Advertising and Wom Effect Towards Consumer Buying Decision Process on Janji Jiwa Coffee in JAKARTA

This Research aimed to analyze Consumer behavior of Janji Jiwa coffee which focuses on the buying decision process from consumers, which includes the influence of Instagram promotion and word of mouth on buying decisions, which mediated by purchase intention and brand awareness from Kopi Janji Jiwa in East Jakarta. This Research survey was conducted through Questionnaire on 100 Janji Jiwa customers who have bought the Janji Jiwa coffee at least 4 times in Janji Jiwa outlets in AEON mall Jakarta Garden City and Transera Waterpark at East Jakarta. All the responses are processed with path analysis using Smart - PLS application. The results finding reveal that between Instagram promotion (Advertising) and word of mouth, Purchase Intention and Brand Awareness, and Buying decision, each of these variables has an effect on each other. But there are no influence between word of mouth on purchase intention. In Addition, Purchase intention and brand awareness are proven to mediate and strengthen the relationship between Instagram promotion and word of mouth towards buying decision.


INTRODUCTION
The COVID-19 pandemic situation, which began in 2020, had an impact on most of the business sectors, including businesses in the food and beverage sector, especially coffee shops. The Pandemic situation also had an impact on changes in consumer behavior, which had to switch from the Dine-in system to the Take Away and Online Delivery systems. So that advertising acts as a communication procedure that influences the audience (viewers, readers or listeners) to buy or take action on products, or services etc. (Abden, 2011). In this case Instagram is one of the social media that is currently widely used by people as a media for online promotion. Janji Jiwa Coffee in advertising or marketing products also uses Instagram social media through Instagram promotion features and swipe up links and Instagram Ads which can make it easier for consumers to make purchases. Social media sites in this case can be said as a good platform for retailers to create experiences and retailers can use the information stored on social media sites to improve user experience with their brands (Chaubey et al., 2016). According to Alhabash et al. (2017) Social media advertising can be defined as any online content designed with the intent to persuade and/or distributed via social media platforms that enable Internet users to be able to access, share, engage with, add to, and co-create. When compared to traditional media, Scott (2015) argues that, social media gives users the ability to not only view and access information, but also share, engage, and create personalized, semi-private, and public multimodal content through a network of friends, followers, and users.
In addition to carrying out advertising promotions on Instagram social media, Word of Mouth from customers also plays an important role in the purchase decision-making process. In this case, consumers often use word of mouth referrals as a tool to reduce the amount of information to be processed and to reduce consumer anxiety (Hung & Li, 2007). Then, Herr et al. (1991) argue that advocacy in the form of word of mouth communication can be the most influential source of information for the purchase of several products / services because it is considered to come from a more reliable source, and can help reduce customer anxiety. Word of Mouth can be a powerful source of information, which can influence and increase Purchase Intention, Brand Awareness, which in turn can influence consumer behavior and consumer perceptions of an item or service (Steffes & Burgee, 2009;Ha, 2004;Page & Lepkowska, 2002).
Then, Lin et al. (2011) consider purchase intention as a psychological state of a consumer who plans to consume a certain brand within a certain period of time and possibilities. Then, for the measurement of Purchase Intention from consumers depending on the consumer's intention to buy certain products that the higher purchase intention (purchase intention) indicates a higher probability of purchase (Stever, 2011). Thus, Yim et al. (2012) stated that purchase intention is the possibility of consumers buying certain products, where the higher the purchase intention, the more likely consumers will buy the product. Then, Belch and Belch (2012) define Purchase Intention as some point in the buying process, that are when consumers have to stop looking for and evaluate information about alternative brands in generating sets and making purchases, generally based on matching the purchase motive with the attributes or characteristics of the brand under consideration.
Brand Awareness is very important in the purchasing decision-making process, because if there is no Brand Awareness there will be no communication and no transactions (Percy, 1987). Then, Gustafson and Chabot (2007) argue that, if an organization has a successful Brand Awareness it means that the organization's products and services have a good reputation in the market and are acceptable. and Brand Awareness also continues how to be aware of current and potential customers of your products and services.
Before making a purchase decision by consumers there are five stages in the decision-making process. That are, problem recognition, information seeking, evaluation of alternatives, purchase decisions and postpurchase behavior (Kotler and Armstrong, 2001). In understanding customer purchasing decisions, marketers must understand the consumption process and product usability in customer perceptions (Pride and Ferrell, 2012).
Previous research conducted by Maria et al. (2019) shows that based on the results of the path analysis conducted, the effectiveness of advertising directly has a positive and significant effect on. purchase intentions. Maria et al. (2019) also proves that the effectiveness of advertising directly has a positive and significant effect on brand awareness. This finding is relevant to the research of Gunawan & Dharmayanti (2014), the result of this research is that Advertising has a positive and significant effect on Brand Awareness. Maria et al. (2019) based on the results of the path analysis conducted, word of mouth directly has a positive and insignificant effect on Brand Awareness. In addition, previous research conducted by Maria et al. (2019) also shows that Brand Awareness directly has a positive and significant effect on purchase intention. Consumers buy products with certain brands that they know because they hope the product they buy is really a product that has been tested for quality and can be used to meet future expectations.

Research conducted by
Then previous research conducted by Wijaya and Yulianti (2020) showed that Purchase Intention had a positive and significant effect on Buying Decision. Ajzen (2006) defines Purchase Intention as a hypothesis to take aspects of motivation that affect online buying behavior.
Previous research conducted by Novansa and Ali (2017) found that Brand Awareness has a positive and significant influence on Consumer Buying Decisions. Shimp (2014) argues that Brand awareness has four dimensions, that are : Top of Mind, Brand Recall, Brand Recognition, and Unaware of Brand. This is in accordance with the results of research conducted by Hasibuan (2012) which states that Brand Awareness, Perception of Quality, and price have a positive and significant relationship to Buying Decision. The price variable is the independent variable that has the greatest influence on the Buying Decision variable. The results of previous research conducted by Novansa and Ali (2017) are in line with the results of research conducted by Kardoyo (2007) which proves that Brand Awareness has a positive and significant effect on Buying Decision. Thus, the higher consumer awareness of a brand, it will affect the buying decision of consumers.
The main problem in this study is the decline in sales in 2020 caused by the PSBB policy which has an impact on changing consumer behavior from dine-in to online delivery. This study wants to discuss the consumer behavior of Kopi Janji Jiwa which focuses on the buying decision process from consumers which includes the influence of Instagram promotion and word of mouth on buying decisions mediated by purchase intention and brand awareness of Janji Jiwa Coffee in East Jakarta as the object of research and consumer behavior as the main theory in this research. So based on the research phenomenon, the formulation of the problem in this study can be translated into several points of problem formulation as follows: (1)

METHODS
This study uses qualitative and quantitative analysis. In this study, the data was processed using the SPSS program by conducting a Pre-Test (Test Validity and Reliability) first. Then, this study uses path analysis to test mediation using Smart PLS 3. The reason for using Path analysis in this study is because according to Ferdinand (2011) The use of the path analysis method allows researchers to test several endogenous variables at once with exogenous variables. In addition, another reason for using path analysis in this study is that there is a mediating effect between exogenous and endogenous variables. The sampling method in this study used non-probability sampling with purposive and snowball sampling techniques. The criteria used in selecting the sample are consumers who have purchased Janji Jiwa coffee in the last 6 months, consumers who have seen advertisements/promotions of Janji Jiwa coffee on Instagram social media, and Consumers whose purchasing decisions are influenced by WOM (friends, family, etc.). The samples of this study were taken from 2 Janji Jiwa Coffee shops located in East Jakarta, that are, the Janji Jiwa coffee shop located at AEON Mall Jakarta Garden City, East Jakarta and the Janji Jiwa Coffee shop located in Transera Waterpark, Harapan Indah, which is on the border East Jakarta area with Bekasi area. The number of samples from this research is 100 samples. In order to obtain interval data, this study uses a 5-point Likert scale, as explained by Sugiyono (2016) that the variables measured by the Likert scale can be translated into variable indicators and used to arrange items in the form of questions or statements. Respondents will be given five answer choices according to the Likert scale, that are: 1 = Strongly Disagree, 2 = Disagree, 3 = Neutral, 4 = Agree, 5 = Strongly Agree.
Before the research data was further processed, a pre-test analysis was conducted which aims to test whether the variables in the research model are valid and reliable, taken from 30 respondents at the beginning of the study. Because Ghozali (2017) argues that it is necessary to take answers from 30 respondents in the first stage of research as a condition to carry out testing of research instruments. In the pre-test there are validity and reliability tests, which according to Sugiyono (2016) there are several criteria to state that a variable can be said to be valid. The first is the value of the validity index 0.3. which means that, the correlation coefficient must be greater than 0.3. In SPSS software the correlation is explained based on the Pearson Correlation value or commonly called r-count. Second, the question item (indicator) of the variable can be said to be valid if the -value , with as the significance level = 0.05 (5%). In SPSS software, -value is explained based on the significant value (Sig. 2 tailed). In the reliability test, there is a provision in measuring the reliability of the questionnaire instrument, namely by calculating the reliability coefficient with a measuring instrument through Cronbach Alpha (α) contained in the SPSS application. Hair et al., (2010) stated that a measurement can be said to be reliable if the Cronbach Alpha is close to 1 or > 0.6. A measuring instrument is said to be reliable and can be processed at a later stage if the Cronbach Alpha value is > 0.6. If the measuring instrument has a Cronbach Alpha value < 0.6 then the measuring instrument can be said to be unreliable.
In this study, the Structural Equation Model (SEM) data analysis was chosen because it was based on research conducted by Joo et al. (2018) stated that the Structural Equation Model (SEM) analysis was also able to analyze the relationship between latent variables and their indicators, as well as latent variables with each other. related to the data analysis tool in this study, this study used Partial Least Square (PLS), using the statistical software SmartPLS 3. Hair et al. (2012) defines Partial Least Square (PLS) as a strong multivariate technique, and uses a component-based approach that can generate model estimates.
Then, Cao et al. (2018) mentions that the analysis using Partial Least Square (PLS) in the analysis of research data makes it possible to assess the measurement model and the structural model simultaneously. In data analysis using Partial Least Square (PLS), there are two stages of analytical approach, that are, the outer model in which convergent and discriminant validity tests are carried out, as well as reliability tests. Also, the inner model which includes the R2 value test (coefficient of determination), as well as the t value for hypothesis testing (Anderson & Gerbing, 1988;Koloseni & Mandari, 2017).

RESULTS AND DISCUSSIONS
In this study obtained 103 respondents who fit the criteria (Purposive Sampling). That are, consumers who have purchased Janji Jiwa coffee in the last 6 months, consumers who have seen advertisements/promotions of Janji Jiwa coffee on Instagram social media, and Consumers whose purchasing decisions are influenced by WOM (friends, family, etc.). In the validity and reliability test conducted on 32 respondents at the beginning of the study using IBM SPSS Statistics 23 in Table I, in the validity test, the value of the Pearson Correlation on each indicator has a value above 0.3 (> 0.3). In addition, the p-value or Sig. (2-tailed) as in the table above when compared with a significance level (α) of 5%, proved to have a value less than 0.05 (< 0.05). So, with the fulfillment of these two criteria, it can be said that all indicators of research variables are declared valid, and able to represent or represent research variables well. Furthermore, in the reliability test using IBM SPSS Statistics 23 in Table II it can be seen that the Cronbach's Alpha values of the 5 research variables marked in red boxes are all above 0.7 (> 0.7). That means, this research variable has passed the reliability test, and each indicator can be relied on as a research measuring tool. In data analysis using Partial Least Square (PLS), in the outer model, Joreskog & Sorbon (1993) states that, in testing the outer model, Confirmatory Factor Analysis (CFA) will be carried out which is used to test the validity and reliability of the measurement model of the variable or construct. which cannot be measured directly. The results of this CFA also show the relationship between indicators and variables or constructs. The convergent validity test is carried out to determine the extent to which the measurement indicators correlate with other measurement indicators in the same latent variable. A variable indicator is declared to meet the convergent validity requirements, if the outer loading value is above 0.7 and has an Average Variance Extracted (AVE) value greater than 0.5 (Hatcher, 1994;Santosa, 2018). According to Santosa (2018), the variable indicator is declared to meet the requirements of convergent validity if the resulting outer loading value is above 0.7. Based on the results of the convergent validity test in Table III, all indicators of latent variables have outer loading values above 0.7, except for indicators BA 2, BD6, and BD 9 whose outer loading values are below 0.7. So it can be said that all indicators of variable measurement have met the criteria of convergent validity. In addition to the outer loading value, the convergent validity test also pays attention to the Average Variance Extracted (AVE) value. Hatcher (1994) also explains that convergent validity determines whether the items on the questionnaire can effectively reflect the appropriate factors, taking into account the Average Variance Extracted (AVE) value which must be greater than 0.5. The value of Average Variance Extracted (AVE) in Table  IV shows that all latent variables in the study have an Average Variance Extracted (AVE) value greater than 0.5, so it can be concluded that all of these variables are considered to have met the criteria of convergent validity). Then, in testing discriminant validity, a comparison of the correlations between variables will be carried out based on the square root value of the Average Variance Extracted (AVE) using the Fornell-Larcker Criterion method. Chin (1998) explains that discriminant validity can be achieved when the square root of the Average Variance Extracted (AVE) latent variable is greater than the correlation between other latent variables in the model. The results of the discriminant validity test in Table V, show that the results of the square root value of the Average Variance Extracted (AVE) of each latent variable or construct, have a greater value than the correlation value with other latent variables or constructs. Thus, from the value of the square root of the Average Variance Extracted (AVE), it shows that the measurement scale for the variables or constructs of this study has good discriminant validity. Furthermore, testing on the outer model using Partial Least Square (PLS) the next stage is the reliability test. The purpose of reliability testing is to determine the level of consistency of a measurement. If the measurement tool is tested several times, and gives relatively the same results or values, it can be said that the measuring instrument is reliable. Reuterberg & Gustafsson (1992) suggested that reliability will be achieved when both the Composite Reliability (CR) and Cronbach's Alpha values are greater than 0.7. The results of the reliability test in Table V show that both values have met the requirements for reliability testing, which has a value greater than 0.7. Thus, it can be concluded that all latent variables or constructs in this study have internal consistency reliability and are considered reliable, so they can be used for hypothesis testing.
In addition to testing the outer model, testing the inner model needs to be done to test the relationship between latent variables or constructs. In testing the inner model, there is an analysis of the value of R-Square (R2) which is used to analyze the coefficient of determination which measures how much influence exogenous variables have on endogenous variables. The result of R2 is 0.67, the research model is categorized as good or strong. Chin (1998) if the R2 result is 0.33, then the model is categorized as moderate, and if the R2 result is 0.19, it indicates that the model is categorized as weak. From the value of R-Square (R2) for Purchase Decisions in Figure  2, it can be explained that Instagram Promotion (Advertising), Word of Mouth, Brand Awareness, and Purchase Intentions are able to explain Purchase Decisions by 68.2%, while the rest is 31.8%, explained by variables or other factors outside this study. So it can be categorized as a good model because the R-Square value is greater than 0.67. Then, in Table 6 (See Appendices) the R-Square (R2) value of Brand Awareness can be interpreted that Instagram Promotion (Advertising) and Word of Mouth can explain Brand Awareness by 60.9%, while the remaining 39.1%, explained by other variables in outside of research. So that it can be categorized as a moderate model because the R-Square value is smaller than 0.67, but this model is still worthy of research because this research model is quite good. Then the R-Square (R2) value of Purchase Intention in Table 6 (See Appendices), it can be interpreted that Instagram Promotion (Advertising) and Word of Mouth can explain Brand Awareness of 62.7%, while the remaining 37.3%. So that it can be categorized as a moderate model because the R-Square value is smaller than 0.67, but this model is still worthy of research because this research model is quite good. Next, analysis of the value of F-Squared (F2) is used to see the effect size. Based on the results of the analysis of the F-square (F2) value in Figure 3, the F-square (F2) value which is colored red, has a value below 0.02, which means that the influence of exogenous variables on the structural model can be ignored because there is no effect. While the value of F-square (F2) which is black, has a value between 0.02 and 0.15, which means that the influence of this variable is relatively weak. For the green F-square (F2) value, it can be explained as follows: first, the influence of the advertising variable on brand awareness is considered sufficient because the F-square (F2) value is above 0.15 (0.450). second, the influence of the Brand Awareness variable on the buying decision is considered sufficient because the F-square (F2) value is above 0.15 (0.217). Third, the influence of the Purchase Intention variable on the buying decision is considered sufficient because the F-square (F2) value is above 0.15 (0.306). Fourth, the influence of the Brand Awareness variable on Purchase Intention is considered sufficient because the F-square (F2) value is above 0.15 (0.237). Then, in the inner model test, there is also a Q-Square (Q2) value which is used to measure prediction relevance, namely to validate the predictive ability of the structural model. This Q-Square (Q2) value also measures how well the observed values generated by the model and the parameter estimates are. The value of Q-Square (Q2) > 0 indicates that the structural model has predictive relevance, while the value of Q-Square (Q2) 0 indicates that the structural model has no predictive relevance. The results of the Stone Geisser test in Figure  4, obtained the value of Q-Square (Q2), and the three values in the table were found to be greater than 0, namely 0.362, 0.361, and 0.436. This means that the observed values have been reconstructed, so that the structural model has predictive relevance, and shows that exogenous variables or constructs, as explanatory variables, are able to predict their endogenous variables or constructs. After testing the inner model, the next step of this research is to analyze path coefficients and test research hypotheses to determine the magnitude of the relationship or influence of latent variables or constructs. Based on the results of the direct effect analysis in Figure 5, the advertising variable has an effect on purchase intention because H0 < t-table (2.195 < 1.98) and p-value > (0.000 > 0.05) so that H1 is accepted. The word of mouth variable has no effect on purchase intention because H0 > t-  In addition to testing direct effects, path analysis also examines indirect effects, namely testing through the role of mediating or intervening variables. Based on the results of the Indirect Effect test in Figure 6 it can be seen that Purchase Intention is proven to mediate the influence of Instagram Promotion (Advertising) on Buying Decision. because the value of H0 < t-table (1.999 < 1.98) and p-value > (0.046 > 0.05) so H7 is accepted. Then, the Brand Awareness variable was proven to mediate the effect of Word of Mouth on buying decision, because the value of H0 < t-table (1.983 < 1.98) and p-value > (0.046 > 0.05) so that H8 was accepted.

CONCLUSION
This study aims to determine the effect of the relationship between Instagram Promotion (advertising) and word of mouth on buying decisions mediated by purchase intention and brand awareness. Based on the research results obtained from 103 respondents, it can be stated that, the relationship of Instagram Promotion (advertising) to Purchase Intention, the effect of the relationship of Purchase Intention to Buying Decision, the influence of the relationship of Instagram Promotion (Advertising) to Brand Awareness, The influence of the relationship of Word of Mouth to Brand Awareness , The influence of the relationship of Brand Awareness has an effect on Buying Decision, each of these relationships influence each other. However, based on the results of the tests carried out, the relationship between the Word of Mouth variable and the Purchase Intention variable proved to have no effect. So, in this case, Word of Mouth is not effective in increasing the Purchase Intention of consumers.
To test the mediating variable in this research, through the role of the Purchase Intention variable as a mediating variable (Intervening). It can be concluded that the purchase intention variable is proven to be able to mediate and strengthen the influence of Instagram Promotion (Advertising) on Buying Decisions. Then, the test results in this study also show that Brand Awareness is proven to be able to mediate and strengthen the influence of Word of Mouth on Buying Decision.
From the results of this study, suggestions and input that the author wants to give and are expected to be considered for future researchers, namely so that future researchers can add other exogenous variables, such as product reviews, consumer reviews, and product quality, so that they are able to find out more and be more specific about how it affects the Buying Decision. Then, further researchers are expected to be able to expand the reach or scope of the research area, not only in the East Jakarta area, but can also examine respondents in other DKI Jakarta areas such as Central Jakarta, South Jakarta, North Jakarta, and West Jakarta, or even can do research in the Greater Jakarta area and outside the Greater Jakarta area, so that it can provide research results that are different from previous research, and the differences in the results of this study are expected to further enrich similar research topics.