Determinant Factors Purchase Decision of Tokopedia with Brand Image as Intervening Variable

This study aims to examine the relationship between variables, especially variables brand ambassador, brand awareness, and brand image as an intervention. The data collection method in this research is purposive sampling method. Questionnaires were distributed to 131 respondents. The results of the questionnaire obtained were studied using path analysis. The results show that only brand awareness has no indirect influence between awareness of the purchase decision process


INTRODUCTION
Indonesia nowadays is currently entering the era of globalization and digitalization.The activities carried out by Indonesian people today are not far from technological and digital elements.The easy transfer of information and data that is very fast causes people's behavior to slowly start to change along with technological developments.At the end of the 20th century, the internet emerged as a medium of information that could not be separated from people's lives and its presence had eliminated space and time boundaries for business and communication activities.As for (gross merchandise value) GMV e-Indonesian commerce in 2015 reached US$ 1.78 billion.Then in 2019 projected grew by (88% CAGR) to US$ 20.9 billion and will increase (48% CAGR) to US$ 82 billion or around IDR 1,148 trillion.This value is equivalent to half of the ecommerce market share in the ASEAN region.
The development of the internet especially in Indonesia itself for many parties it is an opportunity for those who want to engage in online business, thus causing new habits in all aspects of life.The rise of online shopping in society in recent years has encouraged the growth of the logistics sector, especially in Indonesia.This can be seen in a survey conducted by Merchant Machine regarding the fastest growing e-commerce in the world in 2018.
The rise of e-commerce, social media, as well as the presence of shopping services from internet-based transportation services are able to change the way people shop from offline to online.APJII and the Indonesian poll conducted a survey stated that the number of internet users in Indonesia on 2018 increased from 27.91 million (10.12%) to 171.18 million people.This means that the penetration of internet users in Indonesia has increased to 64.8% of the total population of 264.16 million people.The statement supported by the existence of e-commerce sales in Indonesia which grew by 133.5% to US$ 16.5 billion or around Rp. 219 trillion in predictions for 2022.This growth prediction does not rule out the possibility because of the magnitude of technological advances that provide convenience for shopping for consumers from all walks of life.The data supported by e-marketers who prove that there is growth in ecommerce transactions in Indonesia.Namely, the growth of existing transactions reached Rp.Huge open innovation opportunities and shifts in people's lifestyles are one of the strengths of the ecommerce business.Currently, most e-commerce transactions are still carried out using desktops or laptops.From several existing e-commerce sites in Indonesia, according to data from the iPrice Group, Tokopedia was named the e-commerce site with the largest number of monthly web visitors in the third quarter of 2019.It was noted that the total number of Tokopedia web visitors in the third quarter of 2019 was 66 million visitors.With the next ranking filled by Shopee with 56 million visitors and Bukalapak with 43 million visitors.
To carry out the latest strategy, it was recorded on October 7 2019 through the official website Tokopedia.comthat it announced the use of the famous boy group, namely BTS as a brand ambassador.The concept of idolizing and buying things related to them, such as the clothes they wear, or the snacks they usually consume, just like Shopee, which had previously targeted Christiano Ronaldo and the girl group Blackpink as the face of its brand, is perhaps what made Tokopedia finally use a boy group.from South Korea, namely BTS as its brand ambassador.
BTS is a boy group that has achieved success, both in South Korea and internationally.The boy group under Big Hit Entertainment consists of seven personnel namely Kim Nam-joon, Min Yoon-gi, Jung Ho-seok, Kim Seok-jin, Park Ji-min, Kim Taehyung and Jeon Jung-kook.It is stated on the official website of Tokopedia.comthat this celebrity figure was chosen because both the vision and the messages brought by BTS were able to be consistently conveyed and aligned with the vision of Tokopedia.
When viewed in terms of the growth of online shopping activity in the next few years as previously explained, it is hoped that Tokopedia's move to appoint BTS will be able to increase sales rates and active users on Tokopedia.Especially the BTS boy group, with millions of fans in Indonesia and dominated by young people, Tokopedia has a great opportunity to attract more customers and new users.This is supported by APJII data in 2018 which noted that active internet users are at millennial age, namely 20-24 years with a penetration of 88.5%.Then below that there is an age group of 25-29 years with a penetration of 82.7%, an age group of 30-34 years with a penetration of 76.5%, and an age group of 35-39 years with a penetration of 68.5%.Therefore, researchers are interested in knowing what factors are considered by users of the Tokopedia application.

Reasearch Purpose
Referring to the formulation of the problem, the formulated research objectives are as follows: 1.To measure the effect of the brand ambassador variable (X1) on brand image (Y) 2. To measure the influence between brand awareness variables (X2) on brand image (Y) 3. To measure the effect of brand ambassador (X1) and brand awareness (X2) on brand image (Y) simultaneously 4. To measure the effect of the brand ambassador variable (X1) on the product purchasing decision process at Tokopedia (Z) 5. To measure the influence between brand awareness variables (X2) on the product purchasing decision process at Tokopedia (Z) 6.To measure the effect of the brand image variable (Y) on the product purchasing decision process at Tokopedia (Z)

METHODS
This research is a descriptive quantitative research.In this study, researchers used the Likert scale measurement method.The population in this study are Indonesian people who use Tokopedia and know and are exposed to advertisements with Tokopedia's brand ambassadors, namely BTS.In this study, researchers used 131 respondents who used the Tokopedia application and knew and were exposed to advertisements with the Tokopedia brand ambassador and taken by purposive sampling.
This research is intended to build a real picture of a phenomenon that is in the context of research.With descriptive research using the IBM SPSS version 26 program, various information will be collected in order to answer the questions in this research.

Validity Test
The results show that the brand ambassador variable, brand awareness and brand image has valid criteria for all statement items, namely based on the criterion value of Sig <0.05 and r count > r table 0.361.Whereas on variable process decision purchase there is 2 statements that do not have valid criteria, namely PKP statements 2 and PKP 4. Because of this, so that the research can be continued, the researcher conducted a retest by removing the statements on the PKP 2 and PKP 4 indicators.
After retesting by removing statements on indicators.The results of PKP 2 and PKP 4 show that out of the 38 statements given to 30 respondents, the sig value < 0.05 and r count > r table 0.361.It can be concluded that all statement items are valid.

Reliability Test
To determine whether the correlation coefficient is significant or not, the researcher uses the distribution (r table) for alpha 0.05 with degrees of freedom (dk = n -2).According to Nurhasanah (2007), in her explanation it states that if Cronbach Alpha gives a value > 0.70 then the research instrument is considered reliable.
Reliability test in this research shows that all variables have Cronbach's value Alpha above 0.70.So it can be concluded that all variables are declared reliable.

Classic Assumption Test
The magnitude of the value Kolmogorov-Smirnov in the structure normality test I was 0.068 with a significance of 0.200.Because the Kolmogorov-Smirnov significance value is 0.200 > 0.05.So it can be concluded that the data is normally distributed.The magnitude of the value Kolmogorov-Smirnov in the structure normality test II is 0.053 with a significance of 0.200.Because the Kolmogorov-Smirnov significance value is 0.200 > 0.05.Then got concluded that the data is normally distributed.
The significance value obtained is 0.200 (> 0.05) then Ho is accepted, therefore it can be concluded that the residuals in this study have met the assumption of normally distributed data because the significance value is > 0.05.According to Imam Ghozali ( 2013) the multicollinearity test aims to test whether the regression model found a correlation between independent (independent) variables.
The results of processing the multicollinearity test data using SPSS 26 show that the two independent variables do not occur multicollinearity between the independent variables so that it can be stated that a simple linear regression model can be carried out in this study.
Based on the heteroscedasticity test, it shows that the dots are not clustered, but spread out and the distribution is not patterned.It can be said that the independent variable is free from the classical assumption of heteroscedasticity.Based on the processing results in table 2 the path coefficients obtainedis ρyx1X1= 0.175, ρyx2X2= 0.387, ρzyY = 0.360.To search for errors (є2) namely by looking at the R square in the model summary table.The formula to get the residual coefficient is 1.00 -R square.So the error (є2) is = 1.00 -0.647 = 0.353.So, the path analysis equation that is formed is as follows: Z = ρzx1X1+ρzx2X2+ ρzyY + ϵ2 Z = 0.175X1 + 0.387X2 + 0.360Y + 0.353ϵ2 Purchase decision = 0.175 Brand ambassador + 0.387 Brand awareness + 0.360 Brand image + 0.353ϵ2

Hypothesis Testing t
In table 2 the results of the t test can be seen by the two probability values Sig.< 0.05 and t count > t table 1.980 obtained from df = n-2 where df = 131-2 = 129 (closer to df = 120).In the IBM SPSS calculation results in the form of the Coefficients table, it is obtained: 1.There is a relationship between the brand ambassador (X1) and the purchasing decision process (Z).This can be seen in the probability value Sig.0.024 <0.05 and t count 2.286 > t table 1.980.So the hypothesis is that Ha is accepted and H0 is rejected.This means that partially there is a significant influence between brand ambassadors on the purchasing decision process.2. There is a relationship between brand awareness (X2) and the buying decision process (Z).This can be seen in the probability value Sig.0.000 <0.05 and t count 4.894 > t table 1.980.So the hypothesis is that Ha is accepted and H0 is rejected.This means that partially there is a significant influence between brand awareness on the purchasing decision process.
3. There is a relationship between brand image (Y) and the purchasing decision process (Z).This can be seen in the probability value Sig.0.000 <0.05 and t count 5.240 > t table 1.980.So the hypothesis is that Ha is accepted and H0 is rejected.This means that partially there is a significant influence between brand image on the purchasing decision process.Tester Between Sub Variables In table 4.14 the results of the F test can be seen the probability value of Sig. of 0.000 <0.05 and F 44.885 > F table 2.68 which is obtained from df = (k-1) (n-k) where df = df 1 = 4-1 = 3 and df2 = 131-4 = 127.Then the result is significant, that is, Ha is accepted and H0 is rejected.This can be interpreted that there is influence between brand ambassadors (X1), brand awareness (X2), and brand image (Y) simultaneously In table 5 the results of the F test can be seen the probability value of Sig. of 0.000 < 0.05 and F 77.513 > F table 2.68 obtained from df = (k-1)(n-k) where df = df 1 = 4-1 = 3 and df2 = 131-4 = 127.So the result is significant, that is, Ha is accepted and H 0 is rejected.This can be interpreted that there is influence between brand ambassadors (X1), brand awareness (X2), and brand image (Y) on the purchasing decision process (Z) simultaneously.Based on the calculation results above, it is known that the value of the indirect effect is greater than the value of the direct effect.These results prove that indirectly both brand ambassador (X1) and brand awareness (X2) through brand image (Y) have a significant influence on the purchasing decision process variable (Z).This value indicates that the influence of brand ambassador (X1), brand awareness (X2), on brand image (Y) simultaneously is 41.2%.While the remaining 58.8% obtained from 100% -41.2% are other aspects that have an influence on the brand image variable (Y).This value is also the standard error (є1) of the brand image (Y).This value indicates that the effect of brand ambassador (X1), brand awareness (X2), brand image (Y), on the buying decision process (Z) simultaneously is 64.7%.While the remaining 35.3% obtained from 100% -64.7% are other aspects that have an influence on the purchasing decision process variable (Z).This value is also the standard error (є1) of the buying decision process (Z).

The influence of brand ambassadors (X1) on brand image (Y)
The test results in this study state that brand ambassadors have a significant effect on brand image and have a positive value.These results can be seen in the significant value of 0.003 which is smaller than the significant level of 0.05 or the calculated t value which is greater than the t table (2.999 <1.980) and has a positive beta value of 0.285.So that the hypothesis taken is that Ha is accepted and H0 is rejected.This is in line with previous research by Raswen (2016) which stated that brand ambassadors have a positive effect on brand image.In his research, he stated that brand ambassadors are able to help and create stronger emotional relationships between brands or companies and consumers.This can be seen when the announcement of the use of brand ambassadors at Tokopedia made the hashtags for BTS and Tokopedia rank first on Twitter Indonesia.This statement is reinforced by Yolanda's research ( 2017) by proving that brand ambassador variables affect brand image.For him, companies must continue to use brand ambassadors who have good physical appearance and have many talents who can always motivate consumers.

Effect of brand awareness (X2) on brand image (Y)
The test results in this study state that brand awareness has a significant effect on brand image and has a positive value.These results can be seen in the significant value of 0.000 which is smaller than the significant level of 0.05 or the calculated t value which is greater than the t table (4.316 <1.980) and has a positive beta value of 0.410.So that the hypothesis taken is that Ha is accepted and H0 is rejected.
If a brand is in the minds of consumers, the brand will be considered for selection.The role of brand awareness in purchasing decisions depends on the degree to which brand awareness is achieved in the minds of consumers.This is in line with previous research by Tariq, et al. (2017) which states that brand awareness has a positive effect on brand image.It is known that the higher the brand awareness and brand image, the higher the consumer purchasing decisions.This statement is reinforced by the research of Rakhma, et al. ( 2019) by mentioning the same thing.

The influence of brand ambassadors (X1) on the purchasing decision process (Z)
The test results in this study state that brand ambassadors have a significant effect on the purchasing decision process and have a positive value.These results can be seen in the significant value of 0.024 which is smaller than the significant level of 0.05 or the calculated t value which is greater than the t table (2.286 <1.980) and has a positive beta value of 0.175.So that the hypothesis taken is that Ha is accepted and H0 is rejected.This is in line with previous research by Yolanda (2017) which stated that brand ambassadors have a positive effect on the purchasing decision process.His research proves that brand ambassadors have an effect on consumer purchasing decisions.Basically, to increase consumer attention, one of the companies is using well-known celebrities as bait to help consumers in making a decision.This statement is in accordance with Tokopedia when collaborating with indomilk at WIB (Indonesia Shopping Time).It is known that because the BTS brand ambassador was so strong, a day after giving an advertisement for indomilk banana milk packaging with Tokopedia's brand ambassador at that time, namely BTS, it was discovered that in just a few days on the official indomilk website at Tokopedia it had sold out.And this is reinforced by the research of Gloria, et al. (2019), namely mentioning the same significant results.4. Effect of brand awareness (X2) on the purchasing decision process (Z) The test results in this study state that brand awareness has a significant effect on the purchasing decision process and has a positive value.These results can be seen in the significant value of 0.000 which is smaller than the significant level of 0.05 or the calculated t value which is greater than the t table (4.894 <1.980) and has a positive beta value of 0.387.So that the hypothesis taken is that Ha is accepted and H0 is rejected.This is in line with previous research by Cahyani (2016) which stated that brand awareness has a positive effect on the purchasing decision process where during the decision making process consumers tend to choose products and brands that they are aware of or remember.This is an important consideration set, because a brand that is not part of the consideration will not be selected.In other words, brands that have a high top of mind for Tokopedia have the highest probability of being chosen by consumers.This statement is reinforced by Fariha's research (2019) by mentioning the same significance in the buying decision process.

Effect of brand image (Y) on the purchasing decision process (Z)
The test results in this study state that brand image has a significant effect on the purchasing decision process and has a positive value.These results can be seen in the significant value of 0.000 which is smaller than the significant level of 0.05 or the calculated t value which is greater than the t table (5.240 <1.980) and has a positive beta value of 0.360.So that the hypothesis taken is that Ha is accepted and H0 is rejected.This is in line with previous research by Yolanda (2017) which stated that brand image has a positive effect on the purchasing decision process.In his research, he stated that the better the brand image of a product or service, the stronger consumer confidence in the product and encourages the desire to make a purchase decision.This is the same as what is being done by Tokopedia, where currently by providing a campaign for WIB (Indonesian Shopping Time) every 25th until the end of the month to give confidence to consumers that Tokopedia will not lose when it has big discounts.For him the brand image of a product will determine consumer perceptions in assessing and making a purchasing decision.This statement is reinforced by research by Gloria, et al. ( 2019) with a significant mention of the same thing.6. Indirect influence of brand ambassador (X1) on purchasing decision process (Z) through brand image (Y) The test results in this study stated that in the sobel test the brand ambassador variable on the purchasing decision process through brand image was 2.56810.Because the obtained t value is 2.56810> t table is 1.98, with a significant level of 0.05.So it can be concluded that Ha is accepted and H0 is rejected.This means that the brand ambassador (X1) has a significant effect on the purchasing decision process (Z) through brand image (Y).So it can be concluded that there is an indirect effect in the form of mediation between brand ambassadors on the purchasing decision process (Z) of products at Tokopedia.This is in line with previous research conducted by Widjaja (2015) which stated that brand ambassadors have a significant effect on the purchasing decision process through brand image and is strengthened by Larasari et al.'s research. (2018) by mentioning the same thing.In implication, it is true that purchasing decisions when using the Tokopedia brand ambassador were different from before the announcement of the Tokopedia brand ambassador.The concept of idolizing and buying things related to them, such as the clothes they wear, or the snacks they usually consume, just like Shopee, which had previously targeted Christiano Ronaldo and the girl group Blackpink as the face of its brand, is perhaps what made Tokopedia finally use a boy group.South Korean origin, namely BTS. 7. The indirect effect of brand awareness (X2) on the purchasing decision process (Z) through brand image (Y) The test results in this study stated that in the Sobel test the brand awareness variable on the purchasing decision process through brand image was 0.7324056.Because the obtained t value is 0.7324056 < t table 1.98, namely with a significant level of 0.05.So it can be concluded that H0 is accepted and Ha is rejected.This means that the brand ambassador (X1) has no significant effect on the purchasing decision process (Z) through brand image (Y).So it can be concluded that there is no mediating effect between brand awareness and the buying decision process (Z) for products at Tokopedia.This is not in line with previous research conducted by Rakhma, et al. (2019) but this research is in line with Kurnyawati (2014) which states that brand awareness has no significant effect on the purchasing decision process through brand image.Maybe in this case consumers tend to buy products with familiar brands because they feel safe, avoiding various risks with the assumption that familiar product brands are more reliable.

CONCLUSIONS
This study aims to measure the influence of brand ambassador, brand awareness, and brand image variables on the product purchase decision process at Tokopedia in the 2018-2020 period.Based on the results of the analysis in the form of distributing questionnaires namely g-from to 131 people and discussions were carried out using path analysis.So the following conclusions can be drawn: (a) There is a significant influence between the brand ambassador variable (X1) on brand image (Y).(b) There is a significant influence between brand awareness variables (X2) on brand image (Y).(c) There is a significant influence between brand ambassador (X1) and brand awareness (X2) variables on brand image (Y).(d) There is a significant influence between the brand ambassador variable (X1) on the product purchasing decision process at Tokopedia (Z).(e) There is a significant influence between brand awareness variables (X2) on the product purchasing decision process at Tokopedia (Z).(f) There is a significant influence between the brand image variable (Y) on the product purchasing decision process at Tokopedia (Z).(g) There is a significant influence between the brand ambassador (X1), brand awareness (X2) and brand image (Y) variables on the product purchasing decision process at Tokopedia (Z).(h) There is an indirect effect between the brand ambassador variable (X1) on the purchasing decision process (Z) through brand image (Y).(i) There is no indirect effect between brand awareness variables (X2) on the purchasing decision process (Z) through brand image (Y).

Figure 1 .
Figure 1.The World's Fastest Growing E-Commerce Source: Databoks (2019) Based on Google research, Temasek and Bain & Company entitled e-Conomy SEA 2019.For the digital economy in the Southeast Asia (ASEAN) region, it is estimated that there are 360 million internet users who have enormous potential in 2019.As for (gross merchandise value) GMV e-Indonesian commerce in 2015 reached US$ 1.78 billion.Then in 2019 projected grew by (88% CAGR) to US$ 20.9 billion and will increase (48% CAGR) to US$ 82 billion or around IDR 1,148 trillion.This value is equivalent to half of the ecommerce market share in the ASEAN region.The development of the internet especially in Indonesia itself for many parties it is an opportunity for those who want to engage in online business, thus causing new habits in all aspects of life.The rise of online shopping in society in recent years has encouraged the growth of the logistics sector, especially in Indonesia.This can be seen in a survey conducted by Merchant Machine regarding the fastest growing e-commerce in the world in 2018.The rise of e-commerce, social media, as well as the presence of shopping services from internet- 25.1 trillion in 2014 and will increase to Rp. 69.8 trillion in 2016.Similarly, in 2018, the value of digital trade transactions in Indonesia continued to increase to Rp. 144.1 trillion.
To search for errors (є1) namely by looking at the R square in the model summary table.The formula to get the residual coefficient is 1.00 -R square.So the error (є1) is = 1.00 -0.412 = 0.588.