1. Introduction
Since 21st century, online distribution channels have been increasingly utilized as a competitive marketing tool for facilitating consumers’ purchasing behavior in decisionmaking processes (Karray & Sigué, 2018; Singh, 2014). This phenomenon come from novel functions of online distribution channels compared to traditional distribution channels such as wholesalers and retailers. The effectiveness of using an online distribution channel encompasses the easy access to mobile usage platforms (i.e., multi-platforms) and metasearch engines in diverse businesses (Lee, Kwag, & Potluri, 2015; Verhoef, Kannan, & Inman, 2015). In addition, online distribution channels enable companies to distribute and sell their products and/or service through multiple channels rather than a single channel (Neslin et al., 2006; Neslin & Shanker, 2009). The significant value of online distribution channels lies in its potential not only to reach a wide range of consumers but also to enable them to experience unique online reviews generated by other consumers. Online consumer review can be a new element of the marketing communication mix and can serve as a free "sales consultant" to help consumers identify the product that best suits their unique usage conditions (Chen & Xie, 2008)
According to eMarketer (2013), 92% of consumers read online reviews before they make a decision of purchases, and 67% of sales of consumer goods are based on UGC (User Generated Content). The significance of UGC lies in its ability to influence consumers’ attitudes and purchase behaviors; Online reviews are created, shared, and consumed by users and therefore is usually perceived as credible and trustworthy (Lee, Becker, & Potluri, 2016; Mudambi & Schuff, 2010). The impact of online reviews is particularly salient when it comes to experiencing intangible products such as hotels or travel-related products which might not be easy to be evaluated prior to actual consumption (Kwok & Yu, 2016; Litvin, Goldsmith, & Pan, 2008; Xie, Zhang, & Zhang, 2014; Zhu & Zhang, 2010). According to the Travel Trends Report in 2018, 9 out of 10 travelers perceive that online reviews are important in making purchases of travel-related products such as hotel bookings.
Due to the emerging interests on the impact of UGC or online reviews, a considerable body of studies in not only tourism and hospitality contexts but also other contexts (e.g., books, TV shows, and movies)have attempted to examine the performance of online reviews in different forms including review valence or frame (Tang, Fang, & Wang, 2014; Vermeulen & Seegers, 2009), numerical ratings (Moe & Trusov, 2011; Mudambi & Schuff, 2010; Ye, Law, Gu, & Chen, 2011), and text (Liu, 2006; Sonnier, McAlister, & Rutz, 2011). A general consensus suggests that positive online reviews generate higher product sales by enhancing customers’quality expectations and attitudes toward a product, whereas negative online reviews lower these factors. Despite this extensive research, an important knowledge gap remains with regard to the effect of online reviews in the hotel industry. First, less attention has been on examining the effects of neutral online reviews systematically even though in real life, reviews are rarely presented in isolation and most consumers are likely to find a mix of both positive and negative reviews online. Without the examination of neutral reviews, prior studies propose an implicit assumption that it has no or little effect on purchase behavior (Tang et al., 2014). Thus, to reflect the reality of online review sites, it is necessary to compare the performance of different review valence including neutral valence reviews.
Second, some studies regard online reviews as the only information source affecting buying behavior. According to Local Consumer Review (2015), reflecting online reviews is one of the final stages in the purchase path, and 82.5% of tour and active bookings are made directly on company websites (Travel Trend Report by Trekksoft, 2018). Thus, online reviews might be used to re-confirm information of, or attitudes toward products that are formed from other information sources such as company websites or OTAs. This means that to some extent, consumers may form their attitudes toward certain travel products and buying intention before looking over online reviews, which, in turn, influence their pre-formed attitudes and booking intention. In this regard, it is necessary to explore the role of various information sources in facilitating consumers’ pre-formed perceptions on online reviews and compare it to their perception after exposure to online reviews. However, prior studies have mostly focused on examining the direct effect of online reviews on the perception of attitudes and booking intention rather than exploring the change of attitude and booking intention.
In addition, after parsing online reviews, consumers subsequently evaluate the usefulness of online reviews, and most online review sites allow them to vote for the usefulness of reviews, which in turn may influence customers’ attitudes and buying intention (Feng, 2016; Zhao, Wang, Guo, & Law, 2015). That is, customers are expected to form different attitudes and evaluations on given products based on not only reviews’ types or valence, but the degree of the perceived usefulness of reviews. However, prior studies related to the usefulness or helpfulness of online reviews have focused primarily on salient factors influencing customers’ perceived usefulness of reviews (e.g., Mudambi & Schuff, 2010; Racherla & Friske, 2012) rather than examining its effect as an independent or moderating variable on purchasing behaviors such as attitudes toward targeted products and booking intention.
In this line, the objective of this study is twofold. First, this study examines the effects of online reviews on the change of consumers’ attitudes and booking intention. This study focuses mainly on online reviews from OTA sites as a type of online distribution channels such as electronic press releases. To reduce the aforementioned research gaps, this study includes not only positive and negative reviews but also mixed-neutral reviews that contain equal amounts of positive and negative reviews in the context of OTAs. Second, this study also investigates the combined effect of the inclination of online reviews and perceived usefulness of reviews on consumers’ attitude changes. To achieve this goal, this study used an experimental design by creating a mimicked hotel company’s website and online reviews extracted from several OTA sites such as Tripadvisor.com.
2. Literature Review
2.1. Online review and its effect
Online reviews, often used interchangeably as eWOM (Electronic word-of-mouth), online recommendations, or online opinions, have gained importance with the emergence of new technology tools, especially social media (Chen & Xie, 2008). It is referred as all informal communication directed at consumers through Internet-based platforms related to the usage of characteristics of particular services or products (Litvin et al., 2008). Travel-related online reviews generally include the real experiences of travelers, even positive, neutral or negative evaluations. Due to the intangibility of travel-related services, consumers continue to search for more external information until their perceived uncertainty and risk are reduced to a certain level (Chen & Xie, 2008; Sparks & Browning, 2011). In this regard, it has been noted that many tourism businesses attempt to use an OTA as their online distribution channel that possesses past consumers’ experiences and comments on a given service and/or product, enabling potential consumers to determine if their decision might be the best choice (Verhoef, et al., 2015). Since, of all the information sources, consumers tend to turn to non-commercial distribution channels due to its reliability, online customer reviews are likely to influence consumers’ perceptual attitudes and purchasing behavior in decision-making processes (Jeong & Jeon, 2008; Mudambi & Schuff, 2010). However, previous studies (see [Table 1]) have revealed that the effects of online reviews vary in different forms such as rating/numerical evaluation or text, valence (positive, negative or neutral), review volume and purposes (Park & Allen, 2013).
[Table 1] Online review and its effect
For example, Chen (2008) found that recommendation from other consumers has a more powerful influence on product choices than did reviews from firm related advisors. Considering different review frames, a considerable body of study found that booking intention was higher following exposures to positive review valence than negative review valence (e.g., Sparks & Browning, 2011; Tsao, Hsieh, Shih, & Lin, 2015). Some studies used customer’s rating scores as one of the online review types (e.g., Ye, Law, & Gu, 2009; Ye et al., 2011). Using data from a major online travel agency, Ye et al. (2011) found that online reviews have a significant effect on online sales, with 10 percent increase in review ratings boosting online booking by more than five percent. Several studies examined the effect of review quantity. For example, Tsao et al. (2015) found that increasing the number of reviews can amplify the effect of review valence on facilitating booking intentions. Lee, Park, and Han (2008) revealed that as the proportion of negative online consumer reviews increase, high-involvement consumers tend to conform to the perspective of reviewers, depending on the reviews’ quality.
The indirect effects of online reviews on purchase decisions have been often discussed in the literature. For example, Vermeulen and Seegers (2009)’s experimental study revealed that consumers’ mere exposures to online hotel reviews (regardless of review valence: positive or negative) increased awareness and probability that they would include those hotels in their choice sets for decision-making consideration. In this line, online reviews may play an important role in influencing not only purchase decisions like accommodation booking, but also attitudes and awareness about a targeted product.
Among diverse forms of online reviews, this study focuses primarily on understanding the impacts of online review valence. Especially, this study emphasizes the importance of considering the role of neutral review valence in exploring the effects of online reviews. Although there have been the aforementioned various attempts to explore the impact of online reviews, less attention has been made on the effect of neutral reviews in comparison to positive and negative reviews (Purnawirawan, De Pelsmacker, & Dens, 2012; Tang et al., 2014; Tsao et al., 2015). In addition, given that consumers use online reviews as one of their final stages before they make a purchase decision, this study attempts to analyze the impact of online reviews on changing pre-formed attitudes and purchase intention of customers.
2.2. Conceptual framework and hypotheses
2.2.1. Review valence and the change of attitude and booking intention
During the purchasing process, consumers try to find product attribute-value information and recommendations from various information sources. By acting as an informant and recommender, online consumer reviews exert the ability of influencing the decision-making process of consumers. In addition, online reviews provide customers with diverse access to prior experiences on which they can base their attitudes, beliefs, or trust that a firm or product will deliver quality services (Lee et al., 2008; Sparks & Browning, 2011).
Online reviews vary in not only content but also in the valence of the success or failure of the product or service (Sparks & Browning, 2011). Review valence refers to the positive, negative or neutral nature of the statements in the message (Buttle, 1998; Ketelaar, Willemsen, Sleven, & Kerkhof, 2015). The valence of online reviews communicates perceived quality and potential purchase risks. Thus, it can be the main sources of information available to enhance attitudes and facilitate purchasing decisions (Walters, Sparks, & Herington, 2007).
Regarding hotel bookings, willingness to book is dependent on whether a potential consumer forms a positive attitude toward a targeted hotel. Product attitude can have a significant influence on purchase intentions and buying behaviors (Ajzen & Fishbein, 1980). In the process of hotel booking, when potential hotel visitors read relevant travelers’ reviews about a hotel, they form overall evaluations on the expected service quality of the hotel, and their expected satisfaction, which subsequently leads to booking intentions (Tsao et al., 2015; Vermeulen & Seegers, 2009). One predominant belief is that positive online reviews boost product sales like hotel bookings by leading to positive attitudes toward a product because when customers read positive reviews about a product, they infer information about satisfaction, usage experiences, and recommendations (Tang et al., 2014). On the other hand, negative reviews can reduce consumer interests in the product or service, which in turn affect customers’ attitudes and buying intentions. In an empirical study, Ye et al. (2009) confirmed that positive online reviews can significantly increase the number of bookings in a hotel. Lee et al. (2008) found that as the proportion of negative reviews increased, so too did consumers’s negative attitudes.
Tang et al. (2014) argued that positive and negative online reviews both have clear opportunities for consumers to process product-related information. An experimental study by Vermeulen and Seegers (2009) confirmed that the positive, as well as negative reviews, increase the level of consumer awareness of hotels while, additionally, positive reviews increase positive attitudes toward a hotel. Spark and Browing (2011) explored the role of four framing factors that influence hotel booking intention and perception of trust: the target of the review, overall valence of a set of reviews, and the existence of numerical rating. As a result, they found that consumers seem to be more influenced by early negative information, especially when the overall set of reviews is negative. In addition, the study showed that positively framed reviews together with numerical rating details increase booking intention.
[H1] Exposure to positive (vs. negative) online reviews leads to positive (vs. negative) attitude change
[H2] Exposure to positive (vs. negative) online reviews leads to positive (vs. negative) change of booking intention
[H3] Attitude change has a significant effect on positive change of booking intention
As well as the effect of positive and negative reviews, this study also considers the impact of neutral reviews. Even though in real life, reviews are rarely presented in isolation, and for most products consumers are likely to experience a mix of both positive and negative reviews (Purnawirawan et al., 2012), yet most studies in a hospitality context have focused heavily on examining the effect of positive and negative reviews and to our knowledge, there have been only a few studies that tested the effect of neutral online reviews on consumer behavior in a hospitality context (e.g., Duan, Gu, & Whinston, 2008; Lee et al., 2008; Purnawirawan et al., 2012).
Neutral reviews can be distinguished with two types: mixed and indifferent neutral reviews. An indifferent neutral review contains neither positive nor negative claims such that it lacks any dominant attitude and subjective preferences. In contrast, a mixed neutral review contains equal or similar amounts of positive and negative claims leading to balanced evaluations, attitudes, and/or emotions. Distinguishing two types of neutral reviews, Tang et al. (2014) stressed that focusing on only positive and negative online reviews while ignoring its neutral form produces biased results of the online review-performance relationship. In that online reviews can have both positive and negative simultaneously and independently, they criticized one the dimensional view that as a negative (positive) evaluation for a product increases, the positive (negative) evaluation seemingly should decrease. In addition, d’Astous and Touil (1999) claimed that consensus is the most influential external cue, and a stronger consensus among reviews is better able to inspire the trust of consumers and to process information. In this sense, neutral review valence should be separately examined from positive and negative reviews because it contains almost equal amount of positive and negative claims, which means a large number of agreements on and confidence in product evaluation have not been made. Therefore, adopting the notion of mixed neutral reviews, this study also analyzes the effect of positive and negative reviews in the comparison to neutral reviews.
2.2.2. Moderating effect of review usefulness on the relationship between review valence and attitude change
Consumers come to believe or disbelieve the messages of online reviews through personal traits, socialization or purchasing experience. Although the rapid increase of online review sites provides easy access to diverse information about products and services, it also fosters consumer skepticism toward online reviews derived from fake or unauthentic reviews (Racherla & Friske, 2012). Because of different levels of skepticism, individuals may have different levels of usefulness for online reviews. This is because consumers also make an effort to evaluate whether or not a review message provides an accurate representation of the product (Buda, 2003; Racherla & Friske, 2012). In other words, perceived usefulness of an online review may vary depending on not only product type and customer’s personality but also review valence. Recent empirical studies have confirmed that the usefulness of online reviews, which is often measured by the helpfulness votes that a review received, have significant positive effects on travelers’ intention of booking a hotel room (Zhao et al., 2015). In this line, this study strives to elaborate the decision process related to the usefulness of online reviews. That is, this study proposes that the level of usefulness of an online review moderates the effect of an online review on customer’s attitude changes, which in turn influences the change of booking intention.
According to the Elaboration Likelihood Model (ELM) developed by Petty and Cacioppo (1986), attitude change may occur via two routes: the central route and the peripheral route. The ELM can be used to explain the processes that are responsible for changing attitudes and for enhancing the strength of attitudes. The likelihood of elaboration is influenced by an individual’s motivation and ability to process information. The central route is characterized by a high level of motivation and the ability to process the persuasive message, while the peripheral route is characterized by using contextual cues to process the message (Petty, Cacioppo, & Goldman, 1981). Individuals taking the central route think critically about issue-related arguments or opinions and scrutinize the merits and relevance of those arguments before forming an attitude about a product. Conversely, individuals using the peripheral route make less cognitive effort and rely on shortcuts such as the number of arguments and physical attractiveness of endorsers when forming an attitude (Petty & Cacioppo, 1986). In general, attitude change formed through the central route is stronger, more lasting and resistant to contrary information than the peripheral route.
This theory can help to explain the reaction of consumers to online consumer reviews by focusing on the information processing procedures that consumers follow in response to online consumer reviews. If the message is perceived ambiguous, skeptical, or not useful with respect to the customer or if the receiver is unable or not motivated to listen to the message, then the receiver will look for a peripheral cue. In this line, as customers’ perceived usefulness of reviews increase, consumers have greater motivation to understand the salient information. Thus, a higher level of review’s usefulness may lead to a higher level of motivation to process information, which in turn helps customers strengthen their attitudes toward a product consistent with the valence of the online reviews that they perceive. Therefore, this study hypothesizes:
[H4] The level of perceived usefulness of online reviews moderate the effect of online reviews on attitude change
3. Methodology
3.1. Research model
Based on the literature review, this study proposes the research model in Figure 1. By distinguishing three types of review valence: positive, neutral and negative, this study examines the effect of online reviews on changing customer’s attitudes and booking intention. Based on the ELM, the study also analyzes the moderating effect of perceived usefulness of a review on the relationships between online review valence, attitude change, and booking intention change.
[Figure 1] Research mode
3.2. Context and Design
This study used an experimental design method by manipulating hotel websites and online hotel reviews. The hotel industry was selected as a study context because of not only its economic importance, but also its indispensable reliance on social media marketing (Noone, McGuire, & Rohlfs, 2011). The hotel and resort’s official website shown in [Figure 2] was designed and its contents such as photos, text, links, and categories were manipulated from hotel and resort’s websites that actually exist. A new name for the hotel was given because it is assumed that familiarity with existing names of hotels would cause the subjects to draw upon their pre-knowledge and pre-evaluations about the product (Sparks & Browning, 2011). To design review valence frames, we extracted a variety of online reviews including negative and positive reviews from three popular OTAs’ websites: hotels.com, Tripadvior.com, and Agoda.com. For positive and negative review valence, a total of 10 reviews were included in each valence frame (see [Figure 2]). To ensure the reality of online reviews, two contrast reviews were included in each valence frame (e.g., for a negative valence, eight out of ten reviews were negative, while the other two were positive). For the neutral valence frame, this study used the concept of mixed-neutral valence frames (Tang et al., 2014) with equal amounts of positive and negative reviews which is expected to result in balanced evaluations, attitudes, and/or emotions.
[Figure 2] Manipulated hotel website and hotel review (positive and negative)
3.3. Participants
A student sample has been often used in experimental studies that examined the effect of online reviews (e.g., Ladhari & Michaud, 2015; Lee & Gretzel, 2011; Xie, Miao, & Lee). This is because the purpose of such studies is to examine the effect of different types of online reviews rather than to occurrences of a phenomenon in a population. Therefore, a student sample was also deemed appropriate in this study. Students from three universities (Kyonggi, Kyunghee and Sejong University in Seoul, South Korea) who are specializing in tourism and hospitality management were recruited via an e-mail containing a link to the questionnaire. The period of this survey lasted from Jun 1, 2014 to Jun 31, 2014. Samples of 414 respondents were randomly assigned to one of review valences (positive 149, negative: 139, and neutral: 126). 297 were male and 112 were female. The mean of age was 23. Most of participants (97%) indicated they had experience with purchasing travel products such as accommodation, restaurants, transportation, and so on.
3.4. Procedure
At the beginning, all participants were provided with the vacation scenario, telling them they had to imagine that they had already decided on booking a vacation somewhere, and were beginning to look for an appropriate hotel. First, all participants were asked to explore the manipulated hotel’s official websites between 10-12 minutes, and then they answered questions to measure their initial level of attitude and booking intention. After that, randomly assigned customer reviews (positive, neutral, and negative) were shown to participants and they were asked to spend at least 2 to 4 minutes for reading them. After reading the customer reviews, participants answered the same questions for attitude and booking intention. In addition, participants answered questions to measure the usefulness of reviews after the exposure to online reviews.
3.5. Measurements
Booking intention was measured by one item with 10 point-paired anchors (likely-unlikely): I would like to book this hotel & resort for my upcoming vacation). With regard to attitude change (ATC), the scales of attitude were adapted and modified from Bizer, Tormala, Rucker, and Petty (2006) and Purnawirawan et al. (2012). Respondents’ attitudes toward the hotel and resort were evaluated in relation to three items (bad–good, unsatisfactory–satisfactory, and unfavorable–favorable) with 7 point-paired anchors. The questions on personal attitudes towards online consumer reviews were combined as one factor [KMO: 0.864, X2 =1582.958(p<.0001); Cumulative variance explained: 86.016; Eigenvalue; 3.441; Cronbach’s alpha: 0.945]. The values of ATC and booking intention change (BIC) were obtained by subtracting the value before the exposure to online reviews and the value after the exposure (value after exposure to online review minus value before the exposure). Perceive usefulness of reviews (PU) was measured by two items (Bailey & Pearson, 1983; Purnawirawan et al., 2012): I found the reviews useful, the review helped me make a decision regarding this hotel. The items for PU were also combined as one factor [KMO: 0.564, X2 =336.314 (p<.0001); Cumulative variance explained: 87.392; Eigenvalue; 3.441; Cronbach’s alpha: 0.853].
4. Results
The current research model requires a moderation analysis to determine whether the size or sign of the effect of review valences on attitude change depends on a moderator variable, perceived usefulness of reviews. Thus, the hypotheses were tested by a PROCESS modeling technique (model 7) proposed by Hayes (2012). In that the current research is based on a mediated moderation model, the PROCESS technique is considered suitable as it is often used to generate conditional effects in moderation models and conditional indirect effects in moderated mediation models with a single or multiple mediators.
4.1. Manipulation checks
To make sure that the manipulations were reliable and effective, several manipulation checks were conducted. First, the check of review valence was conducted by a 7 point-semantic differential scale question after participants read randomly assigned reviews: ‘the online reviews I have just read are (1: negative–7: positive)’. The result of the ANOVA in table 2 showed significant mean differences among three groups (F=193.72, df=2. p<.001), which showed that the manipulation of review valence was successful. That is, respondents exposed to a positive review perceived it as more positive than those exposed to a negative review. Second, the believability of the manipulated website was tested by a 7 point-Likert scale question: ‘I believe that the website I have just explored is the official website of the hotel’. The mean value was 4.50. These results of manipulation checks confirmed the successful manipulation of the reviews and websites.
[Table 2] The ANOVA test of review valence
4.2. Hypotheses testing
With regards to positive versus neutral reviews, the positive reviews positively increased the degree of attitude change (β=.858, p<.0001). The value of review usefulness was not significant on attitude change (β=.107, p>.05). However, regarding the interaction between review valence and usefulness, the interaction term was statistically significant (β=.475, p<.0001). The moderation effect tested by a Boostrapping technique showed a significant moderating effect of review usefulness on the relationship between review valence and attitude change (see [Table 4]). It indicated that although the usefulness of the review did not have a direct effect on attitude change, the positive reviews with the higher level of perceived review usefulness led to the stronger magnitude of positive attitude change (Bootstrap confidence interval: β=.675~1.520) than the lower level of perceived review usefulness (Bootstrap confidence interval: β=.0723~.498). As expected, the positive reviews led to a positive change on booking intension (β=.959, p<.0001), and attitude change also revealed a significant impact on the positive change of booking intention (β=.793, p<.0001).
[Table 3] Results of model testing-positive vs. neutral
[Table 4] The valence x PU interaction effect for ATC-positive vs. neutral
With regard to negative versus neutral reviews, the negative reviews decreased the degree of positive attitude change (β=-.466, p=.004). Unlike the case of positive versus neutral review group (see [Table 5]), the effect of review usefulness on attitude change was significantly negative (β =-.390, p<.0001). The result may be explained by the concept of negative dominance as one of negativity bias suggested by Rozin and Royzman (2001), which occurs when the holistic perception of positive and negative information is more negative than the sum of the subjective values of individual pieces of information. That is, it may be a plausible explanation that although neutral reviews contained the equal amounts of positive and negative, the positive side would have been more prominent than the negative side, which in turn, results in a negative effect of review usefulness on attitude change. For the interaction effect, the interaction term between review valence and usefulness was not significant on attitude change (β=-.076, p=.646). Interestingly, in the case of negative reviews versus natural reviews, even though the negative reviews showed a significant effect on attitude change, its direct impact was not significant on booking intention’s change (β=-.255, p=.094) and only attitude change had a significant effect on the change of booking intention (β=.802, p<.0001).
[Table 5] Results of model testing-negative vs. neutral
[Table 6] The valence x PU interaction effect for ATC-negative vs. neutral
For positive versus negative review valence, the exposure to positive reviews increased the degree of positive attitude change (see [Table 7]). As expected, the effect of positive reviews in the case of the positive versus negative review group on attitude change was stronger than the one in the positive versus neutral review group (β=.298, p<.0001 for the positive versus negative review group; β=.858, p<.0001 in the positive versus neutral review group). Although the usefulness of the review was not significant on attitude change, its interaction effect (β=.551, p<.0001) was significant. As shown in [Table 8], the result of moderating effect by Boostrapping indicated that attitude change is greater when the positive review contained the higher level of perceived review usefulness. It implies that positive reviews with the higher level of review usefulness lead to a higher degree of positive attitude change. Unlike the case of negative versus neutral review in which the valence of negative reviews did not influence the change of booking intention, positive reviews showed a significant effect on the change of booking intention (β=1.294, p<.0001).
[Table 7] Results of model testing-positive vs. negative
[Table 8] The valence x PU interaction effect for ATC-positive vs. negative
In sum, as many studies supported, positive reviews have a significant effect on the positive change of attitude and booking intention when its effect was compared to not only negative but also neutral reviews. Thus, [hypothesis 1] was accepted. However, when negative reviews were compared to neutral reviews, we could not find its significant effect on booking intention change but its effect on attitude change was significant. By considering the existence of neutral reviews, the study concluded that [hypothesis 2] was partially accepted. In any case of review valence, attitude change revealed a significant effect on the change of booking intention. Thus, [hypothesis 3] was accepted. With regards to the moderating effect of review usefulness on the relationship between review valence and attitude change, when negative and neutral reviews were included in the analysis, the interaction effect was not found. On the other hand, the positive review with the higher level of perceived review usefulness showed a significant effect on attitude change in comparing to both negative and neutral reviews. Thus, [hypothesis 4] was also partially accepted.
5. Discussion and Implications
An ever-increasing number of consumers are relying on online distribution channels to gather information about their product and service because they have easy access to consumers’ product review based on actual user experience. For hotels, OTAs are more than a distribution channel. The hotel industry is strongly influenced by eWOM, and especially online reviews, posted on online distribution channels such as OTAs’ sites, have a significant impact on the purchase decision process of potential consumers (Tsao et al., 2015). By employing an experimental design method, the objective of this research was twofold. First, the current study examined the effects of online reviews on changing consumers’attitudes and booking intention by distinguishing three types of online review valence: positive, negative, and neutral review valence. Second, this study also sought to prove how the perceived usefulness of reviews moderates the influence of review valence on consumers’ attitude change. The findings of the current study found evidence to support the view that the impact of online reviews on purchase decisions depend on the types of the review valence. The results showed that positive reviews have a significant effect on the positive change of attitude and booking intention in comparison to both negative and neutral reviews. However, the effect of negative reviews was not significant on the change of booking intention when its effect was compared to neutral reviews. The interaction effect of usefulness review between review valence and attitude change was significant only when positive reviews were included in the analysis.
5.1. Theocratical Implications
Based on findings, this study theoretically contributes to the existing online review-related literature in the following ways. First, with regard to [hypothesis 1 ]and [hypothesis 2], the study offers new insights into literature related to the performance of neutral and negative reviews. This study found several interesting results related to the impact of negative and neutral reviews by separating online review valence into three types. With limited empirical evidence of neutral review’s impacts, there has been a general assumption that the neutral review has no effect on purchase decisions (Tang et al., 2014). However, this study indicates that although participants perceived neutral reviews more positively than negative reviews, there is no difference in the effect of neutral and negative reviews on the change of booking intention. It seems that the effects of neutral reviews on the change of booking intention are not truly neutral, and rather neutral reviews act as negative reviews. As reviewed, most studies have examined the impact of positive (negative) reviews in comparison to negative (positive) reviews, and a general consensus is that the positive reviews increase product sales, including purchase intention or vice versa. The study also found the powerful impact of positive reviews on the change of attitude and booking intention when compared to any type of review valence (neutral and negative reviews). However, when compared to neutral reviews, surprisingly, the effect of negative reviews was not significant on changing booking intention. The result is supported by the study of East, Hammond, and Lomax (2008) which revealed that consumers ignore advice from a negative review if they are very likely to choose a brand.
Second, review usefulness from negative and neutral review groups showed a negative influence on changing attitude towards the hotel. The result may be explained by the concept of negative dominance as one of negativity bias suggested by Rozin and Royzman (2001), which occurs when the holistic perception of positive and negative information is more negative than the sum of the subjective values of individual pieces of information. That is, it may be a plausible explanation that although neutral reviews contained equal or similar amounts of positive and negative reviews, the positive side would have been more prominent than the negative side, hence the negative effect of review usefulness on attitude change. A similar result was also found in Rachelar and Frisky (2012)’s study where negative reviews are perceived to be more useful than either extremely positive or moderate reviews. Perceived lack of information is another plausible explanation. Zhu and Zhang (2010) found that online reviews are more influential for less popular products and suggested that even one negative review can be detrimental because the role of an online review becomes more salient in a circumstance in which alternative means of information acquisition are relatively scarce. In this line, because this study manipulated a hotel website and gave the hotel a new name for participants, they might have perceived the scarcity of available information about the hotel resulting in sensitive reaction to negative reviews.
Third, regarding to [hypothesis 4], this study also provides empirical evidence to support the effect of online review valence varied by the extent of perceived review usefulness. Especially, the positive reviews with a higher level of perceived review usefulness are more powerful in changing attitude. However, when negative reviews were compared to neutral reviews, the interaction effect was not significant. This means that in case of negative versus neutral reviews, the extent of attitude change is not significantly different according to the level of perceived usefulness of review. Not like other studies in which the value of attitude is measured after reading reviews, this study measured the change of attitude by subtracting an attitude value after exploring a hotel website and the one after reading reviews. Thus, this finding can also be explained by the result of East et al. (2008) which proved that consumers are likely to ignore the claims from a negative review if they are very likely to choose a brand. That is, people exposed to negative reviews had been less willing to change their pre-formed attitudes than those exposed to positive reviews.
5.2. Practical implication
Some managerial implications can be also made. First, regarding [hypothesis 3], managers should appreciate the fact that neutral review valence can be damaging for the perceptions of potential travelers such as attitudes or review usefulness, which in turn influences booking intention change. Thus, it is very crucial to keep the complaints of dissatisfied customers within the hotel rather than being expressed on online review sites such as OTAs (Mauri & Minazzi, 2013; Tsao et al., 2012). Ironically, hotels need to encourage dissatisfied customers to use offline means to complain about service failure such as phone services, which helps reduce the portion of negative reviews on online review sites. According to Harris Interactive, 75% of customers think that it takes too long to reach a live agent on the phone, which results in the increase of customer’s reliance on social media and review sites. Therefore, to reduce the likelihood that customers post their complaints on review sites, a hotel needs to offer diverse communication channels to ensure easy and timely access of dissatisfied customers such as a live chat support that would allow customers to send instant messages to contact customer service.
Second, it is also important to prevent the expansion of negative reviews or complaints from customers once negative reviews have been posted on review sites. As Tsao et al. (2012) founded, repeated exposure to negative reviews is particularly damaging to booking intention and negative reviews are more easily spread than positive reviews. Additionally, as recent studies indicated (Levy, Duan, & Boo, 2013; Sparks, So, & Bradley, 2016; Ye et al., 2011), whether the business posts a response is likely to affect how others perceive the brand and possibly influence their purchase decision. Therefore, monitoring the content of online reviews on OTA sites is strongly required on a regular basis and thus, a hotel can respond to unsatisfied reviewers in a timely manner. The key to successfully managing OTA use is to evaluate their full value as a distribution channel and as a marketing platform versus the benefits of other channels and media outlets. Not only timely responses to negative reviews but also the quality of responses should be ensured. Thus, hotel managers are encouraged to design a set of standard review responses, so-called "canned"responses that help employees reply to reviewers with appropriate ways and guarantee the quality of responses and service consistency.
Lastly, it is important to develop strategies to increase the perceived usefulness of reviews. It is difficult for businesses to know whether the reviews posted are perceived as useful or not. However, several studies (Racherla & Friske, 2012) revealed that there are factors influencing customer’s perceived usefulness of a review that includes reviewer reputation, expertise, and a reviewer’s identity. That is, when reviews are given with a combination of such factors, customer’s perceived review usefulness can be increased. Therefore, it would be an effective way to provide links to hotel’s official websites that can lead customers to positive reviews written by experts or power bloggers. It is also recommended to extract positive reviews that contain factors affecting the usefulness of reviews on other OTA websites and then, display them in the form of pop-ups on an official website.
5.3. Limitation and future research
Despite the contributions of this study, the current study is still subject to a number of limitations. First, this study only considered the valence of mixed-neutral reviews. However, in real life, customers are likely to encounter not only mixed-neutral reviews, but indifferent-neutral reviews which contain neither positive nor negative claims. As Tang et al. (2014) emphasized, ignoring mixed- or indifferentneutral reviews leads to substantial under- or overestimates of the influence of positive and negative reviews. Thus, future studies need to compare the different effect of mixedand indifferent-neutral reviews on business performance. Second, the current study did not consider other factors that influence the change of attitude and booking intention such as review quality, quantity, and characteristics (e.g., gender and expertise). It is a plausible expectation that if those factors are considered with review valence at the same time, the effect of online reviews may be different from the findings of the current study. Third, a new brand name for the manipulated hotel was given to avoid familiarity with existing names of hotels, but a new brand may also mean a less popular product. As Zhu and Zhang (2010)’s study indicated, online reviews are more influential for less popular products or services. In this sense, it is worthwhile to conduct this study in a context in which online reviews for well-known hotel brands are given to study participants. Lastly, even though the effect of online review depends on other considerable factors such as customer’s demographic characteristics (e.g., age), the frequency of online purchasing, and internet proficiency, this study could not consider the control variables in an experimental design. Thus, it is required to develop more specified research design when it comes to considering customer’s characteristics and other salient cognitive influencers or cognitive factors.
Cited by
- Analysis of Korean millennials’ travel expenditure patterns: an almost ideal demand system approach vol.25, pp.1, 2020, https://doi.org/10.1080/10941665.2019.1578810