THE IMPACT OF PAKSERV ON CUSTOMER SATISFACTION AND LOYALTY: STUDY CASE GRABFOOD INDONESIA

Every business needs an excellent service quality to make customer satisfied and loyal. However, the terms of satisfaction and loyal still become the problems that should be tackled like GrabFood in Banda Aceh which should cope the food service delivery delay to the customer. This research uses PAKSERV model to give the solution for improving the service quality, the population in this research is the users of grabfood Banda Aceh and the sample 204 users with purposive sampling. This research uses SEM-PLS with the help of AMOS and SPSS software which found that Tangibility has a significant and positive impact on customer satisfaction along with other variables such as Assurance, Personalization, Sincerity, and Formality. On the other hand, Reliability in this model has not a significant impact to the customer satisfaction even-though it has a positive impact. Thus, it could be interpreted that, GrabFood Banda Aceh should pay more attention to the Tangibility, Assurance, Personalization, Sincerity and Satisfaction.


INTRODUCTION
result showed that only Tangibility has a mean result above 3 (3.04), which means that the respondents choose neutral to the Tangibility, while the other indicators' mean scores are average around two, which means not satisfied.
The quality of service can be improved if using a suitable model. The model for service quality used in this research is PAKSERV, a model developed by Raajpoot to measure the service quality in Pakistan (Raajpoot, 2004). Raajpoot confirmed that Tangibility, Reliability, and Assurance from SERVQUAL related to the service quality standard in Pakistan but added Sincerity, Formality, and Personalization. Thus, this research will look at how to improve Grab Food service quality, which will satisfy and make the customer loyal to Grab Food by using the Pakserv model. (Kotler & Keller, 2016) stated that satisfaction shows the value of product or service performance in order to fulfil the customers' expectation; if the service or product performance fall below customers' expectation, then the customer will not be happy and might not be loyal to the products or service (Kotler & Keller, 2012). Besides, customer satisfaction might be affected by the customer's emotions after using a product or service (Tjiptono, 2019).

METHOD
This research uses a survey approach with a quantitative research type. The survey approach allows the researchers to gather the data from respondents to understand or predict the behaviour of a particular population (Nugroho & Irena, 2017), and the survey result will be combined with the quantitative method. The quantitative method uses the numerical method to analyze numerical data collected from the survey to explain a phenomenon (Nugroho & Irena, 2017).
The data collected in this research will use the questionnaire technique, a set of questions prepared by the researchers that must be answered by the respondents (Sekaran & Bougie, 2016). The questionnaire will be distributed to the users of Grab Food in Banda Aceh with a sample of 204 people and use purposive sampling, a technique that defines the sample with the characters from the researcher (Malhotra, 2009). In addition, the questionnaire will use a Likert scale of 1-5, with the condition value one is not very satisfied, and the value of 5 is delighted.

RESULTS AND DISCUSSION Respondents' Profile
The respondents' demographics are divided into gender, age group, marital status, education, and employment. Of 204 respondents, 60% of them are male while 40% are female, as can be seen from the chart below.

Respondents Gender
Laki-laki Perempuan Moreover, the majority of the age group from 204 respondents is 18-23 years old at 65%, followed by 24-29 years old at 24%. As it can be seen from the pie chart below;

Figure 2: Respondents' Age Group
In addition, marital status in this research is divided into three types: married, unmarried, widow, or widower. Most of the respondents' answered not married yet or unmarried for 84%, followed by married at 14%. As can be seen in the figure below:

Figure 3 Marital Status
Furthermore, the respondents are also divided into education levels (from elementary school to Postgraduate), where the most dominant education level is a graduate student at 77%, while the other education level is below 10% (figure 5), 65% 24% 7% 2% 1% 1%

Figure 4: Respondents' Education
Along with respondents' education level, the respondents' job was dominated by student college with 63%, followed by self-employed at 27%, students at 5%, professionals at 4%, and civil servants at 1% (figure 6).

Figure 5: Respondents' Job
Besides that, since most of the respondents are college students and graduates with unmarried status; thus, this affects the income data. Income in this research is divided into five types, starting from 2 million rupiahs to 3.9 million rupiahs, 4 million rupiahs to 7,9 million rupiahs, 8 million rupiahs to 10 million rupiahs, and above 10 million rupiahs. According to the data, most respondents' income is 2 million to 3.9 million rupiahs with 181 people, followed by 4 million rupiahs to 7.9 million rupiahs with 20 people (figure 7). 2% 2% 77% 1% 5% 11% 2%

Reliability Test
This research provides a reliability test to determine how reliable all of the variables in this test where the variables are tested repeatedly (Idzni et al., 2021). The reliability test uses a Cronbach alpha value of more than 0.7; if the variables in this research have a value of more than 0.7, then the variables will be stated as reliable. Table number 2 showed that the Cronbach alpha value for six variables in this research with 32 variables item is above 0.7, which means all of the variables are reliable.

Confirmatory Factor Analysis
Confirmatory factor analysis is one of the measurement steps to measure the indicators formed as latent variables or constructs (Hair Jr et al., 2014). This research has six exogen and two endogen variables, with 34 indicators total. All of the 34 indicators are tested by using AMOS, and the result is shown in the figure and  The result found that indicators values in Y.1 as customer satisfaction and Z.6 as customer loyalty are below 0.5, which will be eliminated as (Hair, 2010) stated that outer loading values between 0.5 and 0.7 are categorized as weak value and could be considered to be deleted if it can increase the result. However, if the outer loading value is below 0.5, the indicators should be eliminated (Hair, 2010).
After that, this research does secondary confirmatory factor analysis without including the indicators Y.1 and Z.6; the result is shown in the figure below

The goodness of Fit Evaluation
There is no benchmark tool for testing the model's hypotheses in SEM analysis. However, most researchers are using the cut-off value to test whether the models are acceptable or not (Ferdinand, 2014). The result is shown in table 6 below: The feasibility result showed that the marginal fit value is poor. So, this research needs to recalculate by using re-specification analysis with Modification Indices (MI); the method is to unify the most considerable value of MI in each indicator; the re-specification result can be seen in the table and figure below:  The structural equation model (SEM) was employed by using AMOS software. The structural equation model analysis is the step after confirmatory factor analysis; the SEM analysis was analyzed considering Critical Ratio (CR) or T-statistics and P-value. The result can be seen in the figure below;

Hypotheses Test
There are 13 hypotheses in this research that needs to be tested; the test has been done in the SEM analysis, and this part will explain further the hypotheses. The 13 hypotheses will use the critical ratio (CR) or T-statistics and P values to determine whether the hypotheses are accepted. The condition where a hypothesis is accepted if the P-value <0.05. The result will be shown in the table below; From the table above, it can be interpreted that Tangibility has a significant and positive impact on customer satisfaction since the p-value is 0.005 and less than 0.05. the percentage decree of Tangibility towards customer satisfaction is 0.237 or 23.7%. Tangibility is the form of physical service that affects the customer perception; if the customer perception is unsatisfactory, the service or product's value will be inferior. This result has the same result as the research of (Alnaser et al., 2017) with the title "Service Quality in Islamic Banks: The role of PAKSERV model, Customer Satisfaction and Customer Loyalty" showed that Tangibility, Reliability, Assurance, Sincerity, Personalization, and Formality have a significant and positive impact to the customer satisfaction at Islamic Bank in Palestine (Alnaser et al., 2017).
Besides that, Assurance, Personalization, Sincerity, and Formality significantly and positively impact customer satisfaction. The p-value is 0.01 for Sincerity, and the other is 0.00, respectively, making all indicators except for Reliability have a significant and positive impact. The value of these indicators showed that Assurance, Personalization, Sincerity, and Formality have the same output as Tangibility which significantly influences customer satisfaction and the result has the same result as the research conducted by (Alnaser et al., 2017). On the other hand, only Reliability has a p-value above 0.05, which is 0.07, even-though Reliability has a positive impact but is not significant towards customer satisfaction.
In addition, Tangibility, Assurance, Sincerity, Personalization, and Formality have a positive and significant impact on customer loyalty with the p-value 0.00, 0.02, 0.02, 0.003, 0.02, respectively. In contrast, Reliability has not significantly impacted customer loyalty. Furthermore, customer satisfaction will affect customer loyalty; this research shows that the p-value of customer satisfaction is below 0.05 (0.04).

CONCLUSION
This research investigates how the customer of Grab Food in Banda Aceh, Indonesia, could be loyal with the measurement of customer satisfaction and using the PAKSERV model. The previous study found that PAKSERV is a suitable model to use in the Asian market, as (Alnaser et al., 2017) stated that most PAKSERV indicators show a relation between Tangibility, Assurance, Reliability, Personalization, Sincerity, and Formality have a positive and significant impact towards customer satisfaction; thus this research also found that customer satisfaction is the main factor to make a customer loyal. This research is similar to the previous study on Service Quality and PAKSERV towards customer satisfaction. This research gave a broad view of the business people or entrepreneurs prioritizing services as the PAKSERV model uses Tangibility, Assurance, Reliability, Formality, Personalization, and Sincerity to determine how the service should be offered to the customer, which will satisfy the customer and later on make them loyal to a brand or service. The PAKSERV model in this research may be used in every business that operates in Asia since it uses using a different approach to western business.