The Influence Of Product Perceived Quality, Service, Environment And Assortment On Customer Loyalty With Customer Satisfaction As An Intervening Variable In Angkringan UMKM In Mojokerto Regency

. The development of the culinary business in Indonesia began to bloom in the 1970s with the rapid emergence of many restaurants which became an opportunity for companies to develop their business. One of these developments is the emergence of the trend of roadside angkringan which is now reaching millennial generation consumers and has become a new favorite sector that is worth developing. The presence of this angkringan answers the desires and needs of the lower middle class community and has big implications in driving the community economy.This research aims to find out the influence of Product Perceived Quality, Service, Environment, Assortment on Loyalty through Satisfaction. The expected benefit of this research is to increase the body of knowledge in the field of management, especially how big the influence of satisfaction is so that it increases loyalty which will ultimately increase purchases. The research method used is a quantitative method with data processing using AMOS. Data collection was carried out by distributing questionnaires to 130 respondents with the characteristics of male and female respondents aged 18-60 years, who visited several food courts spread across the Mojokerto Regency area at least twice a year. The research results show that Product Perceived Quality with a regression coefficient of 0.426 positively and significantly influences satisfaction; Service with a regression coefficient of 0.242 positively and significantly influences Satisfaction, Environment with a regression coefficient of 0.136 positively and significantly influences Satisfaction. Assortment, with a regression coefficient of 0.748, positively and significantly influences Satisfaction, and Satisfaction with a regression coefficient of 0.633 positively and significantly influences loyalty.


INTRODUCTION
The food and beverage industry is one of the mainstay manufacturing sectors in making a major contribution to national economic growth.Its performance achievements so far have been recorded as consistently positive or have increased from year to year, starting from its role in increasing productivity, investment, exports to employment.This is also influenced by the tourism sector which is increasingly emerging, there are many MSMEs which aim to train independence so that they can improve the prosperity of the economy.This picture further emphasizes that the food industry sector has great potential in increasing state income.Not only in the middle to upper market segmentation or what we more often call restaurants, depots or restaurants, but increasingly it is expanding to the MSME sector, one of which is the angkringan stalls on the side of the road which are increasingly in demand by millennials.
Starting from a trend that is developing in big cities like Jogjakarta, angkringan has its own position in the eyes of the people of Mojokerto who want to enjoy typical culinary and traditional snacks or just cure their longing for their home city, namely Jogjakarta.This moment has been developing since approximately 2017, starting from a few outlets but now has almost spread to all corners of the city of Mojokerto with an average number of visitors of 500 to 1000 people per day.Based on the information above, the author sees the continuity of academic research on customer satisfaction and loyalty.In this research the author included several variables such as product perceived quality, service, environment, assortment.This research aims to determine the extent of the mediating effect of customer satisfaction on the relationship between perceived product quality, service, environment and various factors on loyalty from angkringan stalls spread across Mojokerto Regency.

II. LITERATURE REVIEW Product Perceived Quality
Perceived product quality is a characteristic of a product or service that supports its ability to satisfy customer needs.From the definitions above, it can be concluded that product quality is the ability of a product to fulfill customer desires.Customer desires include product durability, product reliability, ease of use and other valuable attributes.(Kotler and Armstrong, 2012) Services (Service) Service can be viewed as a system consisting of two main components, namely service operations which are often invisible or unknown to customers (back office or backstage) and service delivery which is usually visible or known to customers (often also called front office or frontstage)", (Tjiptono, 2016) The definition of environment is described as an embodiment of the quality of the physical environment which can be considered important in influencing the level of customer satisfaction, especially because customer responses to the physical form of the environment are part of their consumption experience.(Ha and Jang, 2012).

Assortment(Product Diversity)
Ais product completeness which concerns the depth, breadth and quality of the products offered as well as the availability of these products at any time in the store.(Farli Liwe, 2015).Decisions regarding product placement are related to product availability/product diversity in appropriate quantities and in very appropriate locations (Tjiptono, 2015).

Loyalty(Loyalty)
LCustomer loyalty is a customer's deep commitment to re-subscribe or re-purchase selected products or services consistently in the future, even though situational influences and marketing efforts have the potential to cause changes in behavior.(Sangadji and Sopiah, 2013).

Satisfaction(Satisfaction)
Customer satisfaction can be defined simply, namely a condition where customer needs, desires and expectations can be met through the products consumed.Because customer satisfaction really depends on their perceptions and expectations, we as producers or parties providing services need to know several factors that influence this.(Nasution, 2001).

Types of research
This research aims to prove the causal relationship of several variables that have been explained in the previous chapter so that the research is called causal research.The variables in question are product perceived quality to satisfaction, service to satisfaction, environment to satisfaction, assortment to satisfaction and satisfaction to loyalty.The method used in this research is a quantitative method, namely quantitative research, namelyresearch method that uses data processing in the form of numbers as a tool for analyzing and conducting research studies, especially regarding what has been researched (Kasiram, 2008).In this research, the method used is the Structural Equation Model (SEM) method using AMOS 22.0 software.This research model is expected to be able to explain the relationship between variables in order to understand the factors that influence satisfaction and loyalty.

Population and Sample
In this study, the population referred to is all visitors who have visited and become customers at 25 angkringan spread across Mojokerto, while the sample is used to estimate population characteristics (Sumarni and Wahyuni, 2006).The researcher took a sample of 130 people who had visited and become customers and the technique used in sampling this research was Non Probability Sampling using the Snowball Sampling method.Non Probability Sampling is a sampling technique that provides unequal opportunities for each element or member of the population to was chosen as the sample, this is because the number of customers cannot be known with certainty.1423 3. Convert the flow diagram into equations for the research model 4. Selecting the input matrix and estimation technique.The input data used in this research is the covariance matrix or correlation matrix for the entire estimation.The sample size used in this research was 130 respondents who were consumers who had visited and made purchases at least twice a year.The software used is AMOS 22.0 with maximum likelihood estimation.

Assess the possibility of identification problems
Goodness of Fit Index The model identification process is in principle a problem regarding the inability of the model being developed to produce unique estimates.Symptoms of identification problems include: (1) very large standard error for one or more coefficients; (2) the program is unable to produce the information matrix that should be presented; (3) the appearance of strange numbers, such as a negative variance error; (4) the emergence of a very high correlation between the estimated coefficients obtained.
Model evaluation Testing of model suitability is carried out through a review of goodness of fit criteria.In summary, model interpretation and modification.In this final stage, model interpretation and modification of models that do not meet the test requirements will be carried out.The maximum likelihood estimation technique requires the normality assumption to be met.The condition for fulfilling the normality assumption is to use a critical ratio (CR) value of ±2.58 at a significance level of 1%.This means that if in the normality assessment table, the CR value is outside the range of ±2.58, then normality is not met (Ferdinand, 2002).Normality testing is carried out to find out whether the data distribution meets the normality assumption.If the data stated has met the normality requirements then the data is processed further using SEM modeling.This table shows that the minimum value in the CR column for each data is in the range -2.58 to +2.58.Therefore, there is no evidence that the distribution of this data is not normal so this data is suitable for use in further evaluation.

Univariate Outliners
Outlinersare observations that look very different from other observations and appear in the form of extreme values for either a single variable or a combination of variables (Hair et al., 1995in Ferdinand, 2002).The treatment of outliers depends on how they arise.Evaluation of outliers includes analysis of univariate outliers and multivariate outliers.4

Multivariate Outliners
Evaluation of multivariate outliers needs to be carried out because even though the data analyzed shows that there are no outliers at the univariate level, these observations can become outliers if they are combined with each other.This evaluation is carried out using the Mahalonobis Distance test which shows the distance of an observation from the average of all variables in a multidimensional space (Hair et al., 1995;Norusis, 1994;Tabachnuck andFidell, 1996 in Ferdinand 2002).This test was carried out using the Mahalonobis Distance criterion at the p < 0.001 level.Mahalonobis Distance is evaluated using Mahalanobis Distance is part of the Mahalanobis Distance evaluation results which have been sorted based on the Mahalanobis d-squared value.The results of data processing in the table show that the Mahalanobis dsquared value is smaller than 84.4031, so it can be concluded that in this data there are no multivariate outliers.

Evaluation of Multicollinearity and Singularity
Multicollinearity can be detected from the determinant of the covariance matrix.A very small determinant value of the covariance matrix indicates the existence of a multicollinearity or singularity problem (Tabachnick andFidell, 1998 in Ferdinand, 2002).In AMOS software, the application will immediately provide a warning if a singularity occurs in the covariance matrix.From the AMOS test results, it was found that the determinant of the 736281 matrix is far from zero.Therefore, it can be concluded that there is no evidence of multicollinearity or singularity in this combination of data variables, so this data can be analyzed further.

Confirmatory Factor Analysis (Confirmatory Factor Analysis)
Ferdinand (2002) explains that confirmatory factor analysis is aimed at estimating measurement models, testing the unidimensionality of exogenous and endogenous constructs.At this stage, the model will confirm whether the observed variables reflect the factors being analyzed.At the confirmatory analysis stage, exogenous constructs were tested to test the suitability of the model and the unidimensionality of the exogenous constructs.The measurement model in confirmatory analysis of exogenous constructs includes six independent variables used in this research, namely Analysis of Product Perceived Quality, Service, Environment, Assortment, Satisfaction and Loyalty.
Test In this table it can be seen that all CR values for the observation variables of Product Perceived Quality, Service, Environment, Assortment are greater than 2.00.So all of these indicators are significantly indicators of the latent factors they form.Therefore, all indicators are acceptable.The results of testing the Product Perceived Quality, Service, Environment, Assortment model can be seen in the table.Through this table, it can be seen that the loading factor (lamda) value for each variable is greater than 0.40.so these indicators together provide unidimensionality for the latent variable.

Confirmatory Factor Analysis of Endogenous Variables
At the confirmatory analysis stage, tests were carried out on endogenous constructs to test the suitability of the model and the unidimensionality of the endogenous constructs.The measurement model in the confirmatory analysis of endogenous constructs includes two dependent variables used in this research, namely satisfaction and loyalty.The satisfaction variable consists of 3 indicators, loyalty consists of 3 indicators.It is known that all CR values for the Perceived value, Customer satisfaction and Customer loyalty variables are greater than 2.00.Therefore, all of these indicators are significantly indicators of the latent factors they form.Therefore, all indicators are acceptable.The results of testing the Perceived value, Customer satisfaction and Customer loyalty construct models can also be seen in the table.Through this table, it can be seen that the factor loading (lamda) value for each variable is greater than 0.    4.5 shows that the Chi-square in this study has a value of308,805which means it is smaller than the cut off value obtained from the Chi-square table for df 160 with a significance level of 5%, namely 324.305.The significance probability value is above 0.05, which is equal to0.149.Therefore it can be concluded that the model is good.The CMIN/DF, GFI, AGFI, CFI, TLI, and RMSEA values can meet the goodness of fit criteria.

Full Structural Equation Modeling Analysis
Feasibility Testing Index

Hypothesis Testing Results
Looking at the calculation results using confirmatory factors and structural equation models, the model in this research is acceptable.The measurement results have met the goodness of fit criteria, namely chi-square = 308.805;significant probability = 0.149; RMSEA = 0.26; CMIN/DF = 1.087;TLI = 0.984; and CFI = 0.986.Next, based on this fit model, eight hypotheses proposed in this research will be tested.

Hypothesis Test Results
Hypothesis Analysis H1:Product Perceived Qualitysignificant effect on satisfaction.Accepted H2:Product Perceived Quality has a significant effect on loyalty Accepted H3:Service has a significant effect on satisfaction Accepted H4:Service has a significant effect on loyalty Accepted H5:Environmentsignificant effect on satisfaction Accepted H6:Environment has a significant effect on loyalty Accepted H7:Assortment significant effect on satisfaction Accepted H8:Assortment has a significant effect on loyalty Accepted H9:Satisfactionhas a significant effect on loyalty Accepted

IV. DISCUSSION AND RESULT Discussion of Research Results
H1: Product Perceived Quality has a significant effect on satisfaction of Angkringan customers in Mojokerto The estimated parameters between Product Perceived Quality and satisfaction show significant positive results with a CR value = 2.797 and a regression coefficient of 0.426 with an acceptable hypothesis standard, namely CR ≥ ± 2.00 with a significance level of <0.05 (5%), so it can be said that hypothesis 1 ( H1) is accepted.Therefore, hypothesis (H1) has a positive influence on satisfaction and needs to be maintained or improved product quality.
H2: Product Perceived Quality has a significant effect on loyalty among Angkringan customers in Mojokerto The estimated parameters between Product Perceived Quality and loyalty show significant positive results with a value of CR = 2.458 and a regression coefficient of 0.647 with a standard hypothesis that can be rejected, namely CR ≥ ± 2.00 with a significance level of <0.05 (5%), so it can be said that hypothesis 2 ( H2) is accepted.Therefore, hypothesis (H2) has a positive influence on loyalty and needs to be maintained or improved product quality.
H3: Service has a significant effect on satisfaction of Angkringan customers in Mojokerto The estimated parameters between service and satisfaction show significant positive results with a value of CR = 2.171 and a regression coefficient of 0.242 with a standard hypothesis that can be rejected, namely CR ≥ ± 2.00 with a significance level of <0.05 (5%), so it can be said that hypothesis 3 (H3) accepted.Therefore, hypothesis (H3) has a positive influence on satisfaction and the quality of service needs to be maintained or improved.
H4: Service has a significant effect on loyalty of Angkringan customers in Mojokerto The estimated parameters between service and loyalty show significant positive results with a value of CR = 2.013 and a regression coefficient of 0.316 with a standard hypothesis that can be rejected, namely CR ≥ ± http://ijstm.inarah.co.id 1430 2.00 with a significance level of <0.05 (5%), so it can be said that hypothesis 4 (H4) accepted.Therefore, hypothesis (H4) has a positive influence on loyalty and needs to be maintained or improved service quality.
H5: Environment has a significant effect on satisfaction of Angkringan customers in Mojokerto The estimated parameters between Environment and satisfaction show results that are not significant in the positive direction with a value of CR = 0.136 and a regression coefficient of 0.715 with a standard hypothesis that can be rejected, namely CR ≥ ± 2.00 with a significance level of <0.05 (5%), so it can be said that hypothesis 5 (H5 ) accepted.Therefore, hypothesis (H5) has a positive influence on satisfaction and the physical environment needs to be maintained or improved.
H6: Environment has a significant effect on loyalty of Angkringan customers in Mojokerto The estimated parameters between Enviornment and loyalty show significant results in the positive direction with a value of CR = 2.460 and a regression coefficient of 0.645 with a standard hypothesis that can be rejected, namely CR ≥ ±2.00 with a significance level of <0.05 (5%), so it can be said that hypothesis 6 (H6) accepted.Therefore, hypothesis (H6) has a positive influence on loyalty and the physical environment needs to be maintained or improved.
H7: Assortment has a significant effect on satisfaction of Angkringan customers in Mojokerto The estimated parameters between Assortment and satisfaction show significant positive results with a value of CR = 2.322 and a regression coefficient of 0.748 with a standard hypothesis that can be rejected, namely CR ≥ ± 2.00 with a significance level of <0.05 (5%), so it can be said that hypothesis 7 (H7) accepted.Therefore, hypothesis (H7) has a positive influence on loyalty and needs to be maintained or improved product quality.
H8: Assortment has a significant effect on loyalty of Angkringan customers in Mojokerto The estimated parameters between Assortment and loyalty show significant positive results with a value of CR = 2.244 and a regression coefficient of 0.807 with a standard hypothesis that can be rejected, namely CR ≥ ± 2.00 with a significance level of <0.05 (5%), so it can be said that hypothesis 8 (H8) accepted.Therefore, hypothesis (H8) has a positive influence on satisfaction and needs to be maintained or increased in the amount of product diversity.
H9: Satisfaction has a significant effect on loyalty among Angkringan customers in Mojokerto The estimated parameters between Product variety and Perceived value show significant positive results with a value of CR = 2.857 and a regression coefficient of 0.633 with a standard hypothesis that can be rejected, namely CR ≥ ± 2.00 with a significance level of <0.05 (5%), so it can be said that hypothesis 9 ( H9) is accepted.Therefore, hypothesis (H9) has a positive influence on loyalty and needs to be maintained or increased in the amount of product diversity.
V. CONCLUSION 1. Product Perceived Quality has a significant influence on customer satisfaction at Warung Angkringan in Mojokerto 2. Product Perceived Quality has a significant influence on customer loyalty at Warung Angkringan in Mojokerto.
3. Service has a significant influence on customer satisfaction at Warung Angkringan in Mojokerto.4. Service has a significant influence on customer loyalty at Warung Angkringan in Mojokerto.5. Environment has a significant influence on customer satisfaction at Warung Angkringan in Mojokerto.
6. Environment has a significant influence on customer loyalty at Warung Angkringan in Mojokerto.
7. Assortment has a significant influence on customer satisfaction at Warung Angkringan in Mojokerto.
8. Assortment has a significant influence on customer loyalty at Warung Angkringan in Mojokerto.9. Satisfaction has a significant effect on customer loyalty at Warung Angkringan in Mojokerto.
After the model is analyzed through confirmatory factor analysis, each indicator in the fit model can be used to define the latent construct.This means creating a full SEM (Structural Equation Model) model.Apart from that, the model is also analyzed to evaluate the degree of goodness of fit between the model and money data.The measurement model for analyzing full Structural Equation Modeling includes 6 variables used in this research, namely Product Perceived Quality, Service, Environment, Assortment, Satisfaction and Loyalty using 26 indicators.

Fig 4 . 5 .
Fig 4.5.Full Structural Equation Model Table4.5 shows that the Chi-square in this study has a value of308,805which means it is smaller than the cut off value obtained from the Chi-square table for df 160 with a significance level of 5%, namely 324.305.The significance probability value is above 0.05, which is equal to0.149.Therefore it can be concluded that the model is good.The CMIN/DF, GFI, AGFI, CFI, TLI, and RMSEA values can meet the goodness of fit criteria.Feasibility Testing Index .2.3.2.1 Univariate Outliners Testing for the presence of univariate outliners can be done by determining the threshold value that will be categorized as an outliner by converting the research data values into a standard score or what can be called a z-score, which has a mean of zero with a standard deviation by one.For large samples (above 80 observations), the evaluation guideline is that the threshold value of the z-score is in the range -4 to4 (Hair et al., 1995 in Ferdinand, 2002).Based on the results of the conversion to z-score values shown in table4.24, it can be seen that the maximum and minimum values for all variables are in the range -4 to 4, so there are no univariate outliers in this research data.

Factor Weights and Factor Loading Values Product Perceived Quality, Service, Environment and Assortment.
40.So these indicators together provide unidimensionality for the latent variable.