The Influence of Social Media Marketing and Brand Image on Brand Loyalty and its Impact on Consumer Purchase Decision with Store Atmosphere as A Moderation Variable (Case Study of XYZ Mall)

This research aims to obtain the results of the influence of Social Media Marketing, Brand Image, on Brand Loyalty and Consumer Purchase Decision with Store Atmosphere as a moderating variable. In this research, data was obtained by distributing questionnaires. The analytical method used in this research is SEM-PLS using SmartPLS 4 media. The research results show that Social Media Marketing has a significant effect on Brand Loyalty, Social Media Marketing has no significant effect on Consumer Purchase Decision, Brand Image has an effect significant effect on Brand Loyalty, Brand Image has no significant effect on Consumer Purchase Decision, Brand Image with Store Atmosphere as a moderating variable has no significant effect on Consumer Purchase Decision, Brand Loyalty has a significant effect on Consumer Purchase Decision.


INTRODUCTION
Consumer purchasing decisions are an important aspect to increase sales of a business of goods or services such as a mall.According to (Yusuf, 2021), purchasing decisions are thoughts in which individuals evaluate various options and make a choice on a product from many choices.(Gunawan, 2022) stated that a mall is a shopping center or place for exchange and distribution of goods or services that are commercial in nature, involve careful planning and design and also involve a manager, operational management and then customers.If a mall can determine consumer purchasing decisions then this will have an impact on sales value and also visitors' interest in coming to visit the mall.However, on the other hand, if a mall cannot influence consumer purchasing decisions, it will have an impact on sales from Tenants or Outlets which then also have an impact on other entrepreneurs' decisions to rent in the mall area.Malls in general have determined their market share since the establishment of the mall, by considering the location, decoration, filling or tenant area and services or what can be called standards or classes within the mall category.Once the market share has been determined, the thing that needs to be considered or done is how to ensure that visitors from the determined market share can be loyal or loyal to the mall.
There are many malls in Indonesia, one of them is mall XYZ which located in Jakarta, the capital city of Indonesia.Based on Customer Traffic data for the last 5 years which is presented in Figure 1, 2020 was the lowest year with a decrease of 49 percent from 2018 and a decrease of 54 percent from 2019, then from the graph above it can be seen that 2021 and 2022 experienced an increase from 2020 but the 2021 data still experienced a 42 percent

The Influence of Social Media Marketing and Brand Image on Brand Loyalty and its Impact on Consumer Purchase Decision
with Store Atmosphere as A Moderation Variable (Case Study of XYZ Mall) Shelvy Kurniawan 1* , Novian Tanujaya 2 1,2 Management Department, BINUS Business School Undergraduate Program, Bina Nusantara University, Jakarta 11480, Indonesia shelvy.kurniawan001@binus.ac.id; novian.tanujaya@binus.ac.id *Correspondence: shelvy.kurniawan001@binus.ac.id decrease compared to the 2018 Visitor data and a 48 percent decrease from 2019, then the 2022 data experienced a quite significant increase from 2020 and 2021 but still experienced a 15 percent decrease compared to 2018 and a decrease of 24 percent compared to 2019.The drastic decline in 2020 was caused by the world being faced with the Covid 19 problem which had an impact on all areas of business or businesses including shopping centers or malls.Collection of sales data obtained from 2018 to 2022, obtained if in 2018 Outlet or Tenant sales were 974,582,631.Then in 2019 there was a decrease of 25 percent from 2018, in 2020 there was a decrease of 58 percent from 2018, then in 2021, although there was an increase of 23 percent per year, it still experienced a decrease of 48 percent from 2018 and in 2022 there was an increase sales were significant, namely 92 percent but were still unable to return to the sales figures from 2018 with a decrease of 1 percent.This is in accordance with the results obtained from Customer Traffic, which from 2019 to 2020 experienced a drastic decline and began to increase in 2021 and 2022 but has still not returned to sales conditions in 2018.Then the Loyalty Member data for the last 5 years was obtained from the Marketing Communication department.The results showed that every year the number of Loyalty Members experienced growth, although slowing down, but from 2019 the growth that occurred experienced a slowdown in numbers, where in that year there was an increase of 16 percent from in 2018, then experienced a growth slowdown of only 8 percent in 2020, then experienced a growth slowdown of only 4 percent in 2021 and experienced a growth slowdown of only 2 percent in 2022.From the three results obtained, it can be seen that from 2020 to 2022, the increase in the number of Loyalty Members has decreased, which is not in line with the growth in sales and the number of visitors in 2020 to 2022, which has increased.
By considering problems that faced by XYZ mall that was presented in Figure 1 until Figure 3, the researcher wants to analyze the influence of Social Media Marketing, Brand Image, on Brand Loyalty and Consumer Purchase Decision with Store Atmosphere as a moderating variable.The limitations of previous research and recommendations from previous research are as follows.(Ansari, Ansari, Ghori, & Kazi, 2019) state that there are many other variables that can be used, such as brand health, brand loyalty, brand image, consumer engagement and others.Different from previous research, in this research, the author wants to add variables such as Brand Image, Brand Loyalty, and Store Atmosphere as moderating variables to answer limitations and further studies suggested by previous researchers.(Ardiansyah & Sarwoko, 2020) state that the research conducted only focuses on the swimwear category and further analysis is needed by comparing various other levels of involvement, therefore further studies need to be carried out by including Brand Image, Brand Love & Brand Loyalty as a model developed.Different from previous research, in this research the author will examine a different object from previous research, namely shopping centers or malls, which then the author also includes several variables such as Brand Image and Brand Loyalty by also adding Store Atmosphere as a moderating variable between Brand Image and Brand.Loyalty to answer the Further Research submitted.(Ibrahim, 2022) states that their study is the first of its kind, and future researchers are encouraged to replicate this research by adding more contextual factors (moderators) that have not been examined in our research (e.g.cultural differences, country development, and product type) and then future research can also address the antecedents and consequences of SMMA through systematic reviews or meta-analytic reviews.Different from previous research, this research with previous research is that the author will discuss in more depth the influence of Social Media Marketing and Brand Image on Brand Loyalty with its impact on Consumer Purchase Decisions with Store Atmosphere as a moderating variable to answer the limitations and Further Study presented in the research previous.
The research from (Sudaryanto, Nari, Awaliyah, Wulandari, & Hanim, 2019) has the following limitations and provides suggestions for future researchers, that purchasing decisions are influenced by many variables only in research This was researched using five variables so that for further research it is possible to re-evaluate with more variables.The difference that the author will examine to answer the limitations and suggestions of previous researchers is by examining the same dependent variable but with different independent variables such as Social Media Marketing and Brand Image on Consumer Purchase Decisions to answer the limitations and based on the suggestions of previous research.(Hasian & Pramuditha, 2022) suggests that future researchers can expand the research so that more complete information is obtained and can continue or develop this research by looking for other factors that can influence purchasing decisions such as price, promotion, personal, social and product completeness and to continue further research from previous research, the author will use more different variables such as Social Media Marketing and Store Atmosphere to test their impact on Customer Purchase Decision.(Marliawati & Cahyaningdyah, 2020) has several limitations, including the research only applying two independent variables, considering that brand trust has many factors that influence customers to become loyal to the brand.Previous researchers recommended to future researchers to add new variables and/or indicators to create models and answer the problems that occurred in this research.Different from research from (Marliawati & Cahyaningdyah, 2020), in this research the author will conduct research from a different research object, and the author will use several different variables from previous research to re-examine the impact of Brand Image on Brand Loyalty with additional independent variables such as Social Media Content Marketing.(Arman & Shabbir, 2020) said that future researchers can use large samples and can utilize other industries including luxury products.They can also introduce moderators in the model.The sample size in this study is small so that future researchers can use a larger sample.This study shows the results of one company while future researchers can utilize multiple companies.The difference between the research carried out by the author and previous researchers is that this research will be carried out in a different industry, namely shopping centers or malls, then here the author will also introduce the moderator variable Store Atmosphere into the influence of Brand Image on Brand Loyalty and the sample that will be used by the author this time This will be bigger than previous research to answer the limitations and Future Research of previous research.
So with the discovery of limitations, further studies, and recommendations or suggestions from previous research, the author formulates the problem as follows: Does the implementation of Social Media Marketing have a significant influence on Brand Loyalty?Does Brand image have a significant influence on Brand Loyalty?Does the implementation of Social Media Marketing have a significant influence on Consumer Purchase Decisions?Does Brand image have a significant influence on Consumer Purchase Decisions by being moderated by the Store Atmosphere?Does Brand Loyalty have a significant influence on Consumer Purchase Decisions?

Hypothesis 1 (H1): The Influence of Social Media Marketing on Brand Loyalty
Based on the results obtained by (Ibrahim, 2022), shows a strong positive correlation between Social Media Marketing and Brand Loyalty.This shows that the better marketing provided via social media can increase consumer loyalty.

Hypothesis 2 (H2): The Influence of Social Media Marketing on Consumer Purchase Decisions
Based on the research results of (Ansari, Ansari, Ghori, & Kazi, 2019), Social Media Marketing has a significant positive relationship with consumer purchasing decisions.So it can be concluded that Social Media Marketing has an influence on Consumer Purchase Decisions.These results are in line with (Manzoor, Baig, Hashim, & Sami, 2020) which shows that social media marketing has a significant effect of 69.6% on consumer purchasing decisions.

Hypothesis 3 (H3): The Influence of Brand Image on Brand Loyalty
Based on research conducted by (Çelik, 2022), results were obtained which showed a significant positive direct influence between brand image and brand loyalty.The results of this research support the results of several studies in the current literature such as research by (Arman & Shabbir, 2020) which shows that brand image has a positive effect on brand loyalty but is contrary to the results of (Marliawati & Cahyaningdyah, 2020), which states that Brand Image has a direct, but not significant, positive influence on customer brand loyalty.These results prove that improving brand image does not have a big impact in creating customer brand loyalty.

Hypothesis 4 (H4a): Influence of Brand Image on Consumer Purchase Decision
Based on research conducted by (Enrico, Purba, Budiono, & Rajagukguk, 2021) was founded that Brand Image has an influence on Consumer Purchase Decision.

Hypothesis 4 (H4b): Influence of Brand Image on Consumer Purchase Decision with Store Atmosphere as a moderating variable
From research conducted by (Sudaryanto, Nari, Awaliyah, Wulandari, & Hanim, 2019) shows that Store Atmosphere moderates brand image, price and promotion on purchasing decisions, as seen from the positive regression coefficient of.This proves that a better brand image and moderated by Store Atmosphere can improve purchasing decisions.From research by (Supiyandi, Hastjarjo, & Slamet, 2022) states that the variable that has the most influence on purchasing decisions is brand loyalty of all the variables mentioned used.In other words, increased brand loyalty will greatly influence purchasing decisions.

METHODS
The research method that will be used by the author this time is to use associative research with a quantitative approach.According to Sugiyono in (Rukminingsih & Latief, 2020) quantitative research method is defined as a research method based on valid or real data that is used to research a certain population or sample, data collection using research instruments, quantitative data analysis or statistics, with the aim of testing predetermined hypotheses.According to Sugiyono in (Rukminingsih & Latief, 2020) this type of associative research is related to the formulation of existing problems to determine the relationship between variables.The Time Horizon used is cross-sectional which according to Sekaran & Bougie (Nisrina, 2021) research was conducted to research a phenomenon over a certain period of time.
The population of this study are customers who use Instragram social media, where according to Hair formula (Fauzi, Lukitaningsih, & Hutami, 2023), if the research to be conducted has a large population and the number is unknown, then the sample size can be determined at a minimum of 5 to 10 times the number of indicator variables used.In this research, there are 27 indicators as (n) which will then be multiplied by 5 to obtain a sample size that can be used as a basis for research, namely 135 people.The primary data for this research was obtained through the results of respondents' answers in a questionnaire distributed online which was measured using a 5-point Likert scale.Meanwhile, secondary data was obtained from the research results previously as material for literature study.The data obtained was then processed using the Smart PLS 4 application using the Structure Equation Modeling method.

RESULT AND DISCUSSION
The results of the respondent data obtained in this research were 166 respondents.Valid data that meets the criteria as customers, namely those who are followers of Instagram, is 156 respondents.Then the number of samples suitable for use was 156 respondents with the results of the respondent data being dominated by 57.2% male and the remaining 42.8% female customers.From the respondent data, 94% are customers who have visited and 6% have never visited.Apart from that, from the respondent data, it was found that 71.1% had followed social media and 28.9% had not followed social media but social media content had appeared on their social media.Based on the results of calculations carried out with SmartPLS software, the results of the structural equation path diagram are as follows.The data results are then analyzed first to find out whether they meet the test criteria and then continue with the Structure Equation Modeling test.First, the Outer Model Testing is done through validity test, discriminant validity test and composite reliability test.

Validity Test
According to (Hair Jr, Hult, Ringle, Sarstedt, Danks, & Ray, 2021), At this stage the criterion value that can be used is through the Loading Factors value which is at least greater than 0.7 then if the Loading Factors factor value is above 0.7 then the indicator from the questionnaire can be declared Valid.Conversely, if the Loading Factors value is below 0.7 then the questionnaire indicator is declared invalid.Based on the test results in Table 1, the value for each Outer Model is more than 0.7, which is the criterion for the Loading Factor, which can be concluded that all the indicators used meet the criteria and are valid.Then the indicators that can represent all the indicators used as a basis for further discussion are Social Media Marketing (SMM4), Brand Image (BI1), Brand Loyalty (BL3), Store Atmosphere (SA1), and Consumer Purchase Decision (CPD4).

Discriminant Validity Test
The next stage can be tested by looking at the value of the Average Variance Extracted.Fornell & Larcker (Bastian & Saputra, 2022) state that the Average Variance Extracted must be greater than the latent construct highest squared correlation.So that Average Variance Extracted can be applied to all variables so that it can be used to measure the discriminant validity value and then, if the resulting value is greater than the value of each existing latent variable then the data is valid data, and vice versa.If the Average Variance Extracted (AVE) value is equal to 0.5 or greater then the validity value is acceptable, then if the AVE value is less than 0.5 then the validity cannot be accepted.In table 2 it can be seen that the values marked in bold are the square root of Average Variance Extracted or AVE and for elements outside the diagonal it is the correlation between variables.So it can be concluded that if the AVE value above is greater than the correlation between variables and has a value above 0.50, then the validity value of the data above can be accepted, which can then be continued to the Reliability Test stage.

Composite Reliability Test
According to Heale (Bastian & Saputra, 2022) the reliability value is used to measure the consistency value of research data, so that if the data is measured again, it will give close to the same results.Measuring reliability can be done in two ways, namely Cronbach's Alpha and Composite Reliability.Cronbach's Alpha is used to test reliability which produces lower values, so in the research the author used Composite Reliability.According to (Hair Jr, Hult, Ringle, Sarstedt, Danks, & Ray, 2021) they expect that the value that will be obtained from composite reliability will be more than 0.70 so that: If the value of Composite Reliability is > 0.70 then the resulting value is said to be Reliable.If the value of Composite Reliability is < 0.70 then the resulting value is said to be unreliable.In table 3, the Composite Reliability value for each indicator is greater than 0.70, so it can be said that all the data obtained meets the criteria of the Composite Reliability test.After the outer model testing is done, the next is the inner model testing which include R-Square testing, Goodness of Fit Index testing and Significant Test.

R-Square Test
Ghozali (Bastian & Saputra, 2022) states that the r-square function is to evaluate the inner model or measure the level of goodness of fit in a structural type model in the form of values.Apart from that, a low r-square value will indicate that the ability of the independent and dependent variables is limited and if the r-square value is close to 1, it indicates that the variable being tested provides clear and precise data for testing.According to Hair, et al (2021), if the R-Square value is more than 0.67 then the model can be said to be strong.If the R-Square value is between 0.33 -0.67 then the model can be said to be medium or moderate.The R-Square value is less than 0.33, so the model can be said to be weak.Using SmartPLS 4.0, the R-Square value calculation results are obtained as presented in Table 4.In table 4 it can be seen that the R-Square value for Brand Loyalty is 0.671 and Consumer Purchase Decision is 0.673 which can be concluded that the model for these two variables can be said to be strong, because it is 33.9% for the Brand Loyalty variable and 33.7% for Consumer Purchase.Decisions are influenced by variables outside the research.Testing the Structural Model using the Goodness of Fit Index Tenenhaus (Bastian & Saputra, 2022) states that to be able to ascertain the performance value of the Outer Model and Inner Model, researchers can use the Goodness of Fit Index (GoFI).There are several criteria for GoFI itself, namely: 0.10 is the same as small or weak GoFT, 0.25 is the same as medium or moderate GoFT, 0.36 is the same as large or strong GoFT.To find the Goodness of Fit Index value, there is a formula that must be used, the formula used is as follows: √(AVE×R^2), so the AVE from this research itself is (0.755+0.756+0.793+0.738+0.7 66)/4 = 0.952 multiplied by the R-Square value (0.671+ 0.673)/2 = 0.672, then 0.952 multiplied by 0.672 results in a Goodness of Fit Index of 0.639 and is declared large or strong.

Significant Test
Significance testing is carried out to obtain the influence value of each independent to dependent variable used as well as the moderating variable.The method that will be used in this test is Bootstrapping to find out the effect of the Path Coefficient.According to Hair, et al (2021), the significance test is carried out by looking at the T statistics value as follows: If the T statistics value is greater than 1.96 then the relationship between variables can be said to have a significant effect with a P-Value below 0.05, and vice versa if the T statistics value is less than 1.96 with P-Values above 0.05, then the relationship between variables can be said to have no significant effect.

Discussion
Based on statistical calculations carried out by the author, it can be concluded that Social Media Marketing has a significant effect on Brand Loyalty.This is supported by the T-Statistic value showing 3.23 which is greater than 1.96 and the P-Values value showing 0.001 which is smaller than 0.05.This is in line with previous research conducted by (Ibrahim, 2022) whose research shows a strong positive correlation between Social Media Marketing and Brand Loyalty and is also in line with what was conveyed by (Ismail, 2017) that Social Media Marketing has been considered as a tool which is effective in developing relationships between companies and consumers.
Based on statistical calculations carried out by the author, it can be concluded that Social Media Marketing does not have a significant effect on Consumer Purchase Decisions.This is supported by the T-Statistic value showing 1.522 which is smaller than 1.96 and the P-Values value showing 0.128 which is greater than 0.05 and is not in line with the research of (Ansari, Ansari, Ghori, & Kazi, 2019) which states that Social Media Marketing has a significant positive relationship with consumer purchasing decisions, but is in line with the results obtained by (Yaputra, 2020) who states that Social Media Marketing has an insignificant influence on Consumer Purchase Decisions in Department Store.
Based on statistical calculations carried out by the author, it can be concluded that Brand Image has a significant effect on Brand Loyalty.This is supported by the T-Statistic value showing 5.109 which is greater than 1.96 and the P-Values value showing 0.000 which is smaller than 0.05.This is in line with research conducted by (Çelik, 2022), which showed results showing a significant positive direct influence between brand image and brand loyalty.
Based on statistical calculations carried out by the author, it can be concluded that Brand Image does not have a significant effect on Consumer Purchase Decisions.This is supported by the T-Statistic value showing 1.769 which is smaller than 1.96 and the P-Values value showing 0.077 which is greater than 0.05.This is in line with research conducted by (Iskuntianti, Faisal, Naimah, & Sanjaya, 2020) which states that the Brand Image variable has a positive but not significant effect on Consumer Purchase Decisions, and is in line with research conducted by (Pusparini & Khalid, 2023) which shows that brand image has a positive and significant influence on Consumer Purchase Decisions.
Based on the statistical results that have been carried out, it can be concluded that Brand Image does not have a significant effect on Consumer Purchase Decisions with Store Atmosphere as a moderating variable.This is supported by the T-Statistic value showing 1.522 which is smaller than 1.96 and the P-Values value showing 0.125 which is greater than 0.05.then these results are not in line with research conducted by (Sudaryanto, Nari, Awaliyah, Wulandari, & Hanim, 2019) which shows that Store Atmosphere moderating Brand Image on purchasing decisions has a significant influence positive, but in line with research conducted by (Santosa, 2019) who said that Brand Image which is moderated by Store Atmosphere has no effect on Consumer Purchase Decisions.
Furthermore, it can be concluded that Brand Loyalty has a significant influence on Consumer Purchase Decisions.This is supported by the T-Statistic value showing 6.361 which is greater than 1.96 and the P-Values value showing 0.000 which is smaller than 0.05.This is in line with research conducted by (Supiyandi, Hastjarjo, & Slamet, 2022) which states that the variable that has the most influence on purchasing decisions is brand loyalty of all the variables used.In other words, increased brand loyalty will greatly influence purchasing decisions.
Besides the statistical findings, this research also has limitation that is regarding the research place which was only hold in one mall.Thus, future research can do comparative study by holding the research in several malls to compare the result.Then, in a mall of course consist of many stores, this study was studying store atmosphere of the whole mall and not specific to certain store, so future research need to be more specific in measuring the store atmosphere.

CONCLUSION
From the results of data processing that the author has carried out in this research using SmartPLS 4, in accordance with the objectives of this research, it can be concluded that: Social Media Marketing has a significant effect on Brand Loyalty, Social Media Marketing has no significant effect on Consumer Purchase Decision, Brand Image has a significant influence on Brand Loyalty, Brand Image does not have a significant effect on Consumer Purchase Decision, Brand Image with Store Atmosphere as a moderating variable does not have a significant effect on Consumer Purchase Decision, and Brand Loyalty has a significant influence on Consumer Purchase Decisions.Furthermore, several insignificant results were obtained between the influence of Social Media Marketing on Consumer Purchase Decision, the influence of Brand Image on Consumer Purchase Decision and the influence of Brand Image on Consumer Purchase Decision with Store Atmosphere as a Moderation variable.So the author's suggestion is to create content that is more informative but packaged in an interesting and creative way so that it can attract the interest of social media followers to make purchases.By creating content that is more informative and creative it will also have an impact on Brand Image so that this has a relationship to increase the influence of the dependent variables on Consumer Purchase Decisions.

Figure 1 .
Figure 1.Number of Visitors or Customer Traffic of XYZ Mall Source: XYZ Mall (2023)

Figure 3 .
Figure 3. Number of Loyalty Members of XYZ Mall Source: XYZ Mall (2023) The Influence of Social Media Marketing and Brand.... (Shelvy Kurniawan & Novian Tanujaya) Hypothesis 5 (H5): The Influence of Brand Loyalty on Consumer Purchase Decisions

Table 1
Convergent Validity Test

Table 3 .
Composite Reliability Test Results

Table 4 .
R-Square Test Results

Table 5 .
Path Coefficient Test Results, T-Statistics and P-Values H1.The Social Media Marketing variable has a significant effect on Brand Loyalty because the T-Statistic value shows 3.23 which is greater than 1.96 and the P-Values value shows 0.001 which is smaller than 0.05.H2.The Social Media Marketing variable does not have a significant effect on Consumer Purchase Decision because the T-Statistic value shows 1.522 which is smaller than 1.96 and the P-Values value shows 0.128 which is bigger than 0.05.H3.The Brand Image variable has a significant effect on Brand Loyalty because the T-Statistic value shows 5.109 which is greater than 1.96 and the P-Values value shows 0.000 which is smaller than 0.05.H4a.The Brand Image variable does not have a significant effect on Consumer Purchase Decision because the T-Statistic value shows 1.769 which is smaller than 1.96 and the P-Values value shows 0.077 which is bigger than 0.05.H4b.The Brand Image variable with Store Atmosphere as a moderating variable does not have a significant effect on Consumer Purchase Decision because the T-Statistic value shows 1.522 which is smaller than 1.96 and the P-Values value shows 0.125 which is bigger than 0.05.H5.The Brand Loyalty variable has a significant effect on Consumer Purchase Decisions because the T-Statistic value shows 6.361 which is greater than 1.96 and the P-Values value shows 0.000 which is smaller than 0.05.