted in Pematangsiantar City, North Sumatra, Indonesia. The strong reason for choosing this city is that the JAKET application can only be used in Pematangsiantar City. Its services are not yet available in other cities in Indonesia. The population in this study were all users of the JAKET application. Because the population is so large that it cannot be ascertained with certainty, the sample size is determined by multiplying the number of indicators from the eight variables by 5-10 (Augusty, 2006). Based on this provision, the sample size used is 33 x 5 = 165 people. This study uses a non-probability sampling approach with a purposive sampling technique. Using purposive sampling is because selecting samples based on the fulfillment of the research criteria so that they can provide answers that can support this research. One of the criteria used is an active user of the JAKET application with minimum use of it once a month. The data analysis stage includes the outer model analysis by analyzing the validity and reliability and the inner model analysis to test the hypothesis.
Furthermore, the researcher tested the hypothesis using variant-based Structural Equation Modeling (SEM) called Partial Least Square (PLS) and the SmartPLS version 3.0 application as a tool to analyze it. Measurement of variable components used surveys to measure respondents' perceptions. The measuring scale used is the 1-6 Likert scale from strongly disagree to strongly agree (1 = strongly disagree, 2 = disagree, 3 = quite agree, 4 = neutral, 5 = agree, 6 = strongly agree).
Study question items are oriented towards previous research results (See Appendix 1), namely for the construct of exogenous variables consisting of research-oriented performance expectations. Joshi (2018); Martins et al. (2014) comprise 4 items. The next construct is effort expectancy which is research-oriented (Lavenia, 2018;Zhou et al., 2010) and consists of 4 items. The social influences construct research-oriented (Kietzmann et al., 2011;Singh et al., 2020), consisting of 3 items. Facilitating conditions were researchoriented (Venkatesh et al., 2012;, consisting of 4 items. The construct of hedonic motivation is research-oriented (Putri, 2018) and comprising 6 items. Research-oriented by Putri (2018); Venkatesh et al. (2012) consist of 4 items in the habit construct. Furthermore, question items for the construct of perceived risk are research-oriented (Martins et al., 2014;Yang et al., 2012), which consists of International Journal of Global Optimization and Its Application Vol. 1, No. 3, September 2022, pp.202-214. 206 4 items. For endogenous variables, namely behavioral intention, it is research-oriented (Alalwan et al., 2017;Venkatesh et al., 2012), consisting of 4 items.

Results
On the basis of the demography profile of respondents in Table 1, it shows that the majority of JAKET application users are female, with a percentage of (53.33%). From the JAKET application's educational aspect, most users have a high school education (59.83%). From the professional element, students (35%) are the majority of the JAKET application users, with the orientation of using JAKET in a month ranging from 2-5 times (46.67%). Furthermore, for the frequency of use of the JACKET application, most consumers use it two to 5 times (46.67%) a month with an average expenditure of below 100,000 thousand rupiahs (63.33%).

Reliability and Validity Analysis
The feasibility test of the model is carried out to test how a set of latent construct indicators consistently explains each measurement. The reliability of the variables is assessed by Cronbach's Alpha and Composite Reliability values (Chin et al., 2003). The value of each reliability measurement can be accepted if it has a threshold value> 0.70. Furthermore, Convergent validity testing was determined by the loading factor and AVE in which the loading factor should be> 0.7, and the AVE value is 0.5 to meet convergent validity (Hair et al., 2014). Based on the test result's reliability and validity analysis (See Table 2), the reliability value for each latent construct in terms of Cronbach's Alpha and the Composite Reliability value has a threshold value> 0.7. The analysis of the validity value of each manifest variable in terms of the loading factor value also has a threshold value> 0.70. Vol. 1, No. 3, September 2022, pp.202-214. 207  The discriminant validity was tested for this analysis by comparing the average variance extracted square root with the constructs' correlation coefficient. According to Byrne & Van de Vijver (2010), it is acceptable to measure Discriminant Validity if the average variance extracted from the value of the two construction values is higher than the correlation square. The analysis revealed that all the unobserved square root variables of the 10 constructs were greater than the relationship between each pair of latent variables (See Table 2). Therefore, in this analysis, all latent variables' discriminant validity is well accepted and reasonable (Schaupp et al., 2010). The results of further analysis (See Table 2) obtained values. The Rsquare is 0.569 (56.9%), which shows the ability of exogenous variables to explain endogenous variables (behavioral intention) is moderate (Ghozali, 2014).

Hypotheses Test
A significance test was also used to evaluate the exogenous and endogenous variables' relationship to prove the hypothesis testing. The p-value revealed the relevance criteria. Suppose the p-value between the exogenous and endogenous variables is less than 0.05 at a significance range of 5%. In that case, the exogenous variables have a major impact on the endogenous variable. In contrast, if the value is higher than 0.05, the exogenous variables do not significantly develop the endogenous variable. The results of the hypothesis test are presented (See Table 4).

Discussion
This study indicated that performance expectancy positively and significantly affects behavioral intention. One of the conditions felt in the performance expectancy aspect is the ease of use of the JAKET application. Most female users claim to have a user experience that matches the application's performance expectations. The urgency of performance expectancy gives more confidence in users to affect the behavior of using the JAKET application. This study's results align with the studies conducted by  using the Theory of Reasoned Action (TRA) approach. The results show that performance expectancy significantly affects behavioral intention to use mobile banking services in Vietnam. The same thing was conveyed by Septiani et al. (2017) using the Theory of Planned Behavior (TPB) and Diffusion of Innovation (DOI) approaches, it shows that factors of internal perception (performance expectancy) have a significant effect on the behavioral intention of online transportation service.
Furthermore, positive and significant results were obtained for the effect of effort expectancy on behavioral intention. The performance of the JAKET application follows the user's expectations and desires. Applications that run on the JAKET application have simple features that make it easier for users because it does not take long to learn to use the JAKET application. In addition to its simple elements, this application has the advantage of delivering orders that can serve more than one shopping place, increasing the user's perception of the application. The results of this study are in line with the studies conducted by Indrawati & Yusliansyah (2017) by using the Technology acceptance model (TAM) approach. The results of the survey International Journal of Global Optimization and Its Application Vol. 1, No. 3, September 2022, pp.202-214. 209 show that effort expectancy has a significant effect on the behavioral intention of using smartphones as a medium for making non-cash transactions in business processes. Further studies conducted by Suzianti et al. (2018) conveyed the same thing by being oriented towards the Theory of Technology Acceptance Model (TAM) and Theory of Planned Behavior (TPB). They stated that the effort expectancy is considered high enough, so it is hoped that Gojek can maintain its current position by providing a touch of the latest innovations.
The following result shows that social influence has a negative and insignificant effect on behavioral intention. This research result's social influence factor is not the main factor for consumers to use application-based transportation, namely the JAKET application. If you look at the respondents' characteristics in this study, as many as 80% of the respondents came from the millennial generation, which, when related to the adoption theory, occupy the innovators and early adopters (Indrawati & Yusliansyah, 2017). The hallmark of this group is the courage to take risks in adopting new technologies. High curiosity causes this group to use new technology, which is expected to provide solutions to their needs. Therefore, social influence from the surrounding environment is not the main trigger for this group to use new technology. Likewise, based on the results of this study, the results were negative and insignificant for the effect of facilitating conditions on behavioral intention. This result is strengthened by measuring the validity test. The second measurement item for the facilitating conditions variable (FC2) obtained the lowest loading factor value compared to the loading factor for the other models' constructs in this study. It can be interpreted that the JAKET application's condition does not facilitate the needs of consumers, such as support or service compatibility, supporting facilities for the features provided, and the ease of obtaining and accessing internet facilities. Conditions that facilitate users of this application do not support implementing assessment procedures such as administrative, organizational, or technical support (Nikou & Economides, 2017).
Then for influence hedonic motivation against behavioral intention shows a positive and significant influence on behavioral intention. It means that the higher the acquisition of the application's convenience, the higher the motivation for use and will encourage consumer behavior to reuse it. The results of the discussion of this study are supported by the research of Venkatesh et al. (2012), which says Hedonic motivation is a critical determinant of behavioral intention and is considered a more crucial driver when compared to performance expectancy in a non-organizational context. Furthermore, Primasari (2016) concerned Digital Advertising, and hedonic motivation is the third-largest factor in the model tested in influencing behavioral intention. The results of subsequent research indicate that habit has a negative and insignificant effect on behavioral intention. It means that there are still not many consumer habits in using the JAKET application. The habit factor is unable to encourage user habits to use the application. It is because the JAKET application partners are still limited, and cooperation with culinary businesses is minimal compared to the Gojek and Grab applications. Furthermore, only about 28% of users often use the application if it is viewed from respondent characteristics. Most of the users come from entrepreneurial and other professional backgrounds.
Therefore, The jacket application developer needs to make further innovations on the application to reuse joy. These findings suggest that service providers should be positive in encouraging consumers to use online services and gradually eliminate the negative effects of habits (Lu et al., 2011). Meanwhile, the latest research results show that perceived risk positively and significantly affects behavioral intention. The chances of using the JAKET application as a payment system or personal data information are crucial for consumers. Perceptions of risk play an important role in encouraging consumer confidence in using these applications. Based on the hypothesis's results (See Table 4), perceived risk has the second largest value of coefficients after hedonic motivation in influencing behavioral intention to the JAKET application. It proves that user concerns regarding payment security and personal data information are important things for the application developer to pay attention to, so it is hoped that a good security system setup can affect their intentions and usage behavior. The results of this study are in line with the studies conducted by Silalahi et al. (2017), andYuliati et al. (2020) use the Technology Acceptance Model (TAM) approach states that a smaller risk will increase customer usage behavior.

Conclusions
This study uses the UTAUT 2 model on acceptance of JAKET application technology. This study indicated that UTAUT 2 affect the behavioral intention of the JAKET application, including performance expectancy, effort expectancy, hedonic motivation, and perceived risk. It provides important information for JAKET application developers to maintain and improve the quality of their applications. Furthermore, based on the discussion results, it turns out that social influence, facilitating conditions, and habits cannot influence the behavioral intention of the JAKET application. The discussion results emphasize that JAKET International Journal of Global Optimization and Its Application Vol. 1, No. 3, September 2022, pp.202-214. 210 application developers need to improve aspects of social influence, facilitating conditions, and habits to encourage the behavior of using the application on an ongoing basis. Furthermore, application-based transportation benefits the community as users in everyday life, which are very efficient and effective. This study has limitations in its location, which is limited to only one city, namely Pematangsiantar City. Furthermore, research in the field was only carried out once or was cross-sectional regarding data collection. Further research can be carried out in other cities with JAKET online transportation using other technology adoption models. Then for further research, it can be done by combining the Theory Combined TAM and TPB (C-TAM-TPB) approach as well as using the Innovation Diffusion Theory (IDT) approach, which comes from periodic primary data collection, for example, once every six months. Finally, further research can be carried out using a larger number of samples so that generalizations can be made about the acceptance of technology in the application-based transportation sector, especially for the JAKET application. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement: Not applicable.
Acknowledgments: The author would like to thank Sekolah Tinggi Ilmu Ekonomi Sultan Agung, Indonesia for supporting this research and publication. We would also like to thank the reviewers for their constructive comments and suggestions.

Conflicts of Interest:
The authors declare no conflict of interest.