AN APPROACH TO DETERMINING CUSTOMER SATISFACTION IN TRADITIONAL SERBIAN RESTAURANTS

The aim of this paper is to make a proposal for an easy–to–use approach to the evaluation of customer satisfaction in restaurants. In order to provide a reliable way to collect respondents’ real attitudes, an approach based on the use of smaller number of evaluation criteria and interactive questionnaire created in a spreadsheet file is proposed in this paper, whereby an easy-to-understand and simple-touse procedure is proposed for determining weights of criteria. In addition to the said, the proposed approach applies the simplified SERVQUAL-based approach, for which reason a simplified version of the Weighted Sum Method based on the decision maker’s Preferred Levels of Performances is used for the final ranking of the alternatives. The usability of the proposed approach is considered in the case study intended for the evaluation of traditional restaurants in the city of Zajecar.


Introduction
The Serbian word "kafana" originates from the Turkish word "kahvehane", which means "a place for drinking coffee". Such places have emerged in the Balkan region under the influence of the Ottoman Empire in the 16th century.
Under the influence of different cultures, kafana generated its specificity on the Balkan Peninsula, so that it also became a place where food was consumed and later a place where alcoholic drinks were served. Over time, kafanas have increasingly become and have found their place in the social and cultural life, as well as in business. Nowadays, kafanas continue to be a place where you meet your friends, a place for celebrations, talking about and discussing things and so on. Therefore, kafanas could be denoted as traditional Serbian restaurants. Compared with the other types of restaurants, kafanas have similarities to taverns and pubs, as places of a pleasant ambience.
Certain new trends in the restaurant and food industry, as well as the growing presence of various cuisines, have had an impact on traditional Serbian restaurants. Fortunately, in some parts of Serbia, traditional Serbian restaurants somehow still resist unfortunately unstoppable trends.
In the city of Zajecar, located in eastern Serbia, traditional restaurants are successfully resisting the actual trends and it is still possible for you to find good restaurants, such as: "Dva brata" ("The Two Brothers"), "Gradska Mehana" (The City Meyhane"), "Meda" ("The Bear"), "Roko" ("The Roko") and so forth.
The factors influencing the satisfaction of restaurants' customers have been considered in many previous studies. Based on these studies, an approach to the determining of the significance of the relevant factors that influence customer satisfaction is proposed.
The proposed approach also uses the concept of measuring the difference between expectations and perceptions, so it provides an easy identification of the criteria against which customer expectations are not met. Beside all of the above-said, the proposed model can also be used to determine the overall ratings of the considered alternatives, thus making a comparison with competitors.
Based on all of the above-mentioned reasons that have been taken into account, the remaining part of this paper is organized as follows: In Section 2, a review of the relevant research studies is given. After that, in Section 3 and Section 4, the PIPRECIA and the WS PLP methods are considered. In Section 5, an empirical illustration of the evaluation of Serbian traditional restaurants, based on the integrated use of the PIPRECIA and the WS PLP methods, is presented in detail. Finally, the conclusions are given at the end of the paper.

Literature Research
Measuring customer satisfaction could be very important in a competitive environment (e.g. Stepaniuk 2018; Raudeliūnienė et al. 2018). For the purpose of determining that, Parasuraman et al. (1988) proposed the Service Quality and Customer Satisfaction (SERVQUAL) model. On the basis of that model, many others more specialized models have been proposed later, such as: WebQual (Loiacono et al. 2002;Parasuraman et al. 2005), eTailQ Wolfinbarger and Gilly (2003), E-RecS-QUAL (Parasuraman et al. 2005), and eTransQual (Bauer et al. 2006).
The SERVQUAL model was used for determining the levels of customer satisfaction in many different areas. As one of these areas, tourism and hospitality can be mentioned. For example: Saleh and Ryan (1991)  SERVQUAL to determine the gap between clients' and the management's perceptions in the hotel industry, whereas Devi Juwaheer (2004) explore the tourists' perceptions about hotels in Mauritius by using an adapted SQRVQUAL approach. Further, on the basis of the SERVQUAL model, Tribe and Snaith (1998) proposed the HOLSAT model, adapted for determining tourists' satisfaction with their holidays.
Some other studies have also been dedicated to the restaurant industry. For example: Adam et al. (2015) investigates tourist satisfaction with Ghanaian restaurants based on a factor analysis, and Jung and Yoon (2013) investigate the relationship between employees' satisfaction and customers' satisfaction in a family restaurant. Dobrovolskienė et al. (2017) state that decision making is crucial to every aspect of business. Multiple-criteria decision-making (MCDM) is a scientific field that has undergone extremely rapid development over the last two decades. Multiple-criteria decision-making considers situations in which the decision-maker must choose one of the alternatives from a set of available alternatives and which are judged on the basis of a number of criteria. This is why MCDM contributes to easier decision-making and adoption of long-term and lasting solutions.
MCDM has also been successfully applied in the hospitality industry. Chou et al. (2008) and Tzeng (2008) used MCDM models for selecting the restaurant location. Yildiz and Yildiz (2015) proposed a model for evaluating customer satisfaction in restaurants, based on the use of the AHP and TOPSIS methods. In their studies: Duarte Alonso et al. (2013), Chi et al. (2013), Kim et al. (2007), Yuksel and Yuksel (2003) and Jack Kivela (1997) investigate the criteria that have an impact on customer preferences and satisfaction.

The PIPRECIA Method
The The SWARA method has a certain similarity with the prominent AHP method. The first similarity is that both methods can be used to completely solve MCDM problems or to only determine the weight of the criteria; the second is that both methods are based on the use of pairwise comparisons.
However, the computational procedures of the SWARA and the AHP methods significantly differ from one another. Because of that, the SWARA method has some advantages, as well as some disadvantages, in comparison with the AHP method.
As the main disadvantage of the SWARA method, the fact that its computational procedure does not include a procedure for determining the consistency of pairwise comparisons made can be mentioned. Contrary to that, a significantly lower number of pairwise comparisons required for solving an MCDM problem and for determining criteria weights, too, can be mentioned as an advantage of the SWARA method.

ENTREPRENEURSHIP AND SUSTAINABILITY ISSUES
ISSN 2345-0282 (online) http://jssidoi.org/jesi/ 2019 Volume 6 Number 3 (March) http://doi.org/10.9770/jesi.2019.6.3 (5) Its requirement that evaluation criteria should be sorted in descending order according to their expected significances, which can prove to be inadequate in some survey cases, can also be mentioned as the weakness of the SWARA method. Therefore, with the aim of extending the use of the SWARA method in the cases where a consensus on the expected significance of the criteria is not easy to reach,  proposed the use of the following equation for the purpose of determining the importance of criteria as follows: (1) where: sj denotes the comparative importance of the criterion j, and denotes the significance of the criterion j in relation to the j-1 criterion.
In an extension of the SWARA method, proposed under the name of PIPRECIA,  also mention that a lack an integrated procedure for checking the consistency in the ordinary SWARA method can successfully be compensated for by using Kendall's Tau or Spearman's Rank Correlation Coefficient.
Because of all the foregoing, the PIPRECIA method has been chosen to be used in this approach.

The Computational Procedure of the PIPRECIA Method
The computational procedure of the PIPRECIA method can be shown as follows: Step 1. Choose the criteria on the basis of which an evaluation of alternatives will be carried out.
Step 2. Set the value of the relative importance of the criteria by using Eq. (1), starting from the second criterion.
Step 3. Calculate the coefficient kj for the criterion j as follows: (2) Step 4. Calculate the recalculated weight qj for the criterion j as follows: (3) Step 5. Calculate the weights of the criteria as follows: where wj denotes the weight of the criterion j.

The WS PLP Approach
Based on the Weighted Sum Method (Churchman andAckoff, 1954, MacCrimon, 1968),  proposed the Weighted Sum Preferred Levels of Performances (WS PLP) approach.
The simplified computational procedure of the WS PLP approach for solving an MCDM problem that contains the m alternatives that are evaluated based on the n beneficial criteria (a higher value of the performance rating is desirable) can be shown as follows: Step 1. Evaluate the alternatives in relation to the selected criteria.
Step 2. Set the preferred performance ratings for each criterion.
Step 3. Calculate the normalized performance ratings of the alternatives as follows: where: ij x and ij r denote the performance rating and the normalized performance rating of the alternative i in relation to the criterion j, respectively; j x 0 denotes the preferred performance rating of the criterion j; Step 4. Calculate the overall performance rating of the alternatives as follows: where i S denotes the overall performance rating of the alternative i, ; wj is the weight of the criterion j.
In the proposed approach, the alternatives whose i S is greater than or equal to zero make a set of the most appropriate alternatives, out of which one should be selected.

A Case Study
In order to determine the preferences of the passionate visitors of Serbian traditional restaurants, a supervised survey has been performed in the city of Zajecar, located in Serbia, near the Romanian and the Bulgarian borders.
In this study, the five previously mentioned restaurants have been evaluated on the basis of the six criteria adopted from Stanujkic et al. (2016):  C1 -the interior of the building and the friendly atmosphere,  C2 -the helpfulness and friendliness of the staff,  C3 -the variety of traditional food and drinks,  C4 -the quality and taste of the food and drinks, including the manner of serving,  C5 -the appropriate price for the quality of the services provided, and  C6 -other.
In the proposed approach the criterion "other" is used to enable personalization.

ENTREPRENEURSHIP AND SUSTAINABILITY ISSUES
ISSN 2345-0282 (online) http://jssidoi.org/jesi/ 2019 Volume 6 Number 3 (March) http://doi.org/10.9770/jesi.2019.6.3 (5) The survey presented in this study was conducted by e-mail, or more precisely by using an interactive questionnaire created in a spreadsheet file. By using such an approach, the respondents can see the calculated weights of the criteria and can also modify his/her responses if he or she is not satisfied with the obtained results. In addition, by using such an approach, the obtained results can also be presented graphically, which can make easier to understand the procedure used for determining weights of criteria, and thus lead to obtaining more realistic views of the respondents.
The interactive questionnaire was sent to the selected respondents known as the "bohemians" and/or frequent visitors of traditional Serbian restaurants. Out of the approximately 80 sent questionnaires, the 42 of them were returned, out of which only 30 questionnaires were selected as those properly filled in.
The weights of the criteria calculated on the basis of the responses obtained from the two selected respondents are accounted for in Table 1 and Table 2.  The interior of the building and friendly atmosphere 0.15 C2 The helpfulness and friendliness of the staff 1.10 0.17 C3 The variety of traditional food and drinks 0.90 0.16 C4 The quality and taste of the food and drinks, including the manner of serving 1.15 0.18

C5
The appropriate price for the quality of the services provided 0.95 0.17

Source: Own calculations
Some significant descriptive statistical parameters related to the weights of the criteria obtained by the conducted survey are presented in Table 3.  Table 3, the criteria C4 and C5 have significantly higher importance related to the other criteria, i.e. the quality and the taste of the food and the appropriate price are identified as the most significant criteria.
The obtained correlation coefficient between the responses obtained from the respondents and the mean ranges between 0.44 and 0.98. Criterion C6 -"other" also has a high weight, which can be interpreted as follows:  in addition to the criteria C1 -C5 there are other criteria that affect satisfaction of restaurant customers, which can be applied in much more sophisticated models, and  criterion C6 can successfully substitute many less significant criteria and such enable forming an efficient MCDM models based on the use of a smaller number of criteria.
In addition to the conducted research, the respondents also evaluated the five preselected traditional restaurants by using the five-point Likert Scale. The results obtained from the two of the above-mentioned respondents are accounted for in Tables 4 and 5.  Ranges between the maximum and minimum weights of criteria are also not negligible, as previously shown in Table 3. Therefore, the separate ranking list of considered alternatives has been formed for each respondent, in this approach, by using the WS PLP approach.
In this way, the attitudes of the respondents do not drown into the group attitudes, obtained on the basis of the average weight of and average ratings, and remain clear until the end of the evaluation, where the final ranking of the considered alternative was made based on dominance theory.

1134
The results achieved based on all properly filled questionnaires are shown in Table 6. The appearance of the considered alternative in the first position is given in Column I of Table 6. The appearance of the considered alternatives in the second and the third positions is given in Columns II and III of Table 6. According to Column I of Table 5, based on the dominance theory, the best-placed alternative is the alternative labelled as A1.
In this approach, only the appearances on the first position are used for the determination of the best alternative, or more precisely, the most popular traditional restaurant. The appearances in the second, the third, as well as the other positions, could be used for a further analysis.
The overall ratings, obtained by using WS PLP approach, can also be used for various analysis, especially when it is known that WS PLP approach Si < 0 indicates an alternative where expected customers' satisfaction has not been reached yet.

Conclusions
The main objective of this paper is to determine the most significant criteria that have an influence on customers' satisfaction in traditional Serbian restaurants, as well as weights of these criteria, and propose an easy-to-use approach for the evaluation of customers' satisfaction in restaurants.
For that reason, the newly proposed PIPRECIA method, that is an extension of the SWARA method, is proposed for determining the weight of criteria in order to provide an effective and simple-to-use procedure for gathering the attitudes of the examined respondents that will be as realistic as possible.
The gaps between the expected and the achieved satisfaction obtained based on a set of criteria are used to determine the overall performance of any of the considered alternatives, which is done by applying the WS PLP approach. The final ranking of the alternatives is made by referring to dominance theory.
The approach proposed in this paper has significant similarities to the proven SERVQUAL model or models like that one. However, it is based on the use of a significantly smaller number of evaluation criteria, which could allow the forming of the simplest questionnaires that could be more appropriate when preferences and ratings are collected through conducting surveys with ordinary respondents, i.e. those unprepared in advice for surveying.
The usability of the proposed approach has been verified in the case study on the evaluation of traditional Serbian restaurants. The achieved results confirm the efficiency and usability of the proposed approach for solving similar, as well as numerous other, decision-making problems.