The Influence of Service Quality, Destination Image, and Memorable Experience on Revisit Intention with Intervening Variables of Tourist Satisfaction

Purpose: This study aims to build a model for tourists' revisit intention who come to visit tourist destinations. Research methods: This study uses primary data with 400 samples assigned by random sampling. The data were processed using the Structural Equation Model (SEM). Results and discussions: In total (total effect) of each variable, which has the most considerable full effect, is given the destination image followed by service quality, tourist satisfaction, and memorable experience, respectively. For the mediation role, it is found that the tourist satisfaction variable is a mediator of the relationship between an unforgettable experience and revisit intention. Tourist satisfaction is not a mediator of the relationship between service quality and revisit intention and tourist satisfaction, nor is it a mediator of the relationship between destination image and revisit intention. Regarding revisit intention, the analysis results show that this variable is significantly influenced by the variables of service quality, destination image, and tourist satisfaction. Meanwhile, the effect of memorable experience is not statistically significant. Conclusion: The results of this study prove that efforts to increase tourists' revisit intention can be made by improving destination management that considers service quality, destination image, memorable experience, and tourist satisfaction.


Introduction
Currently, Indonesia is focused on promoting the tourism sector in supporting the national economy. The government wants the tourism sector to become the prime mover of the national economy. Tourism itself is classified into the world's largest industry group. It is evidenced by the strategic contribution of domestic income from the tourism sector, a contributor to foreign exchange, and many new jobs for the community. Even the tourism sector can also be a safeguard for the country's economy when a crisis hits the state. The role of the tourism sector includes as a source of foreign exchange. It's because every foreign tourist who comes to visit must convert their currency into rupiah. On the one hand, this conversion will strengthen the rupiah's value and support the country's foreign exchange. The more foreign tourists who come to visit, the more foreign exchange will be collected.
Another thing from tourism activities that can benefit the country's economy is that if the tourist destination in a developing area develops, the site will develop sooner or later. There will be improvements to facilities and infrastructure; there will be built hotels, various restaurants, souvenir shops, and multiple services such as SPA, barbershop, laundry, etc. Furthermore, it will trigger the emergence of diverse employment opportunities, either directly or indirectly, to support tourism activities. Of course, this will then be able to reduce the unemployment rate around tourism destinations.
It must be realized Indonesia itself is a country with tremendous and strategic potential in the tourism sector. Indonesia has a beautiful landscape, and this natural beauty has been recognized by tourists who visit. Not only is its natural beauty attractive, but Indonesia is also unique in its social, cultural, and historical significance, which continues to increase the national tourism competitiveness ranking. Based on the 2019 World Economic Forum, Indonesia is ranked 40 in tourism competitiveness (Prodjo, 2019).
It is also essential to strive to increase the number of tourist visits to a destination. With so many tourist visits, the stretch of tourism will continue to grow. Nevertheless, maintaining the desire of tourists to return to visit (revisit intention) is also essential. Revisit intention is an exciting thing to study when this concept has been transformed from the idea of product marketing in marketing science and then shifted to tourism marketing. Theoretically, revisit intention is a stage from post-purchase or consumption of a product or service. At the post buying stage, satisfied consumers (in this case, tourists) will make a return visit to a tourist destination and even join in saying good things about the tourist destination (words of the mouth).
The revisit intention of tourists is currently increasingly becoming the attention of researchers in the tourism sector. Many studies have examined this revisit intention, among others; Luo et al.  (2019), Seetanah et al., (2020), Sitepu et al., (2020), etc. Luo et al. (2011 stated that experiential marketing's contribution positively and significantly affects revisit intention. Yuniawati & Finardi (2016) explained that customer experience positively and significantly impacts revisit intention. Hung et al. (2016) linked creative experiences and memorability with revisit intention in creating active tourism activities. Loi et al. (2017), in their research, suggest that destination image and tourist satisfaction affect revisit intention. Kim et al. (2017) analyzed the impact of weather and tourist satisfaction on revisit intention. Park et al. (2018) examined the influence of nostalgia and tourist satisfaction related to revisiting intention. (2018) suggest how a tourist destination's image has an essential effect on revisit intention. Rusdin & Rashid (2018) describe decision and service quality in influencing revisit intention. Sitepu et al. (2020) argue that revisit intention is influenced by the implementation of sustainable tourism, tourist satisfaction, perceived experience, and tourist decisions.
From a number of these studies, researchers generally only linked some variables partially, for example, tourist satisfaction with revisit intention or tourist memorable experience with revisit intention or destination image and satisfaction associated with revisit intention. The research gap that concerns researchers is to formulate a revisit intention model that is more complete than the previous models. In this study, three exogenous variables were used: service quality, destination image, and memorable experience. The three variables are thought to have a direct effect on revisit intention. Furthermore, indirectly, the three exogenous variables affecting revisit intention through the intervening tourist satisfaction variable. The objectives of this research activity are 1) To find out and analyze how service quality, destination image, and memorable experience directly affect tourists' revisit intention who come to visit Medan City, 2) To find out and analyze whether service quality, destination image, and memorable experience do not directly affect revisit intention through the intervening tourist satisfaction variable, 3) An alternative model was found for efforts to increase tourist visits to Medan City.

Methodology
This study uses primary data by distributing questionnaires to respondents who are determined by random sampling. The number of respondents used was 400 people, with criteria, among others; Never been to Medan City, aged 17 years and over, come from within the country or abroad. For domestic respondents, the respondents must come from a city other than Medan (excluding districts/cities directly adjacent to Medan, such as Binjai and Deli Serdang. Data collection methods in this study were questionnaires and questions/statements in this research. Uses a Likert scale, which means a scale of 6 to measure a person's attitudes, perceptions, and opinions about various social problems (Sugiyono, 2016).
Furthermore, the data were analyzed through the stages required to examine the Structural Equation Model (SEM). The steps started from the validity test, reliability test, normality test, outlier test, and model modification according to modification indices and the goodness of fit model (GoF) test. GoF itself includes; Chi-Square value (X2) to test whether there is a deviation or level of compatibility between the sample covariance matrix and the model covariance matrix and the importance of goodness of fit indices (GFI), which is a measure of the accuracy of the model in producing the observed matrix covariances.
This GFI value ranges from 0 -1. The closer to number 1, the model is declared the better. Besides, other criteria that must be met are also considered, such as the value of the Adjusted Goodness of Fit Index (AGFI), Root means the square error of approximation (RMSEA), Comparative Fit Index (CFI), and others. A review was also carried out on the standardized total effect output, direct effect, and indirect effect (Kline, 2015). The SEM model, which is built based on the theoretical framework and previous research as follows: (Source: Author data, 2020)

Results and Discussions Results
Of the 400 questionnaires distributed to respondents, 400 were returned and received by researchers, and no questionnaire was damaged. The questionnaire that deserves to be analyzed is as many as 400 questionnaires, and then the response rate is 100%. Based on the results of filling in the respondents from the returned questionnaires, it can be obtained a description of the respondents' characteristics based on gender. Based on gender, 53% of the 400 respondents were male, and 47% female. Based on marital status, most respondents were single as many as 243 people (60.8%), while respondents who were married were 157 people (39.3%). Based on the age of the respondents, the majority of respondents were 20-29 years old, as many as 183 people (45.8%). For ages 30-39, as many as 169 respondents (42.3%). For respondents aged 40-49, as many as 41 people (10.3%) and respondents aged 50-59, as many as seven people (1.8%).
Furthermore, based on the respondents' education level, it can be seen that the number of respondents with a high school education level was 109 people (27.3%). In comparison, respondents with a diploma education level were 158 people (39.5%). There were 106 respondents with an undergraduate education level (26.5%) and 22 respondents with a master's degree (5.5%). Respondents with doctoral education are five people (1.3%). As for tourists' origin, 70% of tourists who became respondents came from within the country, and 30% came from abroad.
The revisit intention variable' in this study is measured using four critical statements in a questionnaire based on four indicators. Based on the table, the overall average mean is 4.609. Of these four indicators, the one with the highest average value is Sense of Place (RI_2)-the taste when visiting a tourist spot is according to what tourists imagine with a value of 4.682. In second place is Novelty Seeking (RI_4)-tourists discover new and unique things in their activities to visit tourist destinations with a mean value of 4.588. In third place is Attachment to Place (RI_3)tourist destinations have provided complete facilities and various facilities in services that tourists want with a mean value of 4.570. While the fourth place is Past visit (RI_1)-tourists are satisfied with the experience they felt on the previous visit with a mean value of 4.568. This study's tourist satisfaction variable is measured using critical statements in a questionnaire based on four indicators. Based on the table, the overall average mean is 5.045. Of the four indicators with the highest average value, namely the price and costs incurred (TS_4), according to tourists' experience when visiting, the expenses and costs incurred for tourism activities are appropriate. The TS_4 value is 5.128. Second is the quality of service (TS_2) tourism facilities and infrastructure at the destination are already good and according to tourist expectations with a mean value of 5.060. Destination Quality (TS_1) of various existing attractions and culinary attractions in tourist destinations is under tourist expectations, being in third place with a mean value of 5.008. The fourth place is emotional closeness (TS_3). The tourists are emotionally interested in this destination, with a mean value of 4.985. This study's service quality variable was measured using five critical statements in a questionnaire based on five indicators. Based on the table, the overall average mean is 4.609. Of the five indicators with the highest average value, namely Reliability (SQ_2), the destination's ability to provide services as promised accurately and reliably has gone well, with a mean value of 4.682. Second is empathy (SQ_5), where people and operators in tourist destinations have communicated well and understand consumer wants, with a mean value of 4.640. Assurance (SQ_4), people and operators in tourist destinations have the competence, credibility, and provide a sense of security to tourists in third place with a mean value of 4.588. The fourth place is Responsiveness (SQ_3); people in tourist destinations are willing to help and provide fast and precise service, with a mean value of 4.570. Meanwhile, in fifth place is Tangible (SQ_1), the appearance of physical facilities, equipment, personnel, and communication materials in tourist destinations is good, with a mean value of 4.568. This study's destination image variable was measured using three critical statements in a questionnaire based on three indicators. Based on the table, the overall average mean is 5.009. Of the three indicators that have the highest average value, namely Cognitive Image (DI_1), according to natural and cultural resource tourists, tourism infrastructure in general, climate, social and natural environment 3s (sea, sand, and sun) are acceptable as destinations, with a mean value of 5.033. The second place is the Unique Image (DI_2); this destination has a uniqueness or uniqueness different from other destinations, with a mean value of 5.015. In the third place is Affective Image (DI_3). This destination has tourist attractions, comfort, value for money, and an exotic atmosphere of a tourist destination that tourists imagine, with a mean value of 4.980. To measure this study's destination image variable, six key statements were used in the questionnaire based on six indicators. Based on the table, the overall average mean was 4.833. Of the six indicators used, the highest mean value is found in Physical Experience and Lifestyle (ME_4), with a mean value of 5.015. In second place is the social identity experience resulting from the reference group or culture (ME_5) with a mean value of 4.880. In the third-order are destination names, visual symbols, verbal slogans, jingles, giving a good impression of a tourist destination (ME_6) with a mean value of 4.875. In fourth place is Cognitive Experience (ME_3), with a mean value of 4.788. The fifth place is Affective Experience (ME_2) with a mean value of 4.742 while in the 6th position is Sensory Experience (ME_1); colors, music, sounds, raw materials, taste, texture give me a distinct impression which adds to my memories of the destination, with a mean value of 4.700. The next stage is the validity test and the reliability test. The validity test is carried out by the convergent validity test to test the construct (indicator), whether it has a high proportion of variance. Meet the criteria if the value of C.R. > 1.96, while the loading factor or standardized loading estimate> 0.5. To facilitate the results of this validity test, it is carried out directly using the confirmatory factor analysis technique as on figure 2 below.

Figure 2. Validity Test Results Using CFA
The reliability test is done using a construct reliability test, which tests the data's reliability and consistency. Fulfills the criteria if the value of construct reliability > 0.7. The value of construct reliability between 0.6 to 0.7 is still acceptable, provided that the construct validity (indicator) in the model is good. Ghozali (2013) explains that a variable's indicator is called reliable if the AVE value is ≥ 0.05 and CR ≥ 0.07. After analyzing the latent variable forming indicators' validity and reliability, the subsequent analysis is the full model Structural Equation Modeling (SEM) analysis. Analysis of the data processing results at the entire model SEM stage was carried out by conducting a model feasibility test and a causality significance test. The path diagram for the complete analysis of the invalid and reliable indicator models has been aborted and is presented in Figure 3. Based on the figure, it can be seen that the model's feasibility test value has shown a fit model. So there is no need to modify the model based on modification indices according to AMOS recommendations.

Discussions
The results of the data normality test shown showed that no CR value was outside +2.58. So it can be concluded that the univariate is good. The normality test was carried out using the critical ratio criteria of ± 2.58 at a significance level of 0.01 (1%) (Ghozali, 2012), so it can be concluded that there is no deviant data. Likewise, these results are supported by the outlier test that has been carried out. Multivariate outliers test in the Structural Equation Modeling analysis, the outliers' evaluation can be seen at the Mahalanobis distance value at the level of p <0.05. Mahalanobis distance is evaluated using the chi-square degree of freedom equal to the number of indicators used in the study. If the Mahalanobis distance is greater than the chi-square value, it is categorized as a multivariate outlier. Based on the Chi-square value with 71 degrees of freedom (number of variable indicators) at a significance level of 0.001, namely 425, the Mahalanobis value that exceeds or is above 425 identifies the existence of multivariate outliers data. Based on the table above, it can be seen that the highest value lies in the third observation of 192.761, which is still below 425. Based on this description, it can be concluded that there are no multivariate outliers from the data used in this study, so all data is free from outliers. Then do the goodness of fit test. The results are presented in Table 7 below, which shows all the essential indicators that the model used is good and meets the required goodness of fit criteria.  Table 8. It can be seen that based on several relationships built between the variables of this study, all relationships are positive, and the majority are statistically significant. There is only a minority relationship that is not statistically significant, among others, the influence of Service Quality on Satisfaction, the impact of Service Quality on Revisit Intention, and the effect of Destination Image on Revisit Intentions. This result is in line with several studies previously stated, among others: This result is in line with some studies previously stated, among others; Luo et al. (2011), Yuniawati & Finardi (2016, Hung et al. (2016), Loi et al. (2017), Kim et al. (2017), Park et al. (2018, Li et al., (2018), Rusdin & Rashid (2018), Sthapit & Björk (2019), Seetanah et al., (2020), Sitepu et al., (2020), etc.  Table 9 shows how the total effect of the variables used in this study, where memorable experience has a sweeping impact on revisit intention of 0.156. The impact of total service quality on revisit intentions is 0.298, and the effect of real destination image on revisit intention is 0.299. The impact of complete tourist satisfaction on revisit intention is 0.286. Thus, the destination image's most considerable total influence is followed by service quality, tourist satisfaction, and memorable experience. For the real effect of each variable used in influencing tourist satisfaction, it can be seen that the largest order starts with the impact of a memorable experience of 0.402, followed by a destination image of 0.390 followed by service quality of 0.053. The direct effect (direct effect) of each variable used to influence revisit intention is presented in Table 10. Memorable experience has a direct impact on revisit intention by 0.041. Service quality has an immediate effect on revisit intention of 0.282. The destination image directly affects revisit intention of 0.187, and Tourist Satisfaction directly impacts revisit intention of 0.286. So that the order of variables that most influences revisit intention is Tourist Satisfaction, Service Quality, Destination Image, and Memorable Experience. For the direct effect of each variable used in influencing tourist satisfaction, it can be seen that the most extensive sequence starts with the impact of a memorable experience of 0.402 followed by a destination image of 0.390 followed by a service quality of 0.053. Meanwhile, the indirect effect of each variable used in influencing revisit intention is presented in Table 11. If sorted based on the most significant influence, then memorable experience has an indirect effect on revisit intention of 0.115. The destination image is having an indirect impact on revisit intention of 0.111. The service quality has an effect of 0.015. Furthermore, it can be concluded that by considering the comparison between this direct and indirect effect: 1) For the memorable experience variable, the direct impact is 0.041, while the indirect effect is 0.115. Thus the immediate effect <indirect effect, it can be said that there is a role of mediation in this model. In other words, tourist satisfaction is a mediator of the relationship between a memorable experience and revisit intention. 2) For the service quality variable, the direct effect value is 0.282, while the indirect effect is 0.015. In this case, the direct impact> indirect effect, it can be said that there is no mediation role in this model. In other words, tourist satisfaction is not a mediator of the relationship between service quality and revisit intention. 3) For the destination image variable, the direct effect value is 0.187, while the indirect effect is 0.11. In this case, the direct impact > indirect effect, it can be said that there is no mediation role in this model. In other words, tourist satisfaction is not a mediator of the relationship between destination image and revisit intention.

Conclusions
The results of this study indicate that the variable service quality is not significant in affecting tourist satisfaction, destination image has a significant effect on tourist satisfaction, and memorable experience has a significant impact on tourist satisfaction. On the other hand, it is significantly influenced by the variables of service quality, destination image, and tourist satisfaction related to revisiting intention. Meanwhile, the effect of memorable experiences is not significant.
For each variable's total effect (total effect), the largest total effect is given respectively by the destination image followed by service quality, tourist satisfaction, and memorable experience. Furthermore, for the real impact of each variable used in influencing tourist satisfaction, it can be seen that the largest order starts with the effects of a memorable experience, followed by a destination image of the size, then followed by service quality. As for determining the role of mediation, it is found that the tourist satisfaction variable is a mediator of the relationship between a memorable experience and revisit intention. Tourist satisfaction is not a mediator of the relationship between service quality and revisit intention and tourist satisfaction, nor is it a mediator of the relationship between destination image and revisit intention.