Exploring Consumers' Acceptance of E-Marketplace Using TAM and Flow Theory

One business models of E-commerce which well-developed is E-Marketplace. The internal competition of E-Marketplace remains tough yet some achieve its successful compared to the others. The parameter of successful E-Marketplace can be seen by the numbers of acceptance by the costumers. Hence, this research on E-marketplace acceptance is significant to be done. This research takes a case study in one of the Indonesian markets, named the ‘Pusat Perbelanjaan Mentaya’ (PPM) regarding the use of the E-marketplace. This research aims to find out the adoption of e-marketplace acceptance. Our research model based on Technology Acceptance Model (TAM) which is modified by adding shopping enjoyment variable (Flow Theory). This research applies quantitative research method; spreading questionnaires and surveys. The data obtained were 240 from the PPM consumers who use E-marketplace application. Then the data would be used to examine the relationship between the variables of the proposed model. The results of this study show that the perception of the use of E-Marketplace application effects the perceived benefits. This due to the ease of operating the E-marketplace service affects users, moreover, sales and purchase transactions become more effective and efficient. In essence, the ease perceived also has an influence of the users’ attitude. This shows that E-marketplace service users feel that using these services increases effectiveness and benefits; they will not refuse to use E-marketplace services. The results in this study that the variable attitude towards use of Behavioral Intention is not accepted.


INTRODUCTION
The necessity to pursue the satisfaction in purchasing and selling has of ways. This is done in order to get the mutual satisfaction in the transaction. Shopping that is currently in demand is online shopping as it can be done anytime and anywhere.
For instance, e-commerce business models, that can be done anytime and anywhere, just like e-marketplace. E-marketplace is an internet-based online media which conducts business activities and transactions between buyers and seller (Putra et al., 2017). Good online communication will increase trust between sellers and consumers (Bao et al., 2016). One of the variables that can influence users to intend to use the E-market place is perceived usefulness (Setiyadi et al., 2019). E-market place needs to offer the same concept as traditional markets. This eases the sellers to meet their customers. The simplicity offered by the E-marketplace makes it faster known, many even describe the E-marketplace as a Department Store.
The expansion of the E-marketplace has almost dominated various parts of the world. The results of the Sharing Vision survey in 2018 stated that 92% of online consumers shop through the marketplace and 50% use social media. The number of internet users in Indonesia reached 88 million until almost the end of 2014, while the total population of Indonesia was around 253 million, hence, it is estimated that internet users reached 34.9% (Fachriyan & Wijaya, 2018). With the increasing number of internet users in Indonesia, it certainly benefits E-Marketplace owners in Indonesia such as Bukalapak and Tokopedia as well as other E-Marketplace holders. Seeing the increasing number of internet users in Indonesia, the opportunity to become an E-Marketplace holder is very possible and profitable. Even since the inception of the E-marketplace, it has provided opportunities for new entrepreneurs to start or expand their business (Wiradinata, 2017).
The high interest of E-Marketplace users in Indonesia makes Indonesian scientists want to know the extent of the role of the Indonesian people in accepting the development of online transactions. In fact, many studies mention that research on e-commerce in the world is currently an interesting topic to research (Farki & Baihaqi, 2016). Topics related to E-Marketplace as carried out by (Suryanto, 2018), (Pandy, 2018), (Lai et al., 2017), (Chong et al., 2018), (Fedorko et al., 2018), (Fayad & Paper, 2015), and (Al-Gahtani, 2011).
Sampit has a very fast economic development. Economic progress is supported by the easy entry and exit of various products, both through seaports and airports. In addition to these conveniences, in Sampit City there is a Shopping Center called Pusat Perbelanjaan Mentaya (PPM), often visited by people and visitors from outside the city to conduct business transactions. Progress in various fields is currently being matched by the development of information technology that can make it easier for everyone, such as in online transactions. Online meetings require a high level of trust because the seller and the buyer do not meet in person.
Based on the problems described above, it is necessary to know the level of public trust in online transactions. This can help develop an online transaction system in addition to the features required by customers. This research is intended to recommend the development of an E-Marketplace application in Sampit city. By conducting research on the community in Sampit and its surroundings regarding the level of trust in conducting online transactions. So that the development of applications can later meet the needs of sellers and buyers on the side.

METHODS
The phases of the research carried out are problem identification, literature study, formulating problems obtained in the identification process, designing models, compiling questionnaires based on interviews and literature studies, testing the validity and reliability of questionnaires, data collection and model validation, analysis of results, conclusions and suggestion.

Sampling and data collection
This research object is the e-marketplace in Sampit and the subject is the user, the data source in this research comes from online and offline surveys. Online samples are conducted through online forums by spreading online questionnaire by google form and offline samples are conducted by distributing questionnaire at Pusat Perbelanjaan Mentaya Sampit. The number of sample size is calculated using Hair formula collecting 224 valid responses from a total of 240 responses, hair formula is used because the population size is not known with certainty.

Instruments and measures
The measurement instrument in this research is the surveys by spreading online and offline questionnaire, the questionnaire includes three sections. The first section had demographic questions about the respondents. The second section asked respondents what kind of e-marketplace they had used and the third section had questions measuring the variables in the proposed model. The questions given in the third section is about perceived ease of use(PEOU), perceived usefulness(PU), attitude toward using(ATU), behavioral intention(BI), actual usage(AU) and shopping enjoyment(SE). The questionnaire scale is using likert scale from 1(Strongly Disagree) to 5(Strongly Agree). Research data is processed and tested using Structural Equation Model (SEM-PLS) and WarpPLS as a tool, there are two stages of model analysis in SEM-PLS which is : measurement model or the outer model and structural model or inner model.

Conceptual Model
The model used in this research is TAM and flow theory, there are six variables in this model; perceived ease of use, perceived usefulness, attitude toward using, behavioral intentions. Intention, actual usage and shopping enjoyment.

General description of the characteristics of the respondent
Respondents in this study are general consumers who have made transactions at the Pusat Perbelanjaan Mentaya (PPM), which is one of the markets in Indonesia. The number of respondents obtained in this study was 240, after sorting there were 224 respondent data that were suitable for analysis. Respondent data to be analyzed is obtained from various characteristics, for example gender, age, occupation, city, ever used e-marketplace, e-marketplace used, how many times a month use e-marketplace, how many hours a month visit e-marketplace, and how often to visit e-marketplace.
In essence to obtaining information about the demographics of the respondents, this research also receieves the information about respondents' experiences in using the E-Marketplace. Based on the data obtained by 240, it is known that 224 respondents have used the e-marketplace. It also obtained the data on the types of e-marketplace used by respondents; 110 respondents used Shopee, 16 respondents used shopee, 10 respondents used Lazada, 5 respondents used Bukalapak, and 18 respondents used other types of e-marketplace. Other respondents have used various types of e-marketplaces.
The research results also obtain information related to the frequency with which someone uses e-marketplace services. It is known that the majority of respondents visit the e-marketplace every few months, namely as many as 61 respondents. It can be concluded that the majority of respondents visit e-marketplaces to shop or look for products / services only when they are in need. Then obtained information about the length of time the respondents visited the e-marketplace on average less than one hour. As seen from how often respondents visit the e-marketplace, it is known that 86 respondents use the e-marketplace normally.

Processing of Survey Results
The result of data collection, shows that Perceived Ease of Use and Perceived Usefulness has the most positive responses., it has mean value of 4.18. Attitude Toward Using variable has the smallest response. A description of the mean value of each variable can be seen in Table 2.

Measuring Model Assessment
To quantify the model, assessing the connection among indicators and latent variables by evaluating reliability and validity. Reliability was evaluated by composite reliability and Cronbach alpha. While validity was analyzed in two components, that is convergent validity and discriminant validity. In table 3, it is known that the p value is significant, which is less than 0.05 and of all loading indicator values there is a loading indicator value that is less than 0.60. In the other cases, according to Sholihin, the requirements for loading values above 0.60 are often not fulfilled, especially in the case of newly developed questionnaires, therefore, the loading indicator value still needs to be maintained. According to Hair indicator with loading below 0.60 it should be removed from the model. However, the decision to remove indicators with loading values below 0.60 should first analyze the impact. This indicator can be removed if it can increase AVE and composite reliability above the limit. From the above statement, it is considered to maintain the indicator with a loading value below 0.60 because the AVE results and the composite reliability indicator with a loading value of 0.60 meet the acceptable threshold. In table 4, the analysis results provide information that the AVE value has met requirements above 0.50, as well as the composite reliability value and Cronbach alpha. There is other way to analyze discriminant validity; by looking at the results of the calculation of the latent variable correlations.  Table 5 shows that The AVE value between latent variables in the non-diagonal column is lower than the AVE value in the diagonal column.. The statement above shows that the discriminant validity is fulfilled. In conclusion, it shows that the results of the measurement model have good validity and reliability values. Figure 3 shows the result of fit model index and quality indices test. The data test produces a significant value because it can be seen that the p value on APC, ARS and AARS is less than 0.0001 and the AVIF value has a value of less than 5. Figure 4 shows result that relationship between ATU and BI is not significant, because the p value of each of these variables is more than 0,05. The complete model evaluation results can be seen in table 6 below. The SEM-PLS test results with WarpPLS also produce the Effect Size value. Effect Size is calculated from the absolute value of the individual contribution of each latent predictor variable to the R-squared value of the latent criterion variable (Cohen, 1988). According to Kock and Hair in their research cited by (Sholihin & Ratmono, 2013) effect size can be grouped into three categories; weak (0.02), medium (0.15), and large (0.35). The effect size value below 0.02 indicates that the influence of the latent predictor variable is very weak from a practical point of view even though it has a significant P value. Table 6 above summarizes the effect size of the latent predictor variables on the latent criterion variable.

Figure 4 The SEM-PLS structural model test results
[H1] -Perceived Ease (PEOU) is related positively to the Perceived Usefulness (PU). The results of the Model Test in table 6 present the relationship between the PEOU variable and the PU variable, which has a path coefficient value of 0.763 and a p value <0.001. These results indicate that the two variables have a significant relationship. Hence, hypothesis 1 in this study is accepted.
In testing the structural model it also shows that the relationship between the predictor variables for the perceived ease of using the e-marketplace and the criterion variable for the perceived benefits of e-marketplace has an effect size of 0.582, this value indicates that the perceived ease of using a large e-marketplace variable from a practical point of view has a P value significant.
[H2] -Perceived Ease (PEOU) is related positively to the attitude toward using (ATU). In Table  6, the Model Test Results presents the relationship between the PEOU variable and the ATU variable, which has a path coefficient value of -0.165 and p value <0.001. These results indicate that the two variables have a significant relationship. Thus Hypothesis 2 in this study is accepted. In testing the structural model it also shows that the relationship between the perceived ease of using e-marketplace predictor variables with the criterion variable of the attitude of using emarketplace has an effect size of 0.073, this value shows that the perceived ease of using emarketplace variable is weak on a practical point of view and has a P value significant.
[H3] -Perceived Usefulness (PU) is related positively to the attitude toward using (ATU). In Table  6, the Model Test Results presents the relationship between the PU variable and the ATU variable, which has a path coefficient value of -0.38 and a p value <0.001. These results indicate that the two variables have a significant relationship. Thus Hypothesis 3 in this study is accepted. In testing the structural model it also illustrates that the relationship between the predictor variables for perceived benefits using e-marketplace and the criterion variable for the attitude of using emarketplace has an effect size of 0.190, this value shows that the perceived benefit variable using e-marketplace is medium on a practical point of view and has a P value. significant.
[H4] -Perceived Usefulness (PU) is related positively to the behavioral intention (BI). In Table 6, the Model Test Results presents the relationship between the PU variable and the BI variable, which has a path coefficient value of 0.037 and a p value <0.001. This result shows that the two variables have a significant relationship. Thus Hypothesis 4 in this study is accepted. In testing the structural model it also shows that the relationship between the predictor variables for perceived benefits using the e-marketplace and the behavioral intention criterion variable has an effect size of 0.026, this value shows that the perceived benefit variable using the e-marketplace is weak in practical terms and has a significant P value.
[H5] -Attitude toward using (ATU) is related positively to the Behavioral intention (BI). In table 6, the Model Test Results presents the relationship between ATU and BI variables, having a path coefficient value of -0.027 and a p value of 0.41. This result shows that the two variables have an insignificant relationship. Thus, hypothesis 5 in this study is not accepted. Then the structural model test shows the relationship between the ATU variable and the BI variable has an effect size of 0.009. This value indicates that the variable relationship is not significant.
[H6] -Behavioral Intention (BI) is related positively to the Actual Usage (AU). The Model Test Results presents the relationship between the BI variable and the AU variable, which has a path coefficient value of 0.66 and a p value <0.001. This result shows that the two variables have a significant relationship. Moreover, hypothesis 6 in this study is accepted. In testing the structural model, it also shows that the relationship between the BI predictor variable and the AU criterion variable has an effect size of 0.952, this value shows that the BI variable is large from a practical point of view and has a significant P value.
[H7] -Shopping Enjoyment (SE) is related positively to the Behavioral Intention (BI). The Model Test Results presents the relationship between the SE variable, has a path coefficient value of 0,968 and a p value <0.001. The result shows that the two variables have a significant relationship. Moreover, hypothesis 7 in this study is accepted.  Hypothesis 1 is known that there is an effect of perceived ease of use (PEOU) on perceived usefulness (PU), this is due to the fact that the ease of operating e-marketplace services affects the results obtained from these uses. Ease of operating e-marketplace services affects users so that sales and purchase transactions using e-marketplace services become more effective, faster. This result is in line with research (Lu et al., 2009), (Pavlou, 2003) which states that convenience perceived ease of use (PEOU) has a very significant effect on perceived usefulness (PU).

Findings
Hypothesis 2 is proven that perceived ease of use (PEOU) has an influence on attitude toward using (ATU). The results of this study indicate that e-marketplace service users feel that using e-marketplace services will increase the effectiveness and efficiency of the transactions they make, such as sales transactions, purchases, search for products or services according to what they want. Therefore, when users feel that using e-marketplace services is useful, comfortable, and provides convenience, they do not refuse to use e-marketplace services. The results of this study support previous research conducted by (Adams et al., 1992), (Igbaria et al., 1997), and(Burton-Jones &Hubona, 2005) suggest that perceived ease of use has an influence on attitude toward using. In addition, this imperis study is in line with other research conducted by (Venkatesh & Davis, 2000), (Bruner & Kumar, 2005), and (Kulviwat et al., 2007) and also suggests that there is a positive influence on perceived ease of use (PEOU) on attitude toward using (ATU).
In hypothesis 3 testing, it can be seen that the perceived usefulness (PU) that have a positive effect on attitude toward using (ATU) are accepted, this shows that when users of e-marketplace services feel that the use of these services increases the effectiveness and efficiency of their business transactions, they will have an accepting attitude in using the e-marketplace service. Consumers in Pusat Perbelanjaan Mentaya feel convenient and happy when using e-marketplace services because it is easier for them to make purchases, sales and find and promote their services or products. Others, if consumers feel that e-marketplace services are not useful, then consumers will not like or reject e-marketplace services to help with their business transactions. The results of this study support previous research by (Kurnia & Chien, 2003), (Pavlou, 2003), and (Lu et al., 2009).
The test results of Hypothesis 4 from the results of data processing, it is known that the path coefficient value in table 6 above is 0.68 and p value <0.001, these results indicate that both the perceived usefulness variable (PU) and the behavioral intention variable (BI) have a relationship. significant. The majority of respondents stated that e-marketplace services that provide benefits when used will encourage respondents to continue using e-marketplace services in the future. The results of this study also support previous research by (Lu et al., 2009), (Lou et al., 2005), and (Setiawan, 2015) which states that perceived usefulness (PU) have a significant effect on behavioral intention (BI).
On hypothesis 5 test results, the variable attitude toward using (ATU) affects the variable behavioral intention (BI) the results are declared not accepted. This is because the attitude in using e-marketplace services does not determine user behavior in the future. If the user's attitude towards e-marketplace services tends to be negative, then users will certainly not use the e-marketplace service in the future and look for other alternatives to replace existing e-marketplace services. This is in line with research (Widyapraba et al., 2016) which states that attitudes in the use of technology have no influence on user behavior in using this technology in the future. This is in stark contrast to research (F. D. Davis, 1989), (F. D. Davis et al., 1989), and (Rodrigues Pinho & Soares, 2011) which states that attitude to use affects the intention to use.
Hypothesis 6 The SEM-PLS calculation results on the inner model or structural model, the relationship between behavioral intention (BI) e-marketplace services and actual usage (AU) of emarketplace services has an effect size of 0.436, this value shows that the behavioral intention variable is very high. large on a practical point of view and has a significant P-value. The conclusion is that the results of this test are accepted, this is supported by research (Lu et al., 2009). On Hypothesis 7 test results, the variable shopping enjoyment(SE) affects the variable behavioral intention(BI) the result declared accepted.

CONCLUSION
This is a study on investigating the acceptance factors that influence user to use an emarketplace applications, this study suggest that the important factor towards user decisions to use an e-marketplace applications influenced by their perceptions on whether the applications is easy to use (perceived ease of use) and will give benefits or not(perceived usefulness). User perceptions on ease of use of an e-marketplace relate to an individual's perceptions on the application navigation and the ability to complete their task effortlessly. User perceptions on usefulness to use an e-marketplace relate to the online services may save energy and time. These findings recommend e-marketplace designers and developers should focus on how to make the emarketplace applications to be easy to use and useful for users. The variables in this study is limited, for the further research to explore other relevant variables to be added to enrich the emarketplace acceptance.

ACKNOWLEDGMENT
We can complete the research on the perception of e-marketplace acceptance; hence we would like to thank all those who supported the implementation of this research. The research was carried out well according to the predetermined methodology. Hopefully the results of this research can add scientific repertoire in the field of information technology. In essence, it is hoped the results of this research can be useful for the progress of Indonesia, especially advances in technology in Sampit.