INTRODUCTION
The increasing emergence of user-generated content (UGC) during the last decade has substantially changed consumer behaviour. The expansion of online reviews on various websites and social media has led to a new form of multidirectional information sharing, known as electronic word of mouth (eWOM). Similar to other industries, the tourism and hospitality industry has to deal with the growing importance of consumer-generated media in the decision-making process of potential guests through online reviews (Li et al. 2019;Sparks and Browning 2011;Zhang, Gao and Zheng 2019). Unlike conventional goods, hotel services cannot be tested before consumption, so customers who lack prior experience of a hotel tend to make conclusions based on past guests' reviews and subsequent hotel responses (Sparks and Browning 2011). Hence the growing interest of scholars and practitioners in this topic has emerged.
The role of consumer-generated media in the hotel industry has been researched in terms of various predictors that lead to eWOM behaviour (Bakshi, Dogra and Gupta 2019;Boonsiritomachai and Sud-On 2020;Liang et al. 2013;Yen and Tang 2019). Numerous studies have examined the impact that online reviews have on potential guests’ bookings intentions (Cheng et al. 2019;Kwok, Xie and Richards 2017;Litvin, Goldsmith and Pan 2018) and trust (Cheng et al. 2019;Fan and Lederman 2018;Zhao et al. 2015). Moreover, research shows that eWOM affects the consideration sets, attitudes, and behavioural intentions of review readers and, ultimately, sales (Chevalier and Mayzlin 2006;Dellarocas, Zhang and Awad 2007).
The reviews that generate even more powerful word of mouth (WOM) and have a dominating effect on booking intention (Wen et al. 2020) are the negative ones. They can easily damage a hotel's reputation, because of the high value and helpfulness they have for Internet users in general (Browning, So and Sparks 2013;Fernandes and Fernandes 2018;Levy, Duan and Boo 2013;Sparks and Browning 2010). Since customers no longer tend to express complaints and dissatisfaction in one-to-one communication, but over user-generated media (Roozen and Readts 2018;Sangpikul 2021), hotel managers nowadays have to cope with online content over which they have almost no influence. Although it is well-known that customers' reviews are available online to many people (Hennig-Thurau et al. 2004), it should be noted that management responses to those reviews also can be seen. Hence, the purpose of responding to review is not only to appease the complaining customer but also to gain the affection of those who are “over-hearing” that conversation on user-generated media (Zhang and Vásquez 2014). Therefore, developing an appropriate strategy for handling online reviews is becoming a significant challenge for companies today.
According toCasado-Díaz et al. (2020), negative reviews need to be responded to because thereby managers create a three-way network between the hotel, current dissatisfied guests, and potential guests (Li, Cui and Peng 2018). Research has shown that in terms of the trustworthiness of online reviews by guests and of content supplied by the provider itself, potential guests consider online reviews more credible (Fernandes and Fernandes 2018;Levy, Duan and Boo 2013;Racherla, Conolly and Christoduolidou 2013). In addition to that, guests find the reviews with managerial responses more useful than the unresponded ones (Kwok and Xie 2016). Hence, management response strategies play a significant role because they affect satisfaction, loyalty, and retention of potential guests (Zhang et al. 2020), as well as the companies’ financial performance (Kim, Lim and Brymer 2015;Lui et al. 2018;Xie, Kwok and Wang 2017).
Only a small number of studies, however, have been conducted observing this point of view. Due to the limited understanding of using appropriate strategies when responding to a negative online review, both the managerial and academic literature call for further research on this topic (Liu et al. 2020;Nghiêm-Phú 2018;Van Noort and Willemsen 2011;Wang and Chaudhry 2018). Existing studies primarily focus on confronting two opposed strategies, such as generic versus personalized (Roozen and Readts 2018;Wei, Miao and Huang 2013) or accommodative versus defensive (Casado-Díaz et al. 2020;Li et al. 2018;Mate, Trupp and Prat 2019). However, recent studies suggested a more flexible approach by measuring different extents of response accommodativeness and defensiveness (Li, Cui and Peng 2018), as well as including face-to-face service recovery strategies (Min, Lim and Magnini 2015).
There is also growing empirical evidence of the effects that managers’ responses have on satisfaction (Einwiller and Steilen 2015;Istanbulluoglu 2017;Liang, Schuckert and Law 2017;Min, Lim and Magnini 2015) booking intention (Fernandes and Fernandes 2018;Xie et al. 2016;Zhang et al. 2019.), hotel performance (Lui et al. 2018;Xie, Kwok and Wang 2017) and subsequent reviews (Wang and Chaudhry 2018). Despite the considerable research on the impact of management responses on financial aspects of the hotel,Wei, Miao and Huang (2013) warn that consumers should be observed from a broader perspective, by using more than solely purchase-based metrics. They emphasize consumer trust as one of the main concerns of successful relationship marketing. Trust is considered a crucial concept in attracting potential guests since it mediates the booking intention (Kim, Ferrin and Rao 2008;Sreejesh and Anusree 2016), and stimulates satisfaction and retention (Chen and Barnes 2007;Wang et al. 2015). However, it has been largely overlooked by researchers, and only a few studies to date have dealt with the trust of hotel review readers (Wei, Miao and Huang 2013;Sparks, So and Bradely 2016).
Therefore, this study attempts to overcome these two research gaps by expanding the previously researched management response strategies and testing their connection to trust. The strategies were observed based on the response content considering the accommodative-defensive continuum (Coombs 1998) and active listening theory that applies to face-to-face service recovery (Gruber 2011;Min, Lim and Magnini 2015). Accordingly, from literature systematization, five main response strategies have emerged: personalization, empathy, confession, denial, and excuse (Lui et al. 2018;Lee and Cranage 2014;Min, Lim and Magnini 2015;Roozen and Readts 2018;Casado-Díaz et al. 2020;Li, Cui and Peng 2018;Einwiller and Steilen 2015). Using a mixed-method approach, all five strategies have been empirically tested, and their influence on consumer trust examined. Therefore, the paper's primary purpose is to identify the main management response strategies to a negative online review considering both their accommodativeness/defensiveness and active listening theory and to examine their influence on the trust of potential guests reading the reviews on user-generated media. By observing these main strategies together, a broader perspective is gained because they can be evaluated relative to each other. Besides the contribution this paper brings to filling the research gap, it provides hotel managers with practical information on which strategies they should use or avoid using when responding to negative reviews to gain the trust of potential guests.
1. LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT
The review of the recent literature shows the increasing interest in management response strategies as their growing importance is recognized. In general, researchers agree on the positive influence of management responses on building relationships with current and prospective customers (Gu and Ye 2014;Liang, Schuckert and Law 2017;Rose and Blodget 2016). However,Lappas, Sabnis and Valkanas (2016) have noticed that only a third of hotel reviews on TripAdvisor are being responded to. Some companies tend not to engage in discussions on user-generated media, which may occur due to certain company policies, unawareness of the review, or a lack of necessary resources. Regardless of the reason, existing research has confirmed that a negative review should be responded to (Casado-Díaz et al. 2020). AlthoughRose and Blodgett (2016) argue that the response's necessity depends on the context and situation controllability, earlier studies have found that an absence of a response to a negative review result in dissatisfying service experience (Bitner et al. 1990). Moreover, research shows that 60% of consumers expect companies to reply to their negative reviews (Van Noort and Willemsen 2011). According toKwok and Xie (2016), the reviews with managerial responses are more helpful to potential guests than unresponded ones. However,Wang and Chaudhry (2018, 21) warn that "not only does the act of responding influence subsequent opinion, but the manner in which responses are crafted plays an essential role".
Several researchers have addressed the issue of choosing the proper response strategy when handling negative online reviews, proving that an outcome may differ depending on the strategy used (Casado-Díaz et al. 2020;Jeong and Lee 2017;Lee and Song 2010;Li, Cui and Peng 2018;Min, Lim and Magnini 2015). According to the consumer inference theory (Kardes, Posavac and Cronely 2004;Kardes et al. 2008), customers make judgements about products and organizations based on limited information which is often obtained from a secondary source (e.g., online review sites) and includes only partial information on product or organization’s characteristics. To generate inferences, they create if-then linkages between the obtained cues and their conclusions (Kardes et al. 2008). In this case, different management responses can be observed as cues that lead new consumers to certain inferences, and this study observes the inference of trust. AsSparks, So and Bradley (2016) explain, responses to online reviews are related to cue-based trust since they refer to stimuli obtained from a single encounter with no prior consumption experience. They found that providing a response to a negative online review yields higher trust of prospective customers than not responding to one. However, it remains unclear how certain types of response content influence the inference of trust.
When considering the content of the response, the widely used accommodative-defensive continuum was observed. Considering this, Coombs (1998) defines seven strategies of crisis communication based on a different level of accommodativeness and defensiveness. Applying this theory to handling negative reviews in hospitality, existing research focused on the two points of the spectrum, accommodative and defensive response (Casado-Díaz et al. 2020;Li et al. 2018;Mate, Trupp and Prat 2019). However, between these two, the marketing literature recognizes another response strategy commonly used by hotel management, namely, excuse strategy (Lui et al. 2018;Lee and Cranage 2014;Liu et al. 2020). Besides the level of agreement, another issue related to response content was raised byMin, Lim and Magnini (2015), who claim that responding to negative online reviews should follow the same rules as handling complaints in person. Referring to the interactional justice theory, they argue that the way the customer is treated is more important than the actual outcome. Hence, they emphasize the importance of active listening techniques through empathy demonstration and partial complaint paraphrasing. Accordingly, based on the accommodative-defensive continuum and active listening theory, five main strategies have emerged from the literature and gained significant attention in recent studies (Einwiller and Steilen 2015;Lui et al. 2018;Lee and Cranage 2014;Min, Lim and Magnini 2015;Roozen and Readts 2018;Casado-Díaz et al. 2020;Li, Cui and Peng 2018): (1) personalization, (2) empathy, (3) confession, (4) denial, and (5) excuse. After a negative service encounter, to recover the trust of both concerned guests and potential guests who read reviews, it is of great value to determine the positive and negative effects every strategy can cause.
1.1. Personalization as a response strategy
Nowadays, it is challenging for companies to nurture their customer relationship management as much as they would like and need to. That may be why many of them resort to using generic responses when responding to online reviews. Generic responses have standardized content without any specific issues addressed in the review (Zhang et al. 2020), and although these responses demand much less time and effort, some studies show that they can result in less guest satisfaction and can lead to a decrease in the booking intention of guests (Roozen and Raedts 2018;Tam and Ho 2005;Wei, Miao and Huang 2013). Several authors (Stevens et al. 2018;Wang and Chaudhry 2018) emphasize the importance of personalization and tailor-made managerial responses to make customers feel appreciated and to increase review effectiveness (Roozen and Readts 2018). Unlike the generic responses, personalized ones are unique and tailored to each customer by addressing the specific issue raised in the review (Wei, Miao and Huang 2013). While responses to guests’ compliments may be standardized (Nghiêm-Phú 2018), negative reviews should be tailored carefully, with every complaint being acknowledged and apologized to. With respect to the active listening theory, which applies to handling complaints in person,Min, Lim and Magnini (2015) suggest that a partial paraphrase of the complaint should be included in the response to sound more personal. This type of response will lead to greater satisfaction (Min, Lim and Magnini 2015) and show the hotels’ sincerity and care (Zhang et al. 2020). SinceWei, Miao and Huang (2013) found that specific management responses also generated higher trust and communication quality when compared to generic ones, the first hypothesis is defined:
H1: Personalization as a response strategy to a negative online review positively influences the trust of potential guests.
1.2. Empathy as a response strategy
When replying to negative comments from guests, several authors suggest including an empathy statement in the response (Einwiller and Steilen 2015;Min, Lim and Magnini 2015). Since empathy is one of the main aspects of the hotel industry (Umasuthan, Park and Ryu 2017), responses like these can ease customers' anger and dissatisfaction (Zhang et al. 2020). Using this strategy, the organization shows an understanding of the emotions complaining customers feel (Min, Lim and Magnini 2015).Gruber (2011) suggests that empathy as an important part of active listening can increase guests’ perception of interactional justice in face-to-face service recovery. According toMin, Lim and Magnini (2015), handling online complaints should follow the same principle and include an empathy statement. Some studies have confirmed that the greater customer satisfaction is achieved by including empathy in management responses (Einwiller and Steilen 2015;Min, Lim and Magnini 2015). On the other hand, research shows that responses without any empathy sound generic and automatic, making the company less trustworthy in the customers’ eyes (Utz, Matzat and Snijders 2009). Research on using empathy as a response strategy for negative hotel reviews is still scarce; however, studies from other fields confirm that empathy positively influences customers’ trust (Bahadur et al. 2020;Feng, Lazar and Preece 2004). Based upon that, the following hypothesis is proposed:
H2: Empathy as a response strategy to a negative online review positively influences the trust of potential guests.
1.3. Confession as a response strategy
When faced with customer complaints, many hoteliers tend to apologize and express regret for the issue raised in the review. This type of response reflects the accommodative side of the aforementioned accommodative-defensive spectrum of responses. Various studies have shown an apology's effectiveness when handling customer complaints (Abramova et al. 2015;Casado-Díaz et al. 2020;Lee and Song 2010), perhaps because the apology is often seen as psychological compensation that helps customers regain composure (Davidow 2000). Therefore, apology (i.e., confession) strategies are commonly used since they imply taking responsibility for the problem that has arisen and expressing remorse (Li, Cui and Peng 2018). Unhappy customers expect an explanation, apology, or compensation for their inconvenience. An apology from a service provider communicates concern to the customer who has experienced a service failure, but the apologizing itself sometimes is not enough to gain the trust of potential customers. While several researchers believe this is the strategy that positively impacts guest satisfaction and booking intentions (Casado-Díaz et al. 2020;Lee and Song 2010), others (Li, Ciu and Peng 2018) warn that the result of this strategy may not always be positive. Some authors have argued that the positive impact of apology may be annulled by the admission of guilt or taking responsibility (Kerkhof, Utz and Beukeboom 2010). Nonetheless, compared to other less accommodative strategies, especially when perceived service failure is concerned, the confession and apology tend to generate more positive results (Casado-Díaz et al. 2020;Lee and Song 2010;Li, Ciu and Peng 2018;Matzat and Sniijders 2012). Hence, the following is proposed:
H3: Confession as a response strategy to a negative online review positively influences the trust of potential guests.
1.4. Denial as a response strategy
On the other hand, there is another side of the accommodative-defensive spectrum referring to companies that opt to defend themselves and deny the problem. This refutative type of response communicates a lack of empathy and reduces the trustworthiness of the company. It angers customers more than if they did not receive any response at all. (Kerkhof, Utz and Beukebook 2010;Dens, Pelsmacker and Purnawirawan 2015). Using this strategy, hotel managers deny the existence of a service failure addressed in the review either directly or indirectly by providing counter-arguments (Abramova et al. 2015;Lui et al. 2018). However, there has been evidence that this defensive strategy can be beneficial in specific settings, such as dealing with ordinary negative reviews, where there is no service failure, but only mismatched preferences or unrealistic expectations of guests (Li, Ciu and Peng 2018). Nonetheless, this strategy has often been observed as an opponent to accommodative strategies and, when it comes to a service failure, most researchers agree that denial yields less positive or even negative effects for prospective customers (Abramova et al. 2015;Einwiller and Steilen 2015;Lee and Song 2010;Matzat and Sniijders 2012). When faced with a denial response, potential customers tend to perceive the hotel service as poor and hotel manager as careless and aggressive (Treviño and Castaño 2013). It is still not clear how this type of response affects the trust of potential guests, but it might be perceived as harmful and unfair from a reader's perspective because it seems that the hotel is only concerned with its own interests, and not the customer’s. Hence, the following is hypothesised:
H4: Denial as a response strategy to a negative online review negatively influences the trust of potential guests.
1.5. Excuse as a response strategy
Apart from giving defensive responses in which the problem is denied, some companies tend to use excuses. Research often does not distinguish these two strategies; however, in contrast to denial, excuse strategy implies recognizing the service failure, but explaining it by uncontrollable factors, thus shifting the blame to a third party (Lui et al. 2018). Although it is considered a defensive strategy, it is placed between confession and denial if observing the accommodative-defensive spectrum (Coombs 1998). It is less aggressive than denial, but hotels that use this type of response want to separate themselves from the problems mentioned in reviews by rejecting responsibility (Lee and Song 2010). Since the management confirms that a problem has occurred but does not acknowledge responsibility for it, this strategy is used when hotels want to maintain their reputation at all costs (Li, Ciu and Peng 2018). Due to the common interchangeable use of terms refereeing to defensive, denial, or excuse strategy, there is a lack of evidence of how excuses affect trust inferences of potential hotel guests. When comparing previous research results for this strategy, the lack of consensus is noticeable. However, some recent studies indicate that this type of response does not cause favourable reactions in customers (Lee and Cranage 2014;Liu et al. 2020). Moreover,Liu et al. (2020) argue that when a review concerns a factor controllable by the hotel, the excuse strategy will not help gain the trust in the hotel. Therefore, the next hypothesis is defined:
H5: Excuse as a response strategy to a negative online review negatively influences the trust of potential guests.
2. METHODOLOGY
The empirical research was conducted from January to April 2019 in two phases. Given the purpose of the study, which aims at identifying main response strategies and testing relationships that have not been tested before, a mixed-method approach was applied for development purposes, as proposed byGreene, Caracelli and Graham (1989). That is to say, the qualitative method was used to help the development of the quantitative. In both phases, a scenario-based technique was used to provide a standardized stimulus for all respondents (Wason, Polonsky and Hyman 2002) and to enable them to express their attitudes in a specific context rather than in an abstract or non-contextualized way (Finch 1987).
The first phase included qualitative research implementing the focus group method to provide a contextual basis for the survey (Bloor et al. 2001) by vetting the proposed scenarios and adapting scales for response strategies. The pre-existing group was chosen, as suggested byKitzinger (1994), to stimulate interaction approximated to naturally occurring discussion, which would generate more ideas. The focus group consisted of 13 graduate students in Tourism Marketing who have successfully attended the Business Communication course in the master’s programme. They were judged as a suitable group for such an exploratory study because previous studies have proven that Internet users have a high education degree and visit travel websites more often (Hyde 2008;Andrlić and Ružić 2010). As well,Lončarić, Bašan and Gligora Marković (2015) have included a students’ group as their respondents because they have more positive attitudes towards the use of the Internet and high Internet literacy (Kah, Lee and Chung 2010).Han, tom Dieck and Jung (2018) have also used tourism students in a focus group as a voluntary and willingly accessible group of participants. The selecting procedure for the focus group members was conducted by the researchers by taking into consideration the skills that the students have successfully acquired in the Business Communication course.
When using scenarios, researchers strongly suggest checking their plausibility and perceived authenticity before the survey (Harrits and Møller 2021). Hence the focus group aimed at ensuring they are crafted as intended. The participants were given five different scenarios. Each one consisted of one real negative review from TripAdvisor and five hypothetical managerial responses, thus making the total of 5 reviews and 25 responses. The hypothetical responses were written according to the five response strategies that were researched. They were crafted following the suggestions from previous studies (Lui et al. 2018;Min, Lim and Magnini 2015;Lee and Cranage 2014;Zhang and Vásquez 2014). The participants were asked to discuss the scenarios and responses. When discussing negative reviews, they were asked to consider reviews’ realism, credibility, and severity. A credible severely negative review was intended to be chosen for the research sinceEsmark Jones et al. (2018) have found that this type of review creates a bigger attitude gap among consumers when they are faced with different types of managerial responses.
Furthermore, since the absence of appropriate scales for observed response strategies was recognized, and focus group is considered an appropriate tool to adapt scales to a new research problem (Nassar-McMillan and Borders 2002), the response strategies were also talked through. Participants were asked about characteristics of each response strategy, namely personalization, empathy, confession, denial, and excuse. Using the obtained information and thorough literature review, the scales were slightly modified by including new adapted items in the initial battery of items fromLee and Cranage (2014) andMin, Lim and Magnini (2015). Participants also discussed the dominant strategy implemented in each provided response. Thereby the suitability of the responses and their compatibility with the applied strategies were examined. According to the focus group results, the most suitable scenario, based on its realism, review severity and credibility, and response strategies’ compatibility, was chosen, and with the implementation of the suggestions for better clarity, it was implemented in further research.
The second phase involved the questionnaire design and, finally, field research using the survey method. This phase aimed at collecting quantitative data to test hypotheses and thereby determine the influence of each strategy on the trust of prospective guests. The questionnaire included the previously chosen scenario with one negative review and five hypothetical responses, each implementing a different response strategy. The same response was used, only with altered sentences, reflecting the response strategies. A set of statements followed each response using a 5-point Likert scale (1 – strongly disagree, 5 – strongly agree) to examine the responses’ realism, suitability of strategies, and trust. The items for specific strategies and the responses’ realism were adapted from previous studies (Lee and Cranage 2014;Lui et al. 2018;Min, Lim and Magnini 2015;Sparks and Browning 2011) and focus group results. Trust items were adapted fromPurnawirawan, Pelsmacker and Dens (2015) andSparks and Browning (2011).
The questionnaire was distributed in person and via social media in order to capture the responses of consumers in their natural setting, as suggested byWu et al. (2014). Due to the nature of the study, a non-random sample was chosen. Since it was considered essential that the sample consists of online users who read the comments on the user-generated media before making a tourist reservation, the homogeneous sample was used. According toCalder and Tybout (1999), the use of maximally homogeneous samples is justified since the study tests the theory of the influence of the response strategies. To reduce a non-response bias the same sample was used for all strategy’s responses. Hence, the non-response bias does not significantly affect the results of the research.
The questionnaire contained screening questions asking whether the respondents read reviews on UGC while searching for accommodation, and only those who read reviews were included in the sample. In total, 87 sets of questions per each response strategy were obtained, encompassing 435 fully completed questionnaires which were included in the analysis.
G*Power 3.1.9.7, a statistical power calculation programme based on Cohen’s sampling formula, was used to determine the post-hoc achieved power. The results are for F tests; Linear multiple regression: Fixed model, R² increase, Effect size f² = 0.15, α err prob = 0.05, number of tested predictors are 2, the outputs are Noncentrality parameter λ = 45, Critical F = 3.0261534, Denominator df = 297 and Power (1-β err prob) = 0.9999931.
The data were analysed using Smart PLS software 3.2. Since these five strategies have not been empirically tested before, to determine each strategy’s influence on trust, each one was tested individually resulting thus in five separate models. After determining the models’ reliability and validity, the hypotheses were tested using the PLS-SEM path method with 5000 bootstrap subsamples. According toHenseler, Ringle, and Sinkovics (2009), the use of PLS-SEM in the analysis is considered justified because it tests the proposed theory while providing stable calculations of the parameters. Also, PLS-SEM supports a small sample of collected data (Hair et al. 2017).
3. RESULTS
The focus group, as the first phase of the research, resulted in choosing the scenario for further analysis based on the criteria of situation realism, review severity and credibility, and the compatibility of responses with selected strategies. In this phase, the scales for specific strategies were revised, resulting in several new items adapted to the nature of the study (PER1, PER4, EMP1, EMP3, CON2, DEN4, EXC3). The items were then incorporated into scales (Appendix I).
The following phase included a survey. Within the research sample, there were 64.3% female respondents and 35.7% male respondents. The majority were aged between 21 and 30 years (68.6%), followed by those aged from 31 to 40 (17.1%) and 41 to 50 (11.4%). The least respondents were less than 20 years old (0.03%), while no respondents were older than 51 years.
At the beginning of the survey, the realism of the chosen scenario was checked. The results have shown that respondents could imagine searching for hotels online and reading their reviews (M=4.60, SD=0.78). They considered the negative review realistic (M=3.74, SD=1.00), and they could easily relate to the guest’s attitude (M=3.72, SD=1.04).
For further data analysis, the PLS-SEM method was used since it is considered to be appropriate when the research aims to predict rather than confirm structural relationships (Hair, Ringle and Sarstedt 2011). To predict the influence of individual response strategies on the potential guests’ trust, five models were developed, and the analysis was performed in two steps. First, the reliability and validity of each model were examined and, then, the hypotheses were tested. The results of the models’ validity and reliability are presented inTable 1.
Outer loadings for all presented variables range from 0.724 to 0.945. They exceed 0.70 as suggested byHair, Ringle and Sarstedt (2011), thus confirming the indicators’ reliability. During the analysis, some variables were excluded due to outer loadings lower than 0.70. The variable PER2 was excluded from the Personalization model; the variables TRU3 and TRU18, from the Empathy model; the variable TRU18, from the Denial model; and the variables EXC4, TRU18, TRU5, and EXC3, from the Excuse model. After eliminating these variables, composite reliability (C.R.) for all constructs ranges from 0.813 to 0.986, thus exceeding the required minimum of 0.70. Thereby the internal consistency reliability has been satisfied.
Convergent validity was assessed by the average variance extracted (AVE) that should be above the threshold of 0.50 (Hair, Ringle and Sarstedt 2011). All AVE values are between 0.553 and 0.848, thus confirming the convergent validity. Discriminant validity was confirmed using the Fornell-Larcker test. In all five models, the square root of the AVE of each latent variable is higher than their correlation with the other variable in a model (Fornell and Larcker 1981). Furthermore, for all constructs, cross-loadings are lower than the loading on the main construct. The HTMT ratio ranges from 0.316 to 0.844, not exceeding, thereby, the maximum of 0.9 recommended byHenseler, Ringle and Sarstedt (2015). Therefore, these results indicate the factors’ convergent and discriminant validity.
Additionally, the realism of each response and the suitability of each strategy was checked. Considering the obtained mean values, respondents considered all managers’ responses to be realistic (mean values from 3.45 to 4.17), and the responses successfully reflected the characteristics of each strategy (mean values from 3.08 to 4.36). Based on that, the necessary criteria are met, and the models’ reliability and validity are confirmed, so further analysis can be performed.
To test hypotheses and examine the influence of each response strategy on potential guests’ trust, standardized path coefficients and significance levels were calculated (Table 2).
All hypotheses are statistically significant at the level of 5%. The results show a significant and positive influence of personalization on trust (b=0.744, t=18.62), supporting, thereby, hypothesis H1. The influence of empathy on trust is also significant and positive (b=0.494, t=5.032), thus supporting hypothesis H2. Furthermore, hypothesis H3 was also supported, since confession significantly and positively influences trust (b=0.808, 18.256). On the other hand, results show that denial significantly and negatively influences trust (b= -0.323, t=2.785), thus supporting hypothesis H4. Finally, the influence of excuse on trust is also significant and negative (b= -0.578, t=6.807), thereby confirming hypothesis H5. Q2 values are all above zero, indicating good reconstruction of values and models’ predictive relevance.
4. DISCUSSION AND CONCLUSION
The expansion of online reviews on user-generated media has made them one of the main concerns of hotel managers today, especially negative reviews as they carry a greater weight than positive ones for customers. Although negative reviews are unwanted, these complaints are also an opportunity to show off great hotel customer service (Stevens et al. 2018). There is agreement among researchers that negative reviews should be responded to because not dealing with them can harm the reputation of a hotel (Lee and Song 2010). However, there is an ongoing debate about what kind of response should be given.
The research aimed to identify the main types of managers’ response strategies to negative online reviews, concerning their accommodativeness/defensiveness and active listening rules, and to examine their influence on the trust of potential guests reading reviews on user-generated media. The results have confirmed that various types of responses yield different, even opposite, levels of trust. This is in accordance with previous studies that have found different outcomes depending on the managers’ response type to the complaints (Casado-Díaz et al. 2020;Roozen and Readts 2018;Li, Cui and Peng 2018). Although it is widely accepted that negative reviews should be responded to, the act of responding itself is not enough to gain the trust of potential guests. The response should be crafted carefully to avoid causing countereffects.
The research has revealed that using personalization as a response strategy to a negative online review leads to the greater trust of potential guests. These results support some previous findings that mostly agree that personalized responses yield more benefits than generic ones (Stevens et al. 2018;Wang and Chaudhry 2018;Roozen and Readts 2018;Min, Lim and Magnini 2015). However, even when the response is adapted to a specific complaint, there is a range of strategies that can be used and they can cause different positive or negative reactions in potential guests reading these responses.
Research results show that using empathy as a response strategy to a negative review also leads to a greater trust level of potential guests. Empathic response implies showing understanding and sincere care for customers’ problems. The results are consistent with previous findings where including empathy statements in a response has resulted in greater guest satisfaction (Einwiller and Steilen 2015;Min, Lim and Magnini 2015). Findings for both personalization and empathy support the theory proposed byMin, Lim and Magnini (2015) that active listening rules for face-to-face service recovery (Gruber 2011) also apply in the online environment.
Besides including empathy, the results show that implementing confession as a response strategy also positively influences the trust of potential guests. This supports some previous findings that have recognized some benefits of using this strategy (Casado-Díaz et al. 2020;Lee and Song 2010). However, this strategy should be used with caution. Some authors (Li, Cui and Peng 2018) argue that in certain cases when the complaint is not the result of service failure, the opposite, defensive strategy, could be more beneficial. The confession strategy means that hotel management recognizes the problem, apologizes, and admits the responsibility. Since the complaints sometimes may be subjective, the managers should distinguish, depending on the type of review, whether they will accept the responsibility for the subject of the complaint.
AlthoughLi, Cui and Peng (2018) found that defending the company in some cases when there was no service failure is better than just apologizing, our findings show that using denial as a response strategy has a negative influence on potential guests’ trust. Moreover,Abramova et al. (2015) have found that even when the problem stated in a review is beyond the hotel’s control, denying it or arguing with the customer will not actually boost guests’ confidence in the hotel. This also relates to some previous findings that have shown negative implications of using this strategy when replying to complaints (Einwiller and Steilen 2015;Lee and Song 2010).
Similar results appear for the excuse strategy, which implies admitting the problem but shifting the blame to a third party or some external factors. Thus, findings have confirmed that using excuse as a response strategy also negatively influences the trust of potential guests. This appears to be contradictory to some previous studies which have found that excuses in managers’ responses can have some positive implications on guests’ trust (Abramova et al. 2015). On the other hand, similar to our results,Lee and Cranage (2014) andLiu et al. (2020) do not advise using external causal explanations because they yield an unfavourable impression of the company. This inconsistency in some findings can be explained by different types of negative reviews addressed in the studies. However, when it comes to a perceived service failure, results have revealed that the use of the personalization, empathy, and confession strategies in managers’ response positively influences trust, whereas the denial and excuse strategies have a negative influence on trusts. These findings yield several useful theoretical and practical implications.
4.1. Theoretical implications
Previous research has mostly focused on the two opposite sides of the accommodative-defensive continuum when replying to negative online reviews. This study adds to the body of knowledge by applying the more flexible approach, gauging different levels of accommodativeness and defensiveness in manager responses, as suggested by previous researchers (Li, Cui and Peng 2018). The paper provides a systematic review of the literature and identifies the three dominant strategies of the accommodative-defensive spectrum, namely, confession, denial, and excuse. Additionally, it adopts the theory proposed byMin, Lim and Magnini (2015), according to which online review responses should follow the same rules as face-to-face service recovery. Hence, two additional strategies arose, personalization and empathy. Since they were observed independently and all strategies do not exclude each other, this conceptualization helps researchers to develop further models when researching managers’ responses to online complaints.
The paper provides empirical evidence on the influence of these five strategies on the trust of potential guests reading these reviews on the Internet. Studies to date have mostly observed managerial responses by confronting two strategies, whereas this research provides a comparison of five main strategies derived from the literature. Such an approach can help scholars to better understand the positive or negative implications of each strategy by observing their influences relative to each other.
Furthermore, including trust as a dependent variable also presents a novelty of this research. Concerning the influence of response content on potential guests’ trust, previous studies have only observed applying response versus no response strategy (Sparks, So and Bradley 2016) or using generic versus personalized responses (Wei, Miao and Huang 2013). Moreover, supporting all five hypotheses in this research has revealed a causal relationship between the response strategies and the trust of potential guests. Since trust is a key concept in customer relationship management, and research has revealed that it is influenced by different response strategies, this could represent a basis for further model development.
4.2. Practical implications
The research has revealed how different types of answers to the same service failure can generate different perceptions of trust. These findings emphasize the importance of careful tailoring the response to each review. AsWen et al. (2020) state, online reviews are an imperative in today’s digital age, and thus should be treated as top priority in hotel marketing. Hotel managers should be educated on how to semantically tailor the response to each type of complaint, or they should engage specialized employees who will deal with online reviews. As stated earlier, both responses and reviews are accessible to everyone, and the improper reaction of management can harm a hotel’s reputation.
To gain readers’ trust, results suggest that responses to complaints should be personalized, tailored individually to each customer, and referring specifically to their problem. When responding to complaints, managers should show sincere care and understanding for the customer’s problems because providing an empathic response will lead to the greater trust of potential guests. Furthermore, hotel managers are advised to use accommodative rather than defensive strategies. More precisely, they should use the confession strategy, whereby they will recognize the problem, accept responsibility and, finally, apologize for the situation. On the other hand, research shows that managers should not deny the existence of a problem or argue with customers because this could give the impression that they do not care about their customers but only about their company. Similarly, for some managers it may sometimes seem practical to avoid taking responsibility for the problem, and they will opt to shift the blame to someone or something else; however, findings suggest that this strategy also does not instil confidence in the hotel.
Despite the clear recommendations deriving from the research, it should be noted that there is no single solution to every negative review. In our research, the complaint was the result of a perceived service failure, which represents the most common scenario; however, hotel managers face a variety of online complaints every day, some of which may be unjustified or even fake. Therefore, confession, denial, and excuse in responses should always be used with caution, taking into consideration the type of review. As trust is the starting point of building a customer relationship, and is related to booking intention, these results provide valuable information on how to react to a negative review to achieve the greater trust of potential guests.
4.3. Limitations and recommendations for future research
This study has certain limitations which could stimulate future research. It should be noted that this research uses a single scenario to test all response types. While on the one hand, this eases the comparability of strategies, on the other hand, it diminishes the generalizability of the results. The managers’ responses, as well as the potential customers’ reactions to these responses, may differ depending on the type of failure that has occurred (Li, Cui and Peng 2018;Abramova et al. 2015). This research involves a severe negative review concerning a perceived service failure; however, hotels deal with many kinds of complaints that may or may not be justified. Therefore, further research needs to examine more deeply the situational context to determine in which situation which strategy would be the most appropriate. Upcoming studies may distinguish reviews depending on the type, objectivity, or controllability of the reason of complaint. In future research, contextual factors such as hotel class or even respondents’ characteristics could also be considered.
Since these strategies have been understudied up to date, this research has observed each strategy independently to get a clearer picture of its influence on trust and provide a basis for future research. However, not all strategies are mutually exclusive. Thus, another stimulative direction for further studies is to examine the combination of proposed strategies. By combining elements of two or three strategies, new ones would emerge, and their influence on trust or booking intention could be tested. Hence, examining the proposed types of strategies in different settings or different combinations could be a rich area for future research.