The Influence of Marketing Mix, Perceived Risk, and Satisfaction on Word of Mouth in Xyz Clinic

The increasing need for health services, peoples who lived in the Pekayon, Bekasi City were given the opportunity to choose the right clinic. Word of mouth is a marketing technique that can be used by clinics. This study aims to analyze the effects of the marketing mix, perceived risk, and satisfaction on word of mouth at XYZ clinic. The research is a descriptive method with a survey using questionnaires and 200 respondents as the sample. Furthermore, the data analysis technique is descriptive with SPSS16.0 software and Structural Equation Model (SEM) with LISREL 8.70. Based on the results, it can be concluded that the marketing mix has a positive effect on perceived risk, marketing mix has a positive effect on satisfaction, perceived risk has a negative effect on satisfaction, marketing mix has a positive effect on word of mouth, perceived risk has a negative effect on word of mouth, and satisfaction has a positive effect on word of mouth. Referring to these conclusions, it can be confirmed that the clinical management of doctor XYZ needs to improve employee services, convenience the patient that this clinic has expert doctors, and utilizing the use of social media as a marketing strategy.


Introduction
In daily basis activities, health is the primary issue but, changes in unhealthy lifestyles particularly in urban areas have a weight on health. The clinic is one of the health infrastructures that is facilely found in sundry regions, including in densely populated housing. When someone goes to a doctor or clinic, they cannot check the quality of each clinic or doctor because they do not have the time or resources, thus someone admit recommendations from their family, friends, or closest people.
One of the marketing strategies is word of mouth. Word of mouth plays an important role for it is essential for brands, organizations, or businesses that have a small budget in marketing their products or services because they require minimal costs compared to use advertising media. People who did word of mouth usually derived by the satisfaction either facilities or services. Thus, satisfaction plays a huge role in word of mouth marketing.
According to Kotler and Keller (2012), high satisfaction makes someone voluntarily spread word of mouth in the form of good news about the company. In addition to satisfaction, risk perception is something that needs to be considered in word of mouth. According to Martin (2017) risk perception theory adds an important aspect that helps to understand information derived from word of mouth and its effects. Also, from the aspect of service providers, the marketing mix needs to be considered in building positive word of mouth. This is important because the marketing mix framework is widely used by marketers as the foundation of marketing planning by marketing practitioners as a universal marketing paradigm (Cengiz and Yayla, 2007).
Based on prior explanation, previous studies and theories had explained word of mouth, satisfaction, risk perception, and marketing mix, yet none of the researchers nowadays has combined these variables for study in the field of health services, mainly clinics. One of the clinics located in Pekayon, Kota Bekasi is the XYZ Clinic which was established in 2004 withstand until now and developing with plans to create inpatient care facilities. One the predictor has been thought what makes XYZ clinic able to survive and continue to grow is through word of mouth marketing techniques to patients by favoring quality services. Although using word of mouth techniques, the average number of patients treated at this clinic only 60-80 patients per day, while the target is 100 patients per day (Figure 1). There are fluctuations of visitor every year and tends to decrease. The most significant decrease was September to November of 2017. This study aims to analyze the effect of the marketing mix, risk perception, and satisfaction with word of mouth in the XYZ clinic and formulate managerial implications that can be applied to the health industry, especially clinics. Various theories related to the variables studied include; marketing mix has a negative effect on risk perception (Yoon, 2010), the marketing mix has a significant effect on satisfaction (Alipour and Darbahaniha, 2018), negative relationship between risk perception and customer satisfaction (Ghotbabadi et al., 2016) , product, price, promotion, personnel, process, and physical evidence affect word of mouth (Syriac, 2013), there is a negative link between risk perception and word of mouth (Tho et al., 2017), satisfaction significantly and positively affects word of mouth (Hsu, 2018;Simanjuntak & Hamimi, 2019). The conceptual framework of research can be seen in Figure 2.

Figure 2. Conceptual Framework
Risk Perception Satisfaction 1. H1: Marketing mix has a significant effect on risk perception 2. H2: Marketing mix has a significant effect on satisfaction 3. H3: Risk perception has a significant effect on satisfaction 4. H4: Marketing mix has a significant effect on word of mouth 5. H5: Risk perception has a significant effect on word of mouth 6. H6: Satisfaction has a significant effect on word of mouth

Methods
The research was located in XYZ clinic in Bekasi City, Indonesia. The study was conducted for three months starting from July 2018 to September 2018. Using a quantitative approach with survey methods, and involving XYZ clinic patient as the respondents. This study used convenience sampling techniques or often called accidental sampling with the criteria of respondents are patients at the XYZ clinic general polyclinic who have treated at least once. The technique is advantageous as it takes relatively little time and cheaper costs (Sumarwan, 2014).
The variables of this study are word of mouth, satisfaction, risk perception, and marketing mix. In this study, the definition of word of mouth variable is sharing information through face to face or social media consisting of five indicator variables. The second variable is satisfaction which is interpreted as the patient's assessment of the quality of service consists of three indicator variables. The third variable is the perception of risk, a concern that can arise when having a treatment at the XYZ clinic, consists of five dimensions (financial, performance, psychological, physical, time), each dimension has two indicators. The last variable is the marketing mix as a value offered by the XYZ clinic, consisting of seven dimensions, namely; product, price, place, promotion, process, each dimension consists of two indicators, while the physical and personnel dimensions each have three indicators. According to Ferdinand (2005), the sample size is 5 to 10 times of the number indicator variables, so the number of samples is between 34 x 5 = 170 people up to 34 x 10 = 340 people. This study used a total sample of 200 people. The data collected in this study were primary data obtained from the main source, by providing self-administered questionnaires. The data then analyzed by descriptive statistical methods, and Structural Equation Modeling (SEM). Descriptive statistics are used to see patient demographic and behavioral characteristics (service, clinic, and social media). SEM analysis is used to analyze the influence between the variables of the marketing mix, perceived risk, and satisfaction of word of mouth.

Respondents' Characteristics
The sample of this study was the XYZ Clinic patients in Bekasi City. Questions about demographics were given, namely gender, age, education, job, and monthly household expenses. Distribution of respondents by sex are respondents with female sex (51.5%) more than male respondents (48.5%). Then, the age category of 25-35 years is the highest number of respondents with 67 people or 33.5 percent. Characteristics of respondents based on education were dominated by high school level with a total of 90 people or 45.0 percent, jobs that had the highest number of respondents were private employees with 81 people or 40.5 percent, and expenditure categories of household respondents per month with the highest number of 118 respondents or 59.0 percent for the IDR 1.500.000-IDR 5.000.000.

Social Media Use Behavior
Based on the study, it was found that as many as 179 people or 89.5 percent of respondents were active users in social media for the past 3 months, and 21 people or 10.5 percent of respondents were not active in using social media. This illustrates that social media can be a communication tool used by the majority of respondents. Based on the results of the questionnaire that was distributed multi-response, the social media that respondents often use is Whatsapp, which is 163 people or 81.5 percent, second place is Facebook as many as 64 people or 32.0 percent, followed by Instagram which is 57 people or 28.5 percent. This finding in line with Simanjuntak and Musyifah (2016) that social media accounts owned by consumers most widely are Line (98%), Facebook (95%), Twitter (93%), BBM or Blackberry Messenger (93%), Whatsapp (90%), and Instagram (90%).

Overall Model Fit
The model fit test can be seen from the complete Root Mean Square Residual/RMR, Root Mean Square Error of Approximation/RMSEA, Goodness of Fit Index/GFI, Adjusted Goodness of Fit Index/AGFI, Normed Fit Index / NFI, Comparative Fit Index / CFI and Relative Fit Index/RFI in Table 2. The overall value of the aspects of the conformity criteria has been included in the Good Fit category so that the overall research model is declared feasible. The results obtained after re-specification, indicate that the overall model fit is included in the Good Fit category.

Measurement Model Fit
The suitability test of the measurement model was carried out by validity and reliability ( Table 3). The results of the validity test show that not all indicators have a t-value of more than 1.96 and the coefficient > 0.5. In the marketing mix variable, the dimensions of personnel and physical evidence each have one invalid indicator. As for risk perception variables, the financial, performance, and physical dimensions each have one invalid indicator, and for word of mouth variables, only one indicator is invalid.   Table 3, marketing mix, risk perception, satisfaction, and word of mouth meet reliable requirements with Construct Reliability (CR)> 0.7. That is, all the variables and indicators have represented well the latent constructs developed. This indicates that each indicator variable is valid for measuring its latent construct. Based on the table above, the value of the Variance Extracted (VE) of the marketing mix, the perceptions of income and word of mouth ≥ 0.50. The Variance Extracted (VE) value of satisfaction shows <0.5 but, still has construct reliability (CR)> 0.6, then the Convergent Validity of the latent construct can still be said to be sufficient.

Structural Model Fit
The next step that must be done is to test the suitability of the structural model fit. In the initial testing phase, the MMPE3, MMPH3, PRFN1, PRPF1, PRFI2 and Y5 indicators included indicators that did not have good validity so that these variables were not included in the next processing phase, namely the validity test. Figure 3 shows SEM measurement model. By using LISREL, the word of mouth variable has a value of R² = 0.56 which means that the percentage effect of satisfaction, risk perception and marketing mix variables on word of mouth variables is 56.0 percent or can be interpreted as variations in satisfaction, risk perception, and marketing mix used in the model is able to explain 56.0 percent of the variation of word of mouth variables, while the remaining 44.0 percent is influenced or explained by other variables not included in the research model such as service value variables in Hsu's research (2018), or loyalty such as Cengiz and Yayla's research (2007).

Statistical Hypothesis Test
In testing the hypothesis, if the path coefficient value > 0.05 with the value of t-value > 1.96 then the influence of certain variables is included in the significant category, but if the value of the path coefficient < 0.05 with t-value < 1.96 then the influence between variables included in the non-significant category (Table 4).  Table 4 explains that all influences between variables have a significant effect, such as the positive influence of marketing mix on risk perception, marketing mix has a positive effect on satisfaction, risk perceptions negatively affect satisfaction, marketing mix has a positive effect on word of mouth, risk perception negatively influences word of mouth, and satisfaction has a positive effect on word of mouth.
In addition to direct influence, this study also has an indirect influence. Indirect effects on this study include marketing mix (MM) effect on word of mouth (Y) through risk perception (PR), marketing mix (MM) effect on word of mouth (Y) through satisfaction (ST), marketing mix (MM) effect on satisfaction (ST) through perceptions of risk (PR) and risk perception (PR) influence on word of mouth (Y) through satisfaction (ST). Table 5 explains the indirect effects. The indirect influence of the marketing mix on word of mouth through perceived risk is -0.13 while the direct effect of the marketing mix on word of mouth is 0.59. The direct effect of the marketing mix on word of mouth has a greater influence. In addition, the indirect effect makes the path coefficient negative. The amount of indirect influence of the marketing mix on word of mouth through satisfaction is 0.26 while the direct effect of the marketing mix on word of mouth is 0.59. Although the direction of both is the same, the direct influence has a higher path coefficient so that the influence is more significant when compared to the indirect influence of the marketing mix on word of mouth. The magnitude of the indirect effect of risk perception on word of mouth through satisfaction is 0.07, while the direct influence of risk perceptions on word of mouth is -0.38, thus the direct influence of risk perceptions on word of mouth has a greater influence. In addition, indirect influence makes the path coefficient positive. The amount of indirect influence from the marketing mix on satisfaction through risk perception is -0.07 while the direct effect of the marketing mix on satisfaction is 0.80, therefore the direct influence of marketing mix on satisfaction has a greater influence. Besides, indirect effects make the path coefficient negative.

Effects of Marketing Mix on Risk Perception
The first hypothesis, namely the marketing mix has a significant effect on risk perception. This can be seen from the marketing mix path coefficient on the perception of risk of 0.34 with t-count of 4.75. So that it can explain that the marketing mix variable has a positive and significant effect on risk perception thus it accepts the first hypothesis. The positive effect of the marketing mix on risk perception, in a previous study conducted by Yoon (2010) found a negative influence between marketing mix and risk perception.

Effects of the Marketing mix on Satisfaction
The second hypothesis, which states that the marketing mix has a significant effect on satisfaction. The link in between variables can be seen from the path coefficient of the marketing mix to satisfaction of 0.80 with t-value of 10.67. This can explain that the marketing mix variable has a positive and significant effect on satisfaction so that it accepts the second hypothesis. This result is supported by previous research conducted by Alipour and Darbahaniha (2018) which shows that the 7P marketing mix element consisting of product, price, place, promotion, personnel, process, and physical evidence has a significant influence on satisfaction.

Effects of Risk Perception on Satisfaction
The third hypothesis, states that perceptions of risk have a significant effect on satisfaction. This can be seen from the perception of risk to satisfaction path coefficient of -0.21 with t-value of 3.03. So, it can explain that the risk perception variable has a negative and significant effect on satisfaction so that it accepts the third hypothesis. Risk perception negatively influences satisfaction. This result is supported by previous research conducted by Ghotbabadi et al. (2016); Yue Chen et al. (2015); Cheng and Lee (2011).

Effects of the Marketing mix on Word of Mouth
The examination of the fourth hypothesis shows the marketing mix has a positive and significant effect on word of mouth. This can be seen from the marketing mix path coefficient on word of mouth at 0.59 with t-value of 6.72. This can explain that the marketing mix variable has a positive and significant effect on word of mouth so that it accepts the fourth hypothesis which states the marketing mix has a significant effect on word of mouth. This result is supported by previous research conducted by Suryani (2013) which explains that product, price, promotion, personnel, process, and physical evidence affect word of mouth; Cengiz and Yayla (2007) show that the components of the marketing mix, namely product, price, promotion, and place.

Effects of Risk Perception on Word of mouth
The results of examining the fifth hypothesis show that perception of risk negatively and significantly affects word of mouth. This can be seen from the path coefficient of perception of risk for word of mouth at -0.38 with t-value of 5.45. This can explain that the risk perception variable has a negative and significant effect on word of mouth so that it accepts the fifth hypothesis which states risk perceptions have a significant effect on word of mouth. Previous research by Tho et al. (2017) found that there was a negative relationship between perceived risk and word of mouth.

Effects of Satisfaction on Word of Mouth
The test of the sixth hypothesis shows satisfaction has a positive and significant effect on word of mouth. This can be seen from the path coefficient of satisfaction with word of mouth at 0.32 with t-value of 4.63 meaning that it accepts the hypothesis. Prior research whose supporting this result are Hsu (2018); Kitapci et al. (2014); Chaniotakis and Lymperopoulos (2009); Cengiz and Yayla (2007).

Managerial Implications
Actions that should be taken by the XYZ clinic are the promotion of consultation (creating events to regarding diseases that are often found in the community), expanding parking lots, mind the hospitality of employees (enforcing a standard smile, greetings, greeting). Meanwhile, to minimize the risk/concern of patients by increasing physician care qualifications, increasing the number of employees. Besides that, it is necessary to do a family approach because the source of information trusted by patients is from the closest people like family/friends/neighbors. This can be improved by utilizing social media accounts that are quite popular based on this research such as Whatsapp, Facebook, Instagram.

Conclusion
Marketing mix, perceived risk, and satisfaction have an influence on word of mouth. This can be seen from significant research results such as; marketing mix has a positive effect on risk perception, marketing mix has a positive effect on satisfaction, risk perception has a negative effect on satisfaction, marketing mix has a positive effect on word of mouth, risk perception has a negative effect on word of mouth, and satisfaction has a positive effect on word of mouth hypothesis accepted. Referring to these conclusions, it can be implied that XYZ clinic management needs to improve employee services, assure patients that the clinic has expert doctors including duty doctors on duty, and utilize the use of mainstream social media such as Whatsapp, Facebook, and Instagram to share information.

Recommendation
This study only discusses the general polyclinic in the XYZ clinic, not including specialist medical services or larger scale such as hospitals. The questionnaire used is a standard questionnaire, but the questionnaire was independently developed by researchers based on research theory and concepts so that there are several indicators that are invalid and then omitted in SEM calculations. In addition, a compilation of data retrieval is also very dependent on this study, the study respondents were patients who were sick and were waiting for doctor's services at the XYZ clinic or the patient's family to take respondents who were waiting, busy and less focused in filling out the research. This factor makes it difficult for researchers to get good answers.
Suggestion for further research is to conduct detailed research on the relationship of the marketing mix in the services/services (product, price, place, promotion, personnel, process, and physical evidence) with risk perceptions especially in the field of health. In this study, the marketing mix has a positive effect on risk perception. This is beyond the prediction that the marketing mix has a negative effect on risk perception. Further research should add indicator variables from the financial, performance, psychological, physical, time-varying dimensions related to perceived risk and add the social dimension of the risk perception variable. This is because respondents prefer to answer neutral rather than giving a good or bad assessment, if further research adds variations in variable indicators it is expected to reduce the answers that are still biased.