CONSUMER BEHAVIOR IN ADOPTING USEETV GO STREAMING &; VIDEO-ON-DEMAND SERVICES USING UNIFIED THEORY OF ACCEPTANCE AND USE OF TECHNOLOGY 2 (UTAUT2)

One way to entertain yourself which has become a new lifestyle pattern is to watch streaming shows or subscription videos online. A McKinsey & Company survey at the end of March said that as many as 45% of respondents spent more on home entertainment during the pandemic. On the other hand, 85% of respondents reduced their spending on outdoor entertainment. Subscription video streaming services (video-on-demand / VoD) are one of the entertainment options that can be done at home. UseeTV GO is an Over-the-Top (OTT) TV and Video mobile application service, using the publicly accessible Internet. The content is available via smartphones or tablets with broadband connections. Currently, UseeTV Go app users are around 300 thousand active users every month, far below Vidio.com and Netflix. With the low Monthly Active of UseeTV Go's Video-on-demand app users in the pre-growth area far below, Netflix, Vidio.com, and user activeness based on time, UseeTV Go is in the Average Performer area. Therefore, UseeTV Go consumer assessment of the variables contained in the modified UTAUT2 model (Performance Expectancy, Effort Expectancy, Social Influence, Facilitating Condition, Hedonic Motivation, Price Value, Habit, and Content) was carried out on Behavioural Intention on UseeTV Go video-on-demand


INTRODUCTION
The increase in online activities is followed by increased spending to meet online needs, one of which is entertainment services. A Mckinsey and Company survey on May 20-22, 2020 of 715 respondents in Indonesia noted that spending on entertainment increased quite high during the pandemic 1 .As many as 37% of respondents spend more money on entertainment facilities at home. One of the most widely accessed means of entertainment is video-on-demand or VoD services. The service, which began to flourish in 2016, is an online video content provider system with a subscription payment mechanism. One of its appeals, the user has the freedom to decide what he wants to see 2 .
One way to entertain yourself which has become a new lifestyle pattern is to watch streaming shows or subscription videos online. A McKinsey & Company survey at the end of March 2021 said that 45% of respondents spent more on home entertainment during the pandemic 3 . On the other hand, 85% of respondents reduced their spending on outdoor entertainment. Subscription video streaming services (video-on-demand / VoD) are one of the entertainment options that can be done at home 4 .
UseeTV Go is the first interactive television service from Indihome in Indonesia in the form of a mobile application 5 . Pay television services that provide a new experience, viewers not only watch television but can also be in control as if they were the director.
In addition to providing quality shows, UseeTV Go also provides a variety of features that do not exist in other cable service providers, such as Pause & Rewind TV, Video on Demand, Video Recorder, and others 6 . Monthly Active users of the UseeTV Go Linear & Video-on-demand TV application are around 200-300 thousand per month or are in the pre-growth area far below, Netflix, Vidio.com, and user activeness based on time, UseeTV Go is in the Average Performer area 7 . This UseeTV Go user is a customer who buys Indihome TV and Internet packages. Of the 8 .3 million Indihome subscribers, there are 3 million users who subscribe to  Indihome TV 8 . Of the 3 million Indihome TV users, there are only 200-300 thousand  actively using the UseeTV Go application as an Indihome TV mobile extension 9 . Of these 200-300 thousand active users are loyal users spending an average of 1-3 hours a day. So the author examines the consumer behavior of UseeTV Go users when using the UseeTV Go application and whether users feel the influence on their performance, effort, motivation, habits, and the influence of the content. With the above conditions, the author examines and analyzes consumer behavior on Linear TV and video-on-demand UseeTV Go and conducts research entitled: "Consumer Behavior in Adopting UseeTV Go Services using the Unified Theory Of Acceptance And Use Of Technology 2 (UTAUT2)". By testing the effect of performance expectancy, effort expectancy, hedonic motivation, habit, and content have a positive and significant effect on users' behavioral intentions to adopt UseeTV Go services and test whether age and gender will moderate those influences 10 .

RESEARCH METHODS
UTAUT is a theory of the acceptance of the latest technology that was first developed by as a continuation of eight previous theories, namely the Theory of Reasoned Action (TRA) 11 , Theory of Planned Behavior (TPB), Technology Acceptance Model (TAM), Motivational Model (MM), Combined TAM-TPB (C-TAM-TPB), Model of Personal Computer Utilization (MPCU), Innovation Diffusion Theory (IDT), and Social Cognitive Theory (SCT). The UTAUT model was developed with four types of core constructs, namely performance expectancy, effort expectancy, social expectancy, and facilitating expectancy. In addition to the four constructs, there are four moderators, namely gender, age, experience, and voluntariness of use. In this theory, it is explained by behavioral intention and user behavior 12 .  Researchers modify the UTAUT2 model according to the needs of this study where researchers do not use the Use Behavior variable because researchers only identify Behavioral Intention. In addition, researchers also added the Content variable 17 , because according to researchers the Content variable is an important variable in Streaming & video-on-demand TV services because Streaming & video-on-demand TV is a service that sells content services over-the-top, so content is the main product in its sales. there is a moderator variable, the UTAUT2 model has three moderator variables, namely Age, Gender, and Experience 18 , but the researcher removed the Experience variable in this study because the Experience variable requires a periodic sampling data method, not through research in one period, but must be done periodically, where the researcher did not do the method in this study. Researchers removed Social Influence because UseeTV Go is currently an extension application of Indihome TV. Researchers also removed the Facilitating Condition because UseeTV Go is currently an extension application from Indihome TV which means it has an internet and mobile phones 19 . Researchers remove the price because the price or price of UseeTV Go is combined with the price of Indihome TV. Researchers removed Use Behavior because it has Active user data on the data application, so what is needed is a measurement of the extent to which someone will use UseeTV Go or Behavioral Intention 20 .

Figure 2. UTAUT2 Modified Model in This Study (Adapted from UTAUT2 model by 21
The following is an explanation of each variable in the model adapted from UTAUT2: 1. Performance expectancy: The level of user confidence that using a system will help users produce maximum work performance.
2. Effort Expectancy: The level of ease that users feel in using a system.
3. Hedonic motivation is the level of pleasure derived from using technology and has been shown to play an important role in determining the acceptance and use of technology. Fun and entertainment are indicators of hedonic motivation 22 .
4. Habit is to indicate the extent to which users tend to use the technology automatically due to previous learning with the habit of using technology as an indicator 23 .

Content is an important variable in video-on-demand services because UseeTV
Go is one of the video-on-demand services that sell content services over-the-top, so content is the main product in selling 24 6. Behavioral Intention is defined as the user's level of desire or intention to use the system continuously assuming that they have access to information. A person will be interested in using a new information technology if the user believes that using information technology will improve his performance, using information technology can be done easily, and the user gets the influence of the surrounding environment in using the information technology 25 .
This study used questionnaires as a primary data collection tool. The sample in this study uses non-probability sampling techniques and uses purposive sampling techniques, which means that sampling techniques use certain considerations so that the sample can present the population 26 . The questionnaire was disseminated using the Google Form online questionnaire. Respondents are obtained by distributing questionnaires through email and application notifications.
Researchers use 2 tests, namely testing measurement models, to test indicators against latent variables, or in other words testing how much an indicator (item) can explain the latent variable. Measurement model testing is carried out to meet the criteria for the validity of the goodness of data. Data collection tools used in research must meet validity criteria because the quality of research results will depend heavily on the data produced. Two stages must be carried out in the measurement model testing stage, namely 22 Venkatesh, Thong, dan Xu. 23 Venkatesh, Thong, dan Xu. 24 Jeon, Sung, dan Kim. 25 Venkatesh, Thong, dan Xu. 26 Prof. Dr. Sugiyono. The second test in PLS is the inner model or commonly called structural model testing (Assessment of the Structural Model), where the purpose of measuring this structural model is to test the influence between one variable and another. This test is done by looking at the t-value and path value to see whether the effect is significant or not. In addition, it can also look at the percentage of the variable described, namely by finding the value of R!, where the dependent latent variable modeled gets the influence of the independent latent variable. Interpretations of the results of R! are 0.67; 0.33; and 0.19; indicating that the models are "Good", "Moderate", and "Weak". In this study, R! of the dependent variable Behavioral Intention 28 .

RESULT AND DISCUSSION
The total respondents obtained in this study were 411 respondents, from a minimum sample level of 400 respondents. The sample criteria needed in this study are subscribers to the UseeTV Go service application with a minimum of 1 active time a week, where the characteristics of respondents are divided into age and gender categories.
The age range used in this study is the range 15-24 (categorized as youngage) and the range 25-60 (categorized as an adult), where the categorization of the range uses the international age classification used by the United Nations (UN) (United Nations, 2004) Of the total 411 respondents, it can be seen in Figure  4.1 and Table 4.1, that the percentage of age categories in this study, is dominated by the Young-Adult category (aged 15-24 years), with a percentage of 58.15%, where the Adult category (aged 25-60 years), has a percentage of 41.85%.
Based on gender, this study divides into 2 categories, namely Male and Female The percentage of UseeTV Go users are more male with a figure of 66.67%, with a total of 274 respondents, whereas the female gender is less with a percentage of 33.33%, with a total of 137 respondents. Researchers use measurement model testing, to test indicators against latent variables, or in other words test how much an indicator (item) can explain the Latin variable. Two stages must be carried out in the measurement model testing stage, namely the validity test and reality test, where three indicators must be done, namely convergent validity, discriminant validity, and reliability. Convergent validity is related to the principle where the indicator of a construct must have a high correlation 29 , explaining that convergent validity 27 Urbach N. 28  where if the loading factor is greater than 0.7, then the item is considered valid.  The loading factor value of each indicator in the Content variable in Table 4.7, shows the number so that all items are declared valid, in addition, the loading factor value ≥ 0,70, of the Behavior Intention variable, is shown in Table 4.8 below. From the entire table above, it can be stated that all indicators on each variable are valid because, for all indicators on each variable, the loading factor value shows a ≥ number of 0.70. For Moderate variables, Z1 (Gender) and Z2 (Age) as a whole have a loading factor value of 1.000 which shows a ≥ number of 0.70 so all variables are said to be valid.
In addition, another test to find out that items in latent variables meet the criteria of Construct Validity is to know the value of AVE (Average Variance Extracted). AVE is used to measure the sound of a unified variable or correlated variable, by comparing these variables with items to measure other variables in a model 30  After calculating AVE using SmartPLS 4.0 software, it can be concluded that all variables in this study meet the criteria of Convergent Validity because all AVE values in each variable are valued above 0.50.
In addition to convergent validity, discriminant validity is also a stage that must be carried out in testing measurement models, explaining that discriminant validity is needed to measure variables that are different from items used to measure other variables. The indicator in discriminant validity is the AVE value, if the square root value of each variable is greater than the correlation between the two variables in the model 31 .  Source : SmartPLS 4.0 (2023) Judging from the AVE root value in the table above, and comparing with the correlation between variables in mode, the variables in this study have discriminant validity, because the variables in this study have a root value (square root) of the AVE per variable, greater than the correlation between variables in the model.
In addition to testing validity, reliability tests must also be carried out to measure the internal consistency of measuring instruments 32 . Reliability tests are carried out using two methods, namely Cronbach's Alpha and Composite Reliability. According to 33 , Cronbach's Alpha measures the lower limit of the reliability value in a construct, while composite reliability measures the actual value of reliability in a construct. Cronbach's Alpha and Composite Reliability values recommended as benchmarks are 0.7. Based on all tests that have been carried out, from the validity and reliability of Cronbach's Alpha and Composite Reliability values as a whole greater than 0.7, this study has met the specified criteria. Therefore, researchers can conclude that all variables and indicators in this study are valid and reliable.

Structural Model Testing
The second test in PLS is the inner model or commonly called structural model testing (Assessment of the Structural Model), where the purpose of measuring this structural model is to test the influence between one variable and another. This test is done by looking at the t-value and path value to see whether the effect is significant or not. In addition, it can also look at the percentage of the variable described, namely by finding the value of R!, where the dependent latent variable modeled gets the influence of the independent latent variable. Interpretations of the results of R! are 0.67; 0.33; and 0.19; indicating that the models are "Good", "Moderate", and "Weak". In this study, the R! of the dependent variable Behavioral Intention is shown in the following From the value of R! found in Table 4.13, the value of R! in the dependent variable construct Behavioral Intention was influenced by 73.2% by the variables Performance Expectancy, Effort Expectancy, Hedonic Motivation, Habit, and Content while the influence of 26.8% was influenced by other factors, which were not studied in this study. The R! result also illustrates that the model is indicated "Good" because the R! value is greater than 0.67 which is included in the "Good" indication This study uses the one-tiled test hypothesis, so if the t-value ≥ 1.64, then Ha is accepted and because the confidence level is 95%, then the P value below 5% or 0.05 can be concluded that there is a significant influence of the independent variable on the dependent variable. Where the t value and P value are obtained by bootstrapping. The inner model path diagram in this study is shown in the following figure: The test results of the structural model, t values, and P values are shown in the Table below, where variables that do not affect positively and significantly are marked in red. This study uses the one-tiled test hypothesis, so if the t-value ≥ 1.64, then Ha is accepted, and because the confidence level is 95%, then the P value below 5% or 0.05 can be concluded that there is a significant influence of the independent variable on the dependent variable.

CONCLUSION
This study aimed to analyze the influence of Performance Expectancy, Effort Expectancy, Hedonic Motivations, and Habit on UseeTV Go's Behavioral Intention. Based on the test and analysis results, the researcher concluded that Performance Expectancy, Effort Expectancy, and Hedonic Motivations have a positive and significant influence on Behavioral Intention. However, Habit did not have a significant impact on Behavioral Intention. Suggestions were also provided, including strategies that competitors could implement to provide interesting video content and TV services for Indihome users and non-Indihome users. Additionally, gender was found to moderate the Content variable. Finally, the study's limitations were discussed, and further research was recommended, including examining the Experience variable longitudinally and researching what content is of interest to UseeTV Go users specifically.