Consumers’ Propensity to Complain Towards Tourist Agencies’ Services: Evidence from Serbia and Croatia

The goal of the research is to determine a statistically significant difference in the tendency to appeal in relation to socio-demographic characteristics of the users of tourist services. The study has been conducted on a group sample of 116 respondents from Serbia and 106 respondents from Croatia. The empirical part of the research was carried out by a survey technique. The results of the conducted empirical research confirmed the findings and claims of foreign authors on the impact of tourist service users' age, monthly income and the number of family members, on their inclination to complain. The main findings of this research indicate that maintaining a long-term loyalty relationships with existing users is becoming increasingly difficult. The paper seeks to contribute to the existing literature by indicating the importance of knowing the effects of influence factors upon a tendency to file a complaint. The obtained results help employees in the first service line to identify aspects of influence on the users' reaction to service failure and thus give a starting point for creating measures for effective resolution of user dissatisfaction. Future researchers are advised to collect additional information on the factors influencing the user’s complaint behaviour.


Introduction
As the most important precondition for achieving long-term profitability is, the existence of a loyal users base, it is considered that theythe users, are actually the most valuable asset of a company. The user represents a primary focus of relationship marketing that is oriented towards attracting, developing and maintaining lasting, strong relationships with service organizations (Bove & Robertson, 2005). For this reason, Ganesh, Arnold and Reynolds (2000) emphasize that organizations should continuously strive to increase the satisfaction level of their customers and thus create the conditions for their long-term loyalty. Identifying and appreciating users' needs and wishes creates the basis for long-term loyalty (Marić, Marinković, Marić & Dimitrovski, 2016). Service users play a key role in the service success. Together with the employees they act as resource integrators and in doing so create values (Leroi-Werelds, Streukens, Van Vaerenbergh, & Grönroos, 2017). However, customer satisfaction is not a universal phenomenon and not everyone achieves the same satisfaction level. The reason for this is that users have different needs, goals and previous experiences (Kim & Chen, 2010). Consumer behaviour is the key to measuring their satisfaction level. Satisfaction is the basic user variable -to develop trust in the travel agency.
However, failures are inevitable, especially in service organizations where human participation is high, and production and consumption are simultaneous. The dissatisfaction which is then created, is a complete opposite to satisfaction and occurs when the users' expectations are not met.
Complaints are a natural consequence of any service activity, because "mistakes are an inevitable characteristic of all human endeavours, including service delivery" (Mousavi & Esfidani, 2013, p. 548). Whether the users of tourist services will complain or not depends on a large number of factors. The most often mentioned factors include a degree of dissatisfaction, significance of tourist services in the consumption system, cost analysis, personal characteristics and willingness to find oneself in unpleasant situations when making a complaint.
Variables such as nationality, age, gender, income or education have been identified as tourist behaviour determinants (Kastenholz, Carneiro, & Eusébio, 2005, p. 5). Given that every consumer has a different purchasing behaviour in accordance with their demographic characteristics, they are likewise characterised by different complaint behaviour (Metehan & Yasemin, 2011).
The survival and further growth of the service organization can be secured by maintaining the existing, attracting new and returning the former consumers. In order to succeed companies have to pay attention to managing customer complaints (Kotler & Keller, 2006). The focus of this paper is directed towards the tourist agency customers' inclination to complain in Serbia and Croatia.
Earlier research on service processes focused more on employees. However, in this paper, attention is directed towards understanding the service users' behaviour after the service process and during the service recovery process. The research subject and problem is the appeal behaviour of users of tourist services from Serbia and Croatia, i.e., the differences in the appeal behaviour caused by their different socio-demographic characteristics. Regardless of the fact that Serbia and Croatia are neighbouring countries that have intertwined cultural -historical heritage, the conducted research results have shown that certain factors that affect the appeal behaviour of the tourist services users from these two countries are different. Thus, this paper points out the need to be familiar with information regarding appeal behaviour as a prerequisite for creating an action plan necessary for the implementation of a strategy to improve the competitiveness of a travel agency.

Literature review
Both theoretical and empirical studies demonstrate that in the conditions of global changes it is increasingly difficult to exercise customer care. When tourists' expectations are high, they are more often likely to assess the tourist service quality as less favourable for themselves. In relation to this, tourists' expectations influence the perception of the tour quality negatively, whereas tourist motive influences the perception of the tour quality positively (Sangjae, Sungil, Doyoung, 2011). The customer contact persons' behaviour is focused on creating consumer value, and it determines the level of consumer orientation (Homburg, Wiesere, Likas & Mikolon, 2011). Trust is the key variable for a successful relationship between a travel agency and a consumer (Alvarez, Casielles & Martín, 2010, p. 146). The papers of many authors (Chumpitaz Caceres, & Paparoidamis, 2007;Alvarez et al., 2010) point out that the key trust component is the degree to which the customers deem that the contact person intends to deal with issues that have positive results for them. Ngai, Heung, Wong & Chan (2007) state in their article that complaints are a clear expression of dissatisfaction, but that they also represent an opportunity to reduce dissatisfaction and encourage cooperation. Barlou and Moller (2007) point out that nobody likes to receive complaints. Staff will be affected by complaints made by customers; they will also deem that their work is not appreciated (Bell, Mengüç & Stefani, 2004). They are, however, another external source for attracting and keeping customers, as well as a way to gather information essential for further progress of the organization (Yilmaz, Varnali & Kasnakoglu, 2016). It is impossible to always provide 100% flawless service, despite the importance (Tao, Karande & Arndt, 2016, p. 266). Service failure is inevitable, and it is of key importance for companies to successfully renew their services (Choi & Choi, 2014, 109).
Contact personnel cannot eliminate complaints, but can learn to respond to them efficiently. Some researchers (Maxham & Netemeyer, 2002;Priluck, 2003) point out that the salesperson's response to failures can either strengthen or deteriorate client relationships (Maxham & Netemeyer, 2002). The costs of creating a new client are far higher than the costs of retaining the existing ones, which is an additional indicator of complaint management significance (Assaf, Josiassen, Cvelbar & Woo, 2015, p. 78). Service recovery includes all activities designed to handle a problem, negative attitudes of dissatisfied consumers and a tendency to retain them all eventually. When an organization initiates recovery, the customer's perception of interactional justice should be improved. The customer can then see active effort as an act of courtesy, and demonstration of sincerity and openness as showing understanding and respect (Smith, Bolton, & Wagner, 1999).
This phenomenon is referred to as service recovery paradox (Michel et al., 2007, p. 2).
There is no consensus regarding the intensity interpretation of complaining behaviour. Some authors suggest that complaining behaviour is directly proportional to dissatisfaction level. Other researchers have pointed out that other factors can be significant, such as consumer characteristics, prices of purchased products/services, problem severity, complaint mechanism simplicity, the organization's reputation in terms of complaint response, etc. (Keng & Liv, 1997, p. 89). Consumers from different countries can be characterized by different complaining behaviour, which is probably based on a complex decision-making process. This process comprises expected outcomes, costs and benefits that are included, feeling of guilt and attitude towards the appellant (Yuksel, Klilinc & Yuksel, 2006). Homburg, Fürst and Koschate (2010) point out the significance of sociodemographic and psychological customer characteristics that influence their tendency to express their dissatisfaction in the form of a complaint. There are different studies dealing with the correlations between demographic characteristics and customers' behaviour. Customers with different demographic characteristics would differ in post-purchase behaviour, especially in complaint behaviour (Metehan & Yasemin, 2011, p.44). Occupation, education, household income, age, gender, status and accommodation are used as variables for defining customers' demographic and socio-economic characteristics, and these variables are widely used for differentiating complaint behaviour (Metehan & Yasemin, 2011;Crie, 2003).
As demographic variables are often associated in literature with customers' propensity to complain, i.e. customer complaint behaviour (CCB), numerous authors have studied their impact on consumer complaints related to different types of products and services. Phau and Baird (2008) found that consumers aged 40 and more expect more from products and services they purchased. Respondents in the age category of 54 years and more potentially have more time on their hands than those who are younger and are therefore ready to spend more time and effort on complaints (Phau & Baird, 2008, p. 593). (2010) support claims that a user with higher education is able to successfully evaluate the handling of complaints by the company and its outcome, compared to those with lower education levels. Also, at the end of gender differences research, men and women differ in their purchasing behaviour. For example, in relation to a man, a woman strives to become more cognitively involved in shopping activities and to take more care of customer service issues (Homburg et al., 2010). In the work of Mousavi and Esfidani (2013), it is pointed out that those who complained are younger by age, have better education and higher income.

Homburg, Fürst and Koschate
Individuals with higher income are more frequent buyers of expensive products and were found to have more complaints. They also have more personal resources, which enables them to express their dissatisfaction more easily. There is a significant difference between men's and women's complaints (Tronvoll, 2007, p. 34). To sum up, literature overview shows that some articles proposed that women tend to complain more, whereas others propose that men do.
Relationship marketing will help travel agencies to stay on the market and fight the increasing competition they face. A literature review indicates that an important prerequisite for the application of relationship marketing in tourism is the growing process of customer complaint management. Users participate in organizational processes by providing information about services in terms of their improvement or need for placement of new ones. Significant customer participation as a value cocreator is in the form of feedback as a business resource.

Research methodology
Following the model of other studies, this paper analyses the influence of socio-demographic characteristics on the propensity to complain to travel agencies. The aim of the conducted research is to determine the statistically significant difference in the tendency to appeal in relation to the sociodemographic characteristics of the users of tourist services (gender, age, education level, monthly income, place of residence, number of family members). For this purpose, research was carried out in Serbia and Croatia using the methods of questioning the users of tourist services in the form of a structured analysis using a questionnaire distributed either personally or via the Internet. The questionnaire created for this research consists of two parts. The first part of the questionnaire covered issues related to the sociodemographic user characteristics (gender, age, education level, monthly income, place of residence, number of family members). The second part of the questionnaire referred to the tendency to file an appeal on travel agency business dealings whose tourist services were used in the last three years. The questions in this part of the questionnaire were presented in the form of claims to which the respondents gave a grade from 1 to 5 on the offered Likert response scale (1 -I completely disagree, 5 -I completely agree). The choice of questions was made based on the overview of relevant literature (Homburg, Fürst & Koschate, 2010;Mattila, 2001;Karatepe, 2006). Before designing the final version, the questionnaire was pre-tested on a sample of 15 respondents. The preliminary analysis showed that the questions were logically conceived and were clear to the respondents. Starting from the basic subject as well as the research goals of this paper, and taking into account the previous research on this topic, the main hypothesis that will be the subject of testing with the intention of proving it, is:

H 0 : There is a statistically significant difference in the tendency to lodge a complaint between the users of tourist services of different sociodemographic characteristics.
Bearing in mind that socio-demographic characteristics of the respondents include: gender, age, education level, monthly income, place of residence, and number of family members, the main H0 hypothesis set should be broken down into the following five auxiliary hypotheses: H 1 : There is a statistically significant difference in the tendency to lodge a complaint between the users of tourist services of different sexes H 2 : Among the users of tourist services of different ages, there is a statistically significant difference in the tendency to lodge a complaint H 3 : There is a statistically significant difference in the tendency to lodge a complaint between the users of tourist services of different education levels H 4 : There is a statistically significant difference in the tendency to lodge a complaint between the users of tourist services of different monthly income H 5 : There is a statistically significant difference in the tendency to lodge a complaint between the users of tourist services of a different place of residence H 6 : There is a statistically significant difference in the tendency to lodge a complaint between the users of tourist services of a different number of family members Empirical research was conducted in Serbia and Croatia between September 2016 and January 2017, covering 116 respondents from Serbia and 106 respondents from Croatia. In the sample there are more persons of the female gender in Serbia (62.1%) and in Croatia (80.2%). In Serbia, the highest percentage (59.5%) of respondents aged 18-27, while in the sample in Croatia the highest number is of those who are 68 years or older (37.7%). The sample in Serbia (44.0%) and in Croatia (60.4%) has the highest number of respondents who have completed higher or university education. The results show that the monthly income amount is much lower in Serbia, where the highest percentage (31.9%) of respondents has incomes up to 200 euros, while in Croatia, the highest percentage of respondents receive over 600 euros. This can be related to the respondents age where it has been shown that the highest percentage of those in Serbia is up to 27 years old, while in the sample from Croatia the highest percentage is of those aged 68 and over that signifies that the sample of Serbia is much younger and income increases with age, or increases with years of service. In Serbia and Croatia, the largest percentage of respondents live in cities (more than 86%). In both countries, the largest percentage of respondents number from 3 to 4 family members, while that percent is much higher in Croatia (76.4% in Croatia and 53.4% in Serbia).
All questions and claims from the questionnaires were processed using the statistical software IBM SPSS 21. Empirical research was carried out using descriptive and inferential statistics. Descriptive statistics used percentages and frequencies. Kolmogorov -Smirnov test and Sapiro -Wilkov test were used to check the schedule normality. Since the result distribution of the continuous variables according to sex and place of residence is not normal, Mann-Whitney's U test will be used to test the auxiliary hypothesis. To test the remaining auxiliary hypotheses (age-related, education level, monthly income, family members), Kruskal Wallis Test will be used as a non-parametric technique because the distribution of results is not normal.

Determining the statistically significant difference in the tendency to file a complaint with respect to users of tourist services gender
Determination of a statistically significant difference in the tendency to lodge a complaint with respect to respondents' sex was investigated using the Mann-Whitney U test, as non-parametric alternatives of the t-test of independent samples (Table 1). The amount of significance should be less than the required limit value of 0.05, in order to consider the result of Z approximation as statistically significant. Mann-Whitney U test is used to examine the differences between two independent groups (in this case, respondents' gender: male and female), by comparing the mean value of the rank and the median ( Table 2). The acquired values of the tendency to lodge a complaint are converted into ranks for both groups (male and female) and then calculated whether the ranks of these groups differ significantly. Mann-Whitney U test did not reveal a statistically significant difference in the tendency to complain with men (Md = 3.0, n = 44) and women (Md = 3.0, n = 72) in Serbia, U = 1472.5, Z = -0.644, p = 0.520, r = 0.06 (small effect). Men and women in Serbia are equally inclined to file a complaint.
Mann-Whitney U test, did not find a statistically significant difference in the tendency to complain with men (Md = 3.0, n = 21) and women (Md = 3.0, n = 85) in Croatia, U = 781.5, Z = -0.937, p = 0.349, r = 0.09 (small effect). Men and women in Croatia are equally prone to file a complaint.
Bearing in mind that Mann-Whitney U test, both in Serbia and Croatia, did not disclose a statistically significant difference in the tendency to lodge a complaint according to the sexes, it can be concluded that the assumed H 1 assumption: there is a statistically significant difference in the tendency between the users of tourist services of different sexes to lodge a complaint, is not supported.

Determination of statistically significant differences in the tendency to file a complaint with respect to the users of tourist services age
Determining the statistically significant difference in the tendency to lodge a complaint with respect to the lifespan of the examinees was investigated using the Kruskal Wallis Test, as nonparametric alternatives to the singlefactor analysis of the variance of different groups (tables 3 and 4). The Kruskal Wallis Test serves to compare the results of a continuous variable (a tendency to file a complaint) for three or more groups (respondents' life expectancy). If the test value is statistically significant, the results are converted into ranks, so the mean ranges and median values are compared.  Kruskal Wallis Test did not reveal a statistically significant difference in the tendency to complain according to the respondents' life expectancy in Serbia (Gp1, n = 69: 18-27, Gp2, n = 26: 28-37, Gp3, n = 11: 38 -47, Gp4, n = 8: 48-57, Gp5, n = 2: 58-67), c2 (4, n = 116) = 7.367, p = 0.118. Respondents of different ages in Serbia are equally inclined to lodge a complaint.
Kruskal Wallis Test revealed a statistically significant difference in the tendency to complain according to the life expectancy of respondents in Croatia (Gp1,Gp2,Gp3,Gp4,Gp5,n = 40: 68 and older), c2 (5, n = 106) = 15.632, p = 0.008. The tendency to lodge a complaint is at the highest level   Kruskal Wallis Test did not reveal a statistically significant difference in the tendency to complain according to the respondents' education level in Croatia (Gp1, n = 17: secondary school, Gp2, n = 64: higher school or university, Gp3, n = 25: master, or doctorate), c2 (2, n = 106) = 0.393, p = 0.822. Respondents of different education levels in Croatia are equally inclined to file a complaint.

Determination of statistically significant differences in the tendency to appeal in relation to the education level of users of tourist services
Bearing in mind that Kruskal Wallis Test in the whole sample (Serbia and Croatia) did not reveal a statistically significant difference in the tendency to appeal according to the respondents' education level, it can be concluded that the assumed H 3 assumption: there is a statistically significant difference between the users of tourist services of different education levels in the tendency to file a complaint, is not supported.  Kruskal Wallis Test did not reveal a statistically significant difference in the tendency to appeal according to the respondents' monthly income in Serbia (Gp1, n = 37: up to 200 euros, Gp2, n = 15: 200-300 euros, Gp3, n = 23: 300 -400 euros, Gp4, n = 11: 400-500 euros, Gp5, n = 16: 500-600 euros, Gp6, n = 14: over 600 euros), c2 (5, n = 116) = 6.745, p = 0.240. Respondents with varying amounts of monthly income in Serbia are equally prone to file a complaint.

Determination of statistically significant differences in the tendency to file a complaint with respect to the amount of monthly income of tourist services users
Kruskal Wallis Test did not reveal a statistically significant difference in the tendency to appeal according to the respondents' monthly income in Croatia (Gp1, n = 2: up to 200 euros, Gp2, n = 7: 200-300 euros, Gp3, n = 11: 300 -400 euros, Gp4, n = 12: 400-500 euros, Gp5, n = 19: 500-600 euros, Gp6, n = 55: over 600 euros), c2 (5, n = 106) = 11.605, p = 0.041. The tendency to file a complaint at the highest level among respondents with a monthly income between 500 and 600 euros (Mean Rank = 69.16, Md = 4.0), and the smallest among respondents with a monthly income between 400 and 500 euros (Mean Rank = 34.96, Md = 3.0).
Bearing in mind that Kruskal Wallis Test in Serbia did not reveal a statistically significant difference in the tendency to file complaints according to the respondents' monthly income amount, while in Croatia it can be concluded that the hypothesis H 4 is established: there is a statistically significant difference between the users of tourist services of different monthly incomes in the tendency to lodge a complaint, partial evidence has been provided.

Determination of a statistically significant difference in the tendency to file a complaint with respect to the place of residence of the tourist service user
Mann-Whitney U test is used to examine the differences between two independent groups (in this case, the respondents' place of residence: city and village) by comparing the mean value of the rank and the median (Table  10). The obtained values of the tendency to lodge a complaint are converted into ranks for both groups (city and village) and then calculate whether the ranks of these groups differ significantly.  Mann-Whitney U test did not reveal a statistically significant difference with the respondents from the city tendency to appeal (Md = 3.0, n = 100) and villages (Md = 2.25, n = 16) in Serbia, U = 583.5, Z = -1.760, p = 0.078, r = 0.16 (small influence). Respondents from the city and villages in Serbia are equally inclined to file a complaint.
Mann-Whitney U test, did not detect a statistically significant difference with the respondents from the city tendency to appeal (Md = 3.0, n = 95) and villages (Md = 3.0, n = 11) in Croatia, U = 338.5, Z = -1.478, p = 0.139, r = 0.14 (small influence). Respondents from the city and villages in Croatia are equally inclined to file a complaint.
Bearing in mind that Mann-Whitney U test, both in Serbia and in Croatia, did not disclose a statistically significant difference in the tendency to lodge a complaint against the respondents' place of residence, it can be concluded that the set H 5 assumption: there is a statistically significant difference 4.6. Determination of statistically significant differences in the tendency to file a complaint with respect to the number of family members of tourist service users  Kruskal Wallis Test revealed a statistically significant difference in the tendency to appeal against the respondents' number of family members in Serbia (Gp1, n = 31: 1-2 members, Gp2, n = 62: 3-4 members, Gp3, n = 23: 5 and more members), c2 (2, n = 116) = 6.683, p = 0.035. The tendency to lodge a complaint at the highest level with respondents with 1-2 family members (Mean Rank = 71,37, Md = 4,0), and the smallest among respondents with 3-4 family members (Mean Rank = 52,61, Md = 3.0).
Kruskal Wallis Test did not reveal a statistically significant difference in the tendency to appeal against the respondents' number of family members in Croatia (Gp1, n = 20: 1-2 members, Gp2, n = 81: 3-4 members, Gp3, n = 5: 5 and more members), c2 (2, n = 106) = 0.680, p = 0.712. Respondents with a different number of household members in Croatia are equally prone to file a complaint.
Bearing in mind that Kruskal Wallis Test in Serbia revealed a statistically significant difference in the tendency to complain according to the respondents' number of family members, while in Croatia it did not, it can be concluded that the hypothesis H 6 is established: there is a statistically significant difference between the users of the tourist service of a different number of family members in the tendency to lodge an appeal, partial evidence has been provided.
It can be concluded that the main hypothesis H 0 is set: there is a statistically significant difference in the tendency to lodge an appeal among the users of tourist services of different socio-demographic characteristics, partial evidence has been provided.
As the appellate behaviour is a field of research that deals with the identification and analysis of all aspects involved in the user's reaction to failure, and the main hypothesis is partially accepted, the results of the conducted research indicate the need for further research. After conducting the research, questions remain which could be a driving force for other research which could: beside socio-demographic characteristics of the respondents, also include psychographic characteristics; alongside a method of structured testing apply methods such as focus group research, observation, in-depth interviews; apart from the tourist services users also include employees in tourist agencies. It would also be interesting to compare the results obtained by research conducted in several selected countries.
In further research, it would be worthwhile to collect additional information on the factors influencing the user's complaint behaviour in order to create an action plan as a mechanism for implementing a strategy to improve the competitiveness of a travel agency. In the conditions of increasingly intense competition, efficient complaint management is of key importance for creating a company's positive reputation. The results of the conducted research can be used by travel agencies as guidelines when analysing past oversights, in removing inefficient measures used in managing past consumer complaints and stimulating complaints as an efficient means of maintaining loyalty.

Conclusion
Although the service failure has the potential to destroy consumer loyalty, a successful application of the service recovery strategy can prevent the turn of customers who have a problem with the service. The prerequisite for the implementation of this strategy is to dispose of the user's complaint information.
The results of the conducted research partially confirmed the existence of a statistically significant difference in the tendency to file a complaint between users of different socio-demographic characteristics. The highest incentives for filing a complaint in Serbia are users with 1-2 family members, and the smallest ones with 3-4 members. Respondents from Serbia of different sex, age, education, income and place of living are equally inclined to lodge complaints with travel agencies. The research results showed that among the respondents from Croatia who file complaints the most are those aged 28-37 years old and with monthly income between 500 and 600 euros. The least among the respondents from Croatia are those who are between 18 and 27 years of age and those with monthly income is between 400-500 euros.
Respondents from Croatia of different sex, education levels, place of living and number of household members are equally inclined to lodge complaints with travel agencies.
Partial acceptance of the main hypothesis is a sufficient sign that there is an influence of the individual socio-demographic characteristics of the users on the tendency to file a complaint. In context of the conducted research, the employees in travel agencies are suggested to implement measures and activities aimed at encouraging complaints in order to facilitate the discovery of dissatisfaction and the use of feedback as a resource for improving the service quality. Travel agencies would by emphasizing the importance and paying attention to the users who are complaining, encourage others to report their dissatisfaction. Once they identify the problem, they need to respond quickly to customers' needs, they must be authorized to use all their capabilities in a given situation and have the authority within certain limits to independently solve the client's problem. Under such conditions, it is more likely that consumers' remain loyal to the company and become part of a system based on loyalty together with the employees.