THE INFLUENCE OF ELECTRONIC WORD OF MOUTH (EWOM), BRAND IMAGE, AND PRICE ON RE-PURCHASE INTENTION OF AIRLINE CUSTOMERS

This research aims to investigate the factors which influence the Citilink customer’s repurchase intention. This research implements Electronic Word of Mouth (eWOM), brand image, and price as the independent variables that influence customer’s repurchase intention. This research applies explanatory research, and the data are collected by employing a survey method. The respondents comprised 180 customers who had purchased Citilink Indonesia minimum in one purchased. The research data and hypothesis are analyzed by using Structural Equation Modeling (SEM) based on Partial Least Square (PLS). Finally, the findings indicated that eWOM and brand image significantly and positively influence Citilink customer’s repurchase intention. In conclusion, the more positive the eWOM from the Citilink Indonesia’s previous customer, the higher the repurchase intention. In line with it, the more positive (good) brand image of Citilink Indonesia, the higher the repurchase intention of the customers.

The need for air transportation services in Indonesia continues to increase, and this is indicated by the number of air transport traffic that continues to increase from 2016 to 2018 (databoks.kata data.co.id, 2018). A large number of airlines operating in Indonesia directly creates intense competition. Airlines compete to provide affordable flight services. The lowcost carrier is becoming more popular because the lower middle class can enjoy travel by airplane at an affordable price.
Flight services serve not only domestic customers, but also overseas customers. Tourist arrivals in Indonesia increased by 6.12 percent year-on-year to 1.27 million people in February 2019 from 1.19 million in the same month a year earlier. Bali was the main destination, receiving 437 thousand tourists, followed by Jakarta (196 Thousand) and Batam (158 Thousand). Tourist arrivals in Indonesia averaged 832,831.29 from 2011 until 2018, reaching an all-time high of 1,536,489 in July of 2018 and a record low of 548,821 in January of 2011 (Trading Economics, 2019).
To deal with the tight competition of air transportation business, the internet has a big role in building positive electronic word of mouth (eWOM). Research also has shown that eWOM may have higher empathy, credibility, and relevance to customers than marketer created sources of information on the Web (Bickart and Schindler, 2001). Hung and Li (2007) explain that eWOM can effectively strengthen brand knowledge, leading to a lower customer perceived risk of the product by decreasing the incident of being deceived.
EWOM has significant influences on decreasing the perceived possibility to be deceived among the reviewed articles, and satisfaction seems to dominate the online repurchase intentions studies (Schau, et al., 2009). EWOM is recognized as an effective tool for building brand awareness, creating hype in the marketplace, influencing purchase decisions, and developing brand loyalty. Almana and Mirza (2013) also state that Saudi internet shoppers are very much influenced by eWOM and that a higher percentage of them are dependent on such online forums when making decisions to repurchase products through the Internet.
The amount of demand for a product makes the company keep its brand image. The brand has a uniqueness which distinguishes between one similar product to another. Brand image is one of the important consideration for people to repurchase the product. Towards the brand image, people have their of the product. Brand Image is a determinant affecting customers' subjective perceptions and consequent behaviors (Ryu et al., 2012). Aghekyan-Simonian (2012) mentioned that, the more favorable the brand image, the more positive that attitude toward the branded product and its attributes.
Brand image aids a consumer in recognizing his/her needs and wants regarding the brand and distinguishes the brand from other rivals (Anwar, et al., 2011). According to Aaker (1991), a brand association has a level of the strength, and the link to a brand will be stronger when the lin link is based on consumers experience or the exposures to communications.
Price is the value of a product for the buyer to buy the product and need consideration to determine the price of a product, and the company must know the purpose of the pricing of the products. Worthy price is one of the marketing mixes that influences customer satisfaction and repeat purchase towards online shopping Price is one of the most important attributes evaluated by consumers (Merinda and Budhi, 2016).
Price perception aims to build consumers regarded as a topic of marketing communications and brand management (Erdil, 2015). In research, Suhaily and Soelasih (2017) states that the price has an impact on the intention to repurchase, and consumer satisfaction has a positive impact on consumer buying intention.
This research can provide information about how important all of these phenomena's above and affects to repurchase intention. Based on the background, the author wants to research: The Influence Analysis of Electronic Word of Mouth (Ewom), Brand Image, and Price on Re-Purchase Intention of Aircraft Customers.
The increase in domestic and foreign tourists has an impact on Indonesian economy growth and low-cost carrier airlines, such as Citilink. Through its vision to become the leading low-cost carrier in the region by providing scheduled, low-cost and safety-focused air transport services, Citilink Indonesia is one of the airplanes company which has operated as a low-cost carrier in Indonesia. So far, Citilink Indonesia is only ranked 3rd in the Top Brand of five airlines in Indonesia. This can be a threat to the company in the face of intense competition.

LITERATURE REVIEW Electronic Word of Mouth (eWOM)
WOM communication is a widely acknowledged as a non-commercial and trustworthy source of information that has a massive influence on consumer attitude formation and purchase behaviour. With the rapid growth of the internet, eWOM has emerged as a way for consumers to engage in noncommercial advertising, share and discuss direct experience about the specific product and brand (Chevalier & Mayzlin, 2006).
EWOM is defined as any statements made by future, present or former customers about a product or enterprise, either positive or negative, and is accessible by anyone online (Henning, et al., 2004). Goldsmith (2008) defined eWOM as every internetbased communication about the usage or characteristics of something (products, services, or a company). This definition is adopted, referring to eWOM as all informal communications for Traveloka consumers through the internet related to the usage or characteristics of booking accommodation.
The arrival and expansion of the internet have extended consumer 's comments posted on the internet and has provided consumer's opportunities to offer their consumption related advises by engaging in electronic word-of-mouth (eWOM). Henning, et al., (2004) explained eWOM by using six dimensions. The dimension is consist of a positive self-enhancement, social benefits, helping the company, advice seeking, concern for another consumer, and venting negative feelings. H1:Electronic Word of Mouth has a positive influence on the customers' repurchasing intention towards Citilink. Peter & Olson (2000) defined brand image is consumer perception and preference for the brand, as reflected by various brand associations that exist in consumer memory. Although brand associations can occur in various forms, they can be differentiated into performance associations and imagery, associations that are related to brand attributes and advantages.

Brand Image
Brand image is a crucial competitive advantage that helps in creating value through differentiating the brand, forming purchasing rationales, constructing sense and feeling, and a significant value for organizations (Aaker,1996;Keller, 2009). Since the fundamental purpose of a brand is to provoke confidence, feeling of trust, strength, durability, security, and exclusivity (Aaker, 1996;Keller, 1993), thus it can be considered an important means of decreasing uncertainty and providing useful information that can help in directing consumer decisionmaking processes (Erdem et al., 2002). Researchers indicated the presence of a significant relationship between brand image and purchase intention (Shukla, 2010;Wu et al., 2011;Lien et al., 2015). H2: Brand Image has a positive influence on the customers' repurchasing intention towards Citilink Price Kotler and Armstrong (2016) stated that the price is the amount of money charged for a product or service or the sum of the values that consumer exchanged for the benefit of having or using a product or service. Therefore, the following indicators for this construct based on Stanton (1994) are: 1. Price affordability 2. Price suitability with the product quality 3. Price competitiveness 4. Price suitability with product benefit.
Price is the amount of money charged for a product or service, or the sum of the values that consumer exchanged for the benefit of having or using a product or service (Kotler and Armstrong, 2016). Variable pricing in the marketing mix is expected on what should consumers have to purchase an item or service usually uses the value of money. According to Tjiptono (2008) in the context of service marketing, put the price can be defined as the amount of money (monetary unit) or another aspect (non-monetary) containing the utility or usefulness in particular that need to get service. H3: Price has a positive influence on the customers' repurchasing intention towards Citilink.

Repurchase Intention
Repurchase intention is to regard as the primary construct to explain customers repurchase behaviors. Yan & Yu (2013) explain that repurchase intention is the subjective probability that a person will buy products or services continuously from the e-supplier or shop in the future. Jiang and Rosenbllom (2005) also explain that customers repurchase behavior or intentions is beneficial to online business.
The intention of buying (repurchase intention) is a commitment to the consumers who formed after consumers make purchases a product or service. This commitment was arising from a positive impression of consumers to a brand, and consum-ers satisfying against such purchase (Hicks et al., 2005). Several authors defined loyalty as consumers' commitment to a particular organization, but Oliver's point of view is one that is widely accepted.
So, loyalty is deeply held commitment to rebuy or re-patronize a preferred product/service consistently in the future, despite situational influences and marketing efforts having the potential to cause switching behavior (Oliver, 1997). Butcher, (2005:15) argued that consumer intentions to buy back is one measure of the success of an enterprise, especially service companies.
In summary, the research model and hypothesis are shown in figure 1.

METHOD
The research method used in this research is a quantitative method. Quantitative method is known as a method of positivistic because it is based on the philosophy of positivism. This method is also called the method of discovery because, with this method, a variety of new science and technology can be found and developed. Moreover, this method is called as quantitative research because the data are in the form of figures and statistical analysis (Sugiyono, 2014).
The research design that the researcher used in this research is explanatory research. According to Malhotra (2010), explanatory research is one kind of research that has the main goal to present insight and understanding of the situation of the problems faced by researchers. The reason for choosing this type of research is to test and analyze the effect of eWord of Mouth, price, brand image, and repurchase intention variables.
The population is a generalization region consisting of objects or subjects that have certain characteristics that are applied by the researcher to learn and then make a conclusion based on research. Based on understanding above, the researcher takes a conclusion that populations are object or subject wherein a certain region and fulfill specific terms and condition that has related issues with the researcher's problems. The population that the researcher used is a citizen who already uses Citilink Indonesia in Malang City.
The sample is a part of the population and characteristic that consist of the total population itself. Sampling technique in this research used non-probability approach. The meaning of non-probability samples is the technique of taking a sample that does not give opportunities/chance to each element or every population for becoming a sample. This research used the purposive technique. Purposive sampling is used to get an easy understanding of sample criteria that relevant to the researcher's purpose, where it is expected that the selected sample has accurate information for the researcher. The criteria in sampling's respondent in this research are: 1. The registered ages user either men or female who can book Citilink Indonesia by their own identities. 2. Have already use Citilink Indonesia at least once.
The determination of total sample according on Roscoe (1975) that quoted by Uma Sekaran who provide guidance for determine the amount of sample, sample size for every researcher in ranged of 30 -500, if the factors used in the study are many, the sample size is at least 10 times more than the number of factors. In this research, the total of indicator variables is 17x10, so the standard of the minimum respondent is 170 respondents and already fulfill a standard minimum that is appointed. Data were collected by using primary data methods. For primary data, a close-ended questionnaire as an instrument was administered by the researcher to obtain data from respondents. The questionnaire is divided into two, where the first part consisted of the background information, and the second part comprised of items that measure the influence of eWOM, Brand Image, and Price toward purchasing Intention. The data collection method is using an online questionnaire method. The researcher uploads the questionnaire in the google docs to get more accurate data. The questionnaire will be completed by the respondent of Citilink Indonesia user in Malang.
This study adopts structural equation modeling (SEM) for data analysis. The hypotheses that have been formulated in this research are tested by using Partial Least Squares (PLS) Version 2.0 as a multivariate statistical technique that makes comparisons between multiple dependent variables and multiple independent variables. PLS can simultaneously test the measurement model as well as the structural model.

RESULT AND DISCUSSION
The respondents in this research include the citizen who already flew with Citilink Indonesia in Malang at least once and booked the ticket using their own identity. It has previously described that this research utilizes survey method by distributing questionnaires to the respondents. The researcher conducts the data collection for one week by directly distributing online questionnaires by applying Google Forms. The researcher distributed the Google Forms by sending the links to close friends and family who lived in Malang and asked them to share to their network.
The number of questionnaires distributed online is 200 and received as many as 200. After checking, 20 questionnaires are invalid for research data, because: 1. The respondents do not fill questions or statements 2. The respondents never use Citilink 3. The respondents are not living in Malang 4. Inconsistent answers given to the question result in a biased answer Thus, The level of respondent's rate in this research is 90% as the total questionnaires that can be processed as a sample of this research is 180.
The table below shows that the number of respondents (n) in this research is 180. The minimum and maximum values describe the statement in response to the item in the questionnaires. The minimum value of 1 indicates a minimum of respondents who are giving value on each statement on four variables such as EWOM, brand image, price, and repurchase intention. Meanwhile, five indicates that respondents provide maximum value of 5 for each statement on all variables.
The mean value in Table 2 is used to determine the average opinion given by respondents on each item statement for each variable. The available data show that the mean value for each variable is greater than 3.00. It shows that the average respondents agree to the overall statement in each variable in this research. Standard deviation value indicates a measure of deviation between responses to the questions in the questionnaire. A small value indicates that the data used will cluster around the average value.
If the amount of standard deviation does not exceed the average value, then there are no outliers (Sujianto, 2009). Based on Table 2, the standard deviation has no value that exceeds the value of the mean or the average on each construct. Thus, it can be concluded that there are no outliers or the distant difference between the response and the question among respondents.
Analysis of the evaluation model in this research applies Partial Least Square (PLS) 2.0 to estimate parameters and to predict the relationship of causality. Evaluation of the model is performed with three stages, including the testing of convergent validity, testing of discriminant validity, and testing of reliability.

Convergent Validity
Assessment in convergent validity testing is conducted based on the value of Average Variance Extracted (AVE), commonality, and the value of factor loading. Rule of thumb for both parameter AVE and commonality is more than (>) 0.50, and more than (>) 0.70 for the value of factor loading. Additionally, indicators with a loading factor value between 0.5-0.7 should not be removed as long as the AVE and communality values are still above 0.5 (Abdillah and Hartono, 2015). Hair et al. (2006) in Abdillah and Hartono (2015) stated that the rule of thumb is typically applied to make initial examination of the matrix factor, where ± 0.30 is considered as having met the minimum level, loading ± 0.40 is considered better, and loading more than (>) 0.50 is considered significantly practical. Besides the AVE and communality parameters in Table 3, convergent validity can be determined based on the following loading factors as presented in Table 4.

EWOM: Electronic Word of Mouth; BI: Brand Image; P: Price; RI: Repurchase Intention.
Based on Table 3 above, it can be seen that the value of AVE and Communality in each construct is more than 0.5. Similarly, the outer loading test results in Table 4 shows that all indicators value is above 0.7. Thus, based on the processing results, it can be concluded that the convergent validity has been fulfilled.

Discriminant Validity
After assessing convergent validity, the next step is to measure the discriminant validity. Discriminant validity of the assessment is based on the value of the Cross Loading in Table 5, which is more than 0.7 in one variable or construct. Based on the table, it is concluded that the discriminant validity is met for each indicator in each variable reaching over 0.7. Despite the same conditions as the previous loading factor assessment, if it is a value of lower than 0.7, it is still considered valid because they have other parameters with a value of more than 0.5.

Reliability Testing
After construct validity is accomplished, and valid data are obtained, reliability takes place for further testing. Reliability test can be performed by using two methods of Cronbach's Alpha value, whose value must be more than (>) 0.6, and Composite Reliability value, whose value should be more than (>) 0.7. According to the algorithm table above, all variables have the value of Cronbach's Alpha, which is more than (>) 0.6 and Composite Reliability of more than (>) 0.7. Hence, the data and the results of the measurements are considered reliable. Based on the test results of convergent validity, discriminant validity, and reliability testing can be concluded by using algorithm models in Figure 2 below.  Hypothesis 1 states that Electronic Word of Mouth has a positive influence on the customers' repurchasing intention towards Citilink. From Table  6, it can be seen that the value of T-Statistics of Electronic Word of Mouth is 2.783870, which is greater than the value of T-table (1.64). It indicates that Electronic Word of Mouth has a positive influence on the customers' repurchase intention towards Citilink Indonesia. Based on the results, it is determined that Hypothesis 1 is supported.
Hypothesis 2 states that Brand Image has a positive influence on the customers' repurchasing intention towards Citilink. From Table 6, it can be seen that the value of T-Statistics of brand image is 8.227170, which is greater than the value of T-table (1.64). It indicates that brand image has a positive influence on the customers' repurchase intention towards Citilink Indonesia. Based on the results, it is determined that Hypothesis 2 is supported.
Hypothesis 3 states that price has a positive influence on the customers' repurchasing intention towards Citilink. From Table 6, it can be seen that the value of T-Statistics of the price is 1.555562, which is lower than the value of T-table (1.64). It indicates that price has no significant influence on the customers' repurchase intention towards Citilink Indonesia. Based on the results, it is determined that Hypothesis 3 is not supported.
The summary of hypotheses testing results is illustrated in the following Table. Based on the hypothesis testing results above, it is indicated that EWOM and brand image positively influence Citilink customers' repurchase intention. However, the results also demonstrate that Price construct has a negative influence on Citilink customers' repurchase intention. Based on the results, the researcher conducts a validity finding by seeking an explanation for the results of the already tested hypotheses. The validity of the discoveries is analyzed by using journals and research models to support the statement that has previously been disclosed.
After a test of convergent validity, discriminant validity, and reliability testing, the next step is to perform the hypothesis testing. Based on the data processing, the form of total influences is illustrated in the following Table 6.
In one-tailed hypothesis testing, if the coefficient path shown by the T-Statistic is more than or equal (  ) 1.64, then the alternative hypothesis can be stated as supported. Nevertheless, if the statistical value of T-Statistic is less than or equal (  ) 1.64, then the alternative hypothesis is stated as not supported.  Any positive or negative statement made by potential, actual, or former customers about product or company which is made available to multiple of the people and institutes via the Internet (Hennig-Thurau et al., 2014). This construct is related to the opinion or testimonial from customers for experiencing product or services. However, if there are so many good electronic words of mouth, the more the customers' intention to repurchase it.
Hypothesis 1 in this research states that electronic word of mouth has a positive influence on the Citilink customers' repurchase intention. Mostly, the respondents of this research are 18-20 years old. In this range of age, people are already supported with a gadget to encourage their productivity. Together with it, people can always gather information about a product before repurchasing it. In line with it, the woman respondent is greater than then man respondents, which means that a woman has much consideration before repurchasing a product. Automatically, good eWOM can lead to the higher customer intention of a product. The results show that the value of its t-Statistics is 2.783870, which are greater than the value of t-table (1.64). Based on the result, it is concluded that electronic word of mouth has a positive influence on the Citilink customers' repurchase intention. Thus, hypothesis 1 is supported. This result is consistent with the researcher conducted by Chevalier and Mayzlin (2006).
In conclusion, the more positive the electronic word of mouth, the more the Citilink customers' repurchase intention. However, the more negative the electronic word of mouth, the less Citilink customers' repurchase intention. Thus, EWOM becomes one of the determinants of Citilink customers' repurchase intention.
Brand image is a representation of the overall perception of the brand and is formed from past information and experience of the brand. Customers typically rely on a small number of trusted product/service providers that consistently deliver high quality precuts or services (Cousins and Menguc, 2006). The underlying goal is for the brand is to evoke feelings, trust, confidence, security, strength, durability, speed, status, and exclusivity (Aaker, 1996;Keller, 1993). Ultimately, a strong product/service brand conveys the core value proposition of both the organization and its product/service in a way that resonates with customers.
Hypothesis 2 in this research states that brand image has a positive influence on the Citilink customers' repurchase intention. This caused by mostly, the respondents of this research are women, and women have many considerations more than men. Automatically, brand image is being one of the important things to be considered. The better the brand image, the more it fulfills the consideration of a woman. The results show that the value of its t-Statistics is 8.227170, which are greater than the value of t-table (1.64). Based on the result, it is concluded that brand image has a positive influence on the Citilink customers' repurchase intention. Thus, hypothesis 3 is supported. This result is consistent with researches conducted by Shukla (2010), Wang and Yang (2010), and Wu et al. (2011). Wang and Yang (2010) investigated the impact of brand credibility on consumers' brand purchase intention focusing on China's automobile industry. They proposed that brand image play a positive role in this research. Wu et al. (2011) also investigated the direct influences of store image and service quality on brand image and purchase intention for a private label brand. Their study revealed that brand image has a direct and positive influence on purchase intention.
In conclusion, the better and strengthen the brand image, the more the Citilink customers' repurchase intention. In contrast, the worse and weak the brand image, the less the Citilink customers' repurchase intention. Thus, the brand image becomes one of the determinants of Citilink customers' repurchase intention.
Customers are often willing to pay a premium for customized products because their needs are better met. Many consumers use price as a qualitysignaling to reflect in the conventional wisdom that "you get what you pay for." If a company facilitated the creation of a co-production process to make a product tailored to the customer's needs and the product adds great value, price becomes a less important factor (Wind and Rangaswanny, 2001). In this research, the price reflects the quality of the product, so it can increase the intention of Citilink customers' to repurchase it.
Hypothesis 3 in this research states that price has a positive influence on Citilink customers' repurchase intention. Mostly, the respondents of this research have ever flown with Citilink once. It means that they had already experienced Citilink Indonesia and can rate the quality of the Citilink itself with the price offered by Citilink Indonesia. The result shows that the value of its t-Statistics is 1.555562, which is lower than the value of t-table (1.64). Based on the result, it can be concluded that the price does not have a significant influence on Citilink customers' repurchase intention. Thus, hypothesis 3 is not supported.
In conclusion, as long as the price is equal with the good service provided by Citilink Indonesia, it will not be a problem that matter. Because, in the airline business, the customer will give no regret to pay the high price for their safety.

CONCLUSIONS AND RECOMENDATIONS
This research aims to investigate the influence of EWOM, brand image, and price towards Citilink customers' repurchase intention. Thus, it can be concluded that: 1. The findings show that Electronic Word of Mouth (EWOM) on the Citilink's customer repurchase intention is positive and significant. This result indicates that how great of a potential impact EWOM can have on the consumer repurchase intention. In this research, EWOM successfully influences how people see Citilink Indonesia using an experimental study of consumers' use of online recommendation sources. The various platforms which facilitate EWOM, such as social media, discussion boards, and other communication tools are also increasingly being recognized for how they influence the adoption and use of Citilink Indonesia. At this time, EWOM message will be an important reference for consumers to obtain information about Citilink Indonesia and also to establish their repurchase intention. 2. The brand image also gives a positive and significant influence on Citilink Indonesia custom-ers' repurchase intention. In this case, the brand image successfully evokes feelings of trust, confidence, security, strength, and exclusivity. It happens because the strong brand image of Citilink Indonesia conveys the core value proposition of both the organization and its service in a way that resonates with customers. 3. Price has no significant influence on Citilink Indonesia customers' repurchase intention. This condition happens because neither the Citilink price is suddenly getting an increase or decrease, as long as they can give a good value for their customers, the customers will still buy or repurchase Citilink Indonesia whatever happens.