Examining the Impact of eWOM-Triggered Customer-to-Customer Interactions on Travelers’ Repurchase and Social Media Engagement

Electronic word-of-mouth (eWOM) communication on social media has revolutionized how travelers search for and share information and how they interact with one another digitally. This research examines the effects of eWOM-triggered customer-to-customer (C2C) interactions on travelers’ post-eWOM behaviors (i.e., repurchase and customer engagement) in a cross-cultural context. Drawing upon cognitive dissonance theory, a scenario-based experiment was conducted using a sample of 461 African tourists with recent intracontinental travel experience. Our findings suggest that a customer’s repurchase intention and engagement in social media C2C interactions are significantly influenced when their eWOM is challenged by other customers. Compared with individualistic cultures, such a phenomenon is more effective in collectivistic cultures, particularly when a customer shares negative eWOM. Customers in collectivistic cultures are more likely to appreciate consensus with other customers, and they tend to expend more effort toward solving dissonance. The theoretical and managerial implications of these findings are discussed.


Introduction
In today's digital age, a variety of electronic word-of-mouth (eWOM henceforth) media outlets support the sharing of consumers' opinions about their service experiences and interactions. In the hotel sector, eWOM plays an essential role in shaping travelers' behaviors. Per recent reports, 94% of travelers state that eWOM information is important for choosing their destination and accommodation, and 80% of consumers would share their hotel experiences if asked to do so (Condor Ferries 2020; TripAdvisor 2020). Over the last decade, eWOM has quickly evolved from a "nice-to-have" to a "must have" component of customer experience . Moreover, ever-growing social media platforms enable customers' eWOM engagement to go beyond simple eWOM-sharing and -receiving behaviors to include opportunities for interacting with different parties through eWOM communication (e.g., service providers and peer customers) (Narangajavana Kaosiri et al. 2019;Yakhlef and Nordin 2021). Such eWOM-triggered interactions allow customers and businesses to co-create an online persuasive environment by providing clarity and addressing problems through online conversations (Sijoria, Mukherjee, and Datta 2019). Even in the post-purchase stage, customer-to-customer (C2C henceforth) eWOM communication significantly impacts customers' post-purchase evaluation and behaviors (Antón, Camarero, and Garrido 2019;Chen et al. 2018). Furthermore, as digitalization progresses, eWOM has been observed not only in developed economies but also in emerging markets. Specifically, the latest research suggests that consumers increasingly engage in eWOM communication in African countries, while eWOM plays an irreplaceable role in African travelers' hotel consumption (Ledikwe, Stiehler-Mulder, and Roberts-Lombard 2020;Ukpabi et al. 2018). Against such background, this research focuses on two questions: how does eWOM-triggered C2C interactions affect consumers' post-eWOM behaviors? And given the diverse cultures of Africa, how does the effects of eWOM-triggered C2C interactions differ across cultures? To address these research questions, this research examines the effects of eWOM-triggered customer-to-customer (C2C) interactions on travelers' post-eWOM behaviors (i.e., repurchase and customer engagement) in a cross-cultural context. Our inquiry is timely and important for several reasons.
First, although consumers' eWOM activities in the hotel sector have permeated various online media, most existing research focuses on review sites. Hennig-Thurau, Wiertz, and Feldhaus (2015) divided eWOM media into two different types: online opinion sites (e.g., TripAdvisor) and social media platforms (e.g., Facebook). The two media types are significantly different in shaping customers' eWOM experiences in terms of anonymity, synchronicity, and interactivity (Liu et al. 2021a;You and Joshi 2020). However, the extant research has failed to recognize the differences between the types of eWOM media; thus, social media eWOM has attracted limited attention (Bigné, William, and Soria-Olivas 2020;Mulvey, Lever, and Elliot 2020). Meanwhile, given the concerns about fake reviews on opinion sites in recent years, social media sites now play increasingly important roles in travelers' eWOM communication (Aghakhani, Karimi, and Salehan 2018;Digital Marketing Institute 2020). Due to the significance of media differences and unique dynamics in eWOM communication on social media (e.g., Assaker and O'Connor 2020;Marder et al. 2019), it is important to advance our understanding of the eWOM environment by focusing on social media sites.
Additionally, customers' eWOM engagement has evolved from one direction (e.g., eWOM-giving and -receiving) into interactive conversations between customers and businesses and/or other customers . Increasing academic research and industry practices have captured eWOMtriggered business-to-customer (B2C) interactions in which hotel service providers' responses to customers' eWOM have significantly affected their post-eWOM evaluations and behaviors (e.g., Bhandari, Rodgers, and Pan 2021;Li, Cui, and Peng 2017). However, eWOM-triggered C2C interactions remain underexplored (Azer and Alexander 2020; Lee and Lee 2017) because they are a media-specific process. While social media sites encourage both B2C and C2C interactions, most online opinion sites only support B2C communication (Kanje et al. 2020;Wang and Chaudhry 2018). Aforementioned, as the majority of eWOM research in the hotel sector focused on online opinion sites (e.g., TripAdvisor and Yelp; see Bigné, William, and Soria-Olivas 2020;Dedeoğlu 2019;Shin, Perdue, and Pandelaere 2020 ), the C2C interactions on social media platforms (e.g., Facebook and Twitter) have been largely neglected (Bigné, William, and Soria-Olivas 2020;Liu et al. 2019). The eWOM-triggered C2C interactions that transpire on social media are important for several reasons. For example, eWOM received from other customers may influence one's evaluation and behavioral responses toward a service, even in the post-purchase stage (Hess and Ring 2016). Therefore, eWOM-triggered C2C interactions represent a unique touchpoint of eWOM information (i.e., peer customers' responses to focal customers' eWOM-giving) and advance our understanding of eWOM engagement throughout the customer journey. In such interactions, customers may not have united views on a hotel's service, but the impact of the view (in) consistency on the focal customer remains unexplored (Wu et al. 2016). This is important because, even in the post-purchase/post-eWOM stage, a focal customer's evaluation and behaviors toward the hotel may vary. It is essential to understand how different responses from other customers affect the focal customer's post-eWOM behaviors (Dedeoğlu 2019;Filieri and McLeay 2014;Liu et al. 2020). Furthermore, from the managerial perspective, service providers have limited control over C2C interactions on social media (Dineva et al. 2020) but understanding the impact of those interactions can help them develop appropriate intervention strategies.
Finally, between 2000 and 2021, the number of Internet users in Africa grew 130 times, which is nearly 150 times faster than the rest of the world (Internet World Stats 2021). Meanwhile, nearly 10% of global Facebook users are in Africa. Despite ever-developing digitalization, academic research has only captured a limited view of the digital transition of consumer behaviors in African economies. The hotel sector in Africa has been growing quickly and paying increasing attention to eWOM (Ledikwe, Stiehler-Mulder, and Roberts-Lombard 2020;Ukpabi et al. 2018). Prior research has noted that customers' eWOM communication varies worldwide (Bigné, William, and Soria-Olivas 2020;Park and Jeon 2018). Therefore, understanding African customers' eWOM behaviors has both theoretical and managerial implications. The growth of its hotel sector is due to the important role that tourism plays in many African economies (Adeola and Evans 2019;Rogerson 2007). Although tourism development varies across different countries, it has become a powerful economic tool for Africa (Christie et al. 2014;Signé and Johnson 2018). Specifically, intracontinental tourism has become an irreplaceable component of the African tourism economy, particularly in the upcoming post-COVID age (Florio 2020;United Nations 2017). However, existing research has often assumed a homogeneity of African countries and ignored the impact of culture on consumer behaviors (Binns, Dixon, and Nel 2012;Ukpabi et al. 2018). Africa comprises diverse cultures, and travelers' tourism and hotel consumption are heavily influenced by their cultural backgrounds (Binns, Dixon, and Nel 2012;Ukpabi et al. 2018). Given the growing popularity of eWOM communication and the significance of cultural diversity in Africa, this research focuses on African intracontinental travelers' eWOM engagement to contribute important empirical evidence of eWOM in African economies and to highlight the impact of cross-cultural nuances on consumers' eWOM engagement.
In summary, to fill in the important theoretical and practical gaps identified above, this research examines the effects of eWOM-triggered C2C interactions on travelers' post-eWOM evaluations of a hotel and their behavioral engagement with a social media platform from an African cross-cultural perspective. Next, we explain the theoretical foundations and concepts under consideration. We then develop the hypotheses, elucidate the research method, and present the data analysis results. Finally, we discuss the implications and provide suggestions for future research.

eWOM-Triggered C2C Interactions and Post-eWOM Evaluations
Early eWOM research argues that customers engage in eWOM through pre-purchase eWOM-seeking and post-purchase eWOM-giving behaviors (Toder-Alon, Brunel, and Fournier 2014). While mobile technology dominates the consumer market, it is easier than ever for consumers to access an abundance of consumption-related information at any point in the customer journey (Rosario, de Valck and Sotgiu 2020). Customers' eWOM engagement behaviors have become increasingly diverse. More specifically, with interactive technology, eWOM-giving in the post-purchase stage is no longer the end of the eWOM cycle (Ngarmwongnoi et al. 2020). After one shares eWOM on a service provider's social media page, other peer customers can interact with the focal customer by replying to his/her eWOM post to share their thoughts and experiences ). According to Liu et al. (2020), in the post-purchase/post-eWOM stage, where the consumer has gained personal experience with a service, eWOM information about the same service received from other customers still influences the focal customer's post-purchase evaluations and behaviors. Meanwhile, existing research has found that hotel service providers' responses to customers' eWOM influence those customers' repurchase of the service and eWOM media engagement (Wei, Miao, and Huang 2013;. Therefore, we expect that other customers' responses to the focal customer's eWOM would also influence his/her repurchase and social media engagement. In the post-purchase stage, the consistency/inconsistency between customers' experience and eWOM information shapes their behaviors (Azer and Alexander 2020;Liu et al. 2020;Wu et al. 2016). With eWOM-triggered C2C interactions, another customer's support/critique of the focal customer's eWOM could be seen as consistency/inconsistency between the two parties' opinions. Therefore, supportive/ critical responses from other customers could influence the focal customer's post-eWOM evaluation and behaviors. Such effects could be explained by the theory of cognitive dissonance.

Theory of Cognitive Dissonance
The theory of cognitive dissonance posits that human beings strive for internal psychological consistency (Festinger 1957). The fundamental proposition is that cognitive dissonance occurs when conflicting attitudes, beliefs, or values produce mental discomfort and psychological stress (Festinger 1957(Festinger , 1962. Therefore, a person who experiences cognitive dissonance is motivated to restore balance and solve contractions. Festinger (1957) argued that people may engage in two coping mechanisms with cognitive dissonance: they may either make additional effort toward balancing their contradictory thoughts and readjusting their beliefs or blindly believe whatever they want to believe to eliminate the need for additional effort. Notably, the choice of coping mechanism is often a context-specific decision (Festinger 1962;Geschwender 1967).
Cognitive dissonance effects have often been observed in consumer behaviors throughout the decision-making process (Bawa and Kansal 2008;Cohen and Goldberg 1970;Kassarjian and Cohen 1965), when consumers continuously absorb information. Thus, cognitive dissonance occurs when a customer receives inconsistent information from two or more sources, such as advertising (Bawa and Kansal 2008), newspapers (Mao and Oppewal 2010), prior experience (Li and Murphy 2013), and social influence (Tanford and Montgomery 2015), especially those arising from consumer eWOM engagement (Azer and Alexander 2020;Wei, Miao, and Huang 2013). eWOM is an important information source in the digital age, and any (in)consistency between eWOM and other information or between different types of eWOM shapes customers' perceptions and behaviors. As stated previously, dissonance in eWOM/WOM often stems from inconsistencies between personal beliefs and eWOM/WOM information or between different information cues. Table 1 summarizes previous eWOM studies that focus on the impact of information (in)consistency on consumers' cognition and behaviors. As shown in Table 1, most previous eWOM research on information (in)consistency focuses on the prepurchase stage and explores the effects of eWOM (in)consistency on customers' purchase decisions (e.g., Bigne et al. 2021;Chakravarty, Liu, and Mazumdar 2010;Tanford and Montgomery 2015). The studies that emphasized the postpurchase stage examined how (in)consistencies between personal experience and eWOM received in the post-purchase phase affect customers' eWOM engagement (Liu et al. 2020;Wu et al. 2016). Furthermore, Shin, Perdue, and Pandelaere (2020) demonstrated how hotel response personalization moderates the influence of opinion inconsistencies on consumer empowerment. The authors pointed out that previous research, including theirs, examined customer-to-business interaction and called for an understanding of interactive engagement in C2C relationships.
Thus, the current research goes beyond extant studies that examined eWOM (in)consistency (e.g., Book, Tanford, and  Pre-purchase In negative review scenarios, price reductions were not effective in offsetting the effect of unanimity. In contrast, higher prices enhanced willingness to pay in positive review scenarios. Shin, Perdue, and Pandelaere (2020) Empowerment theory Online reviews versus hotel response personalization Post-purchase but pre-eWOM For both repeat and prospective hotel guests, hotel response personalization triggers more empowerment when the review is negative compared with positive review scenarios.
Dual-process theory Multiple individual reviews Pre-purchase Consistency of multiple reviews is perceived as more helpful than any inconsistency, regardless of whether the consistency is positive or negative. Wu et al. (2016) Self-enhancement theory Online reviews versus personal experience Post-purchase but pre-eWOM Powerless consumers are more likely to post positive reviews when the forum consensus is positive. By contrast, powerful consumers are more likely to post positive reviews when the aggregated consensus is negative. Such effects only hold for positive (vs. negative) service experiences.
Chen 2016; Shin, Perdue, and Pandelaere 2020; Wu et al. 2016) by focusing on C2C interactions in the post-eWOM stage and examining how a focal customer's behaviors are influenced by other customers' responses to his/her eWOMgiving. We posit that cognitive dissonance occurs when other customers challenge/criticize (vs. support) eWOM shared by the focal customer. Such dissonance drives the focal customer to reevaluate inconsistent information cues and reshape his/her behaviors. Considering the nature of social media eWOM in the hotel sector, this study specifically investigates the impact of cognitive dissonance on focal customer's repurchase intention and social media engagement with the hotel service provider. Our approach is expected to advance the theoretical understanding of cognitive dissonance effects in eWOM research.

Cognitive Dissonance Effects in the Cross-cultural Context
Another focus of this research is the cross-cultural perspective within Africa. Pioneering research found that cognitive dissonance effects vary across different cultural backgrounds (Hiniker 1969;Hoshino-Browne et al. 2005). However, previous research that applied cognitive dissonance theory to understand travel purchases in the eWOM context (e.g., Tanford and Montgomery 2015) did not consider cultural influences. According to Mariani, Borghi, and Okumus (2020), insufficient attention has been paid to the effect of cultural differences in the online reviews' context. Meanwhile, Triandis (2001) suggested that among Hofstede's (2011) fundamental cultural dimensions (i.e., power distance, collectivism vs. individualism, femininity vs. masculinity, uncertainty avoidance, indulgence vs. restraints, and long-term vs. shortterm orientation), cognitive dissonance effects are likely to vary according to the collectivism/individualism dimension because people from individualistic and collectivist cultures view inconsistencies differently and would thus also respond to dissonance differently (Kastenmüller et al. 2010). The collectivism/individualism dimension is also arguably the most important dimension for examining cultural disparities in international advertising (Diehl, Terlutter, and Weinberg 2003), and it remains the most important dimension of selfconstrual social behavior (Triandis 2002). In addition, the way in which people construct their identities in social settings is determined by both their personal and social identities. Social media enables eWOM and interactive experiences to transpire among users with strong social affinities (Pan and Chiou 2011;Leventhal, Hollebeek, and Chen 2014). Thus, collectivism/individualism are important cultural elements for examining social identity construction in the social media context. Compared with collectivists, people in individualistic societies are less mindful of inconsistencies and less willing to restore the resulting dissonance (Hoshino-Browne 2012). Therefore, customers' dissonance coping strategies could differ in individualistic and collectivist cultures. Moreover, despite the effects on individuals' responses to dissonance, collectivism/individualism may also shape people's reactions to cohesiveness. More specifically, people in collectivist cultures appreciate consistency in ideas, beliefs, and values between themselves and others more than people in individualist cultures (Petrova, Cialdini, and Sills 2007). Through the social identity perspective, such effects result from collectivists' preference for shared opinions and placement of greater value on consensus within the social group (Kitayama and Imada 2008). Conversely, individualists prefer to distinguish themselves from others, and they place less value on consistent opinions with others. Additionally, the extant literature reveals that collectivism/ individualism significantly influences customers' processing and response to eWOM information (Choi and Kim 2019;Filieri and Mariani 2021;Luo et al. 2014). Filieri and Mariani (2021) examined how cultural values determine consumer's evaluation of review helpfulness. They found that consumers from individualistic cultures write more helpful reviews. This is because collectivists strive for opinion consensus and avoid writing reviews that deviate from the opinions expressed by the majority. By contrast, individualistic consumers are more inclined not to conform to others' opinions because they express their personal opinions and emotions which may contradict the reviews written by the majority. Drawing on these arguments, we expect that when other customers respond to a focal customer's eWOM to support or criticize his/her views, the focal customer's behavioral responses to other customers' support and critiques would differ according to his/her cultural backgrounds. Given African countries' diverse cultures, including their distinctiveness in terms of collectivism/individualism (Binns, Dixon, and Nel 2012; Izogo, Mpinganjira and Ogba 2020), we take an intracontinental cross-cultural perspective to further highlight the nuanced differences of eWOM-triggered C2C interactions between collectivistic and individualistic cultures. Our unique cultural angle contributes to developing crosscultural insights into cognitive dissonance and advancing the understanding of customers' social media eWOM engagement.
Drawing upon the theory of cognitive dissonance and considering the cultural traits between collectivism and individualism, this research specifically examines the impact of eWOM-triggered customers' responses (i.e., supportive vs. critical) on focal customers' post-eWOM repurchase intention and social media engagement with a hotel. We specifically focus on an African perspective, given Africa's cultural diversity and the boom of eWOM practices in the African hotel sector. The constructs and phenomena under consideration in this study and the rationale behind the linkages between the individual concepts are outlined in Figure 1.

eWOM-Triggered C2C Interactions and Customer Engagement
Social media as eWOM platforms (e.g., Facebook) not only allow travelers to share their own consumption experience but also provide dynamic online environments in which they can interact with peer customers and the consumed hospitality brand in the post-eWOM stage (Liu et al. 2021a). When hospitality service providers make consistent efforts to develop an interactive community and enhance customer engagement on social media, they build and maintain customer loyalty (So et al. 2016). Customer engagement refers to "the intensity of an individual's participation in and connection with an organization's offerings or organizational activities, which either the customer or the organization initiates" (Vivek, Beatty, and Morgan 2012, 133). In the tourism and hospitality context, customer engagement with a business occurs in both online and offline settings (Taheri, Jafari, and O'Gorman 2014;So, Wei, and Martin 2020). Considering the digital nature of eWOM, we focus on online customer engagement in this study (Kanje et al. 2020). Customer engagement on social media can be categorized into active and passive engagement (Dijkmans, Kerkhof, and Beukeboom 2015). More precisely, active engagement is behavioral and involves activities, such as writing online reviews, WOM, and helping other customers, while passive engagement involves minimal investment and merely observing online brand communications (Kunz et al. 2017). Customer engagement types can either be negatively-or positively-valenced (Leventhal, Hollebeek, and Chen 2014). Existing research has revealed that customer engagement could potentially be influenced by the customer's personal experience with the hospitality brand (i.e., positive vs. negative) (e.g., Azer and Alexander 2020; Bitter and Grabner-Kräuter 2016; Pan and Chiou 2011) and peer customers' thoughts about the hospitality brand (e.g., opinion consistency vs. opinion inconsistency) (Chang and Wu 2014). Notably, the valence of eWOM refers to the positive or negative nature of the eWOM message and is drawn from the valence of the consumption experience (Liu et al. 2020;Tsao et al. 2015). Thus, by drawing upon the theory of cognitive dissonance and the specific focus on eWOM-elicited interactions between customers, we postulate that when one shares eWOM about a hotel on social media, his/her customer engagement with the hotel is influenced by the valence of his/her eWOM-sharing (i.e., positive vs. negative) and the responses received from other customers (i.e., supportive vs. critical).
Social media supports consumer interactions and engagement (Filieri and McLeay 2014) because it provides an avenue for consumers to engage with a firm, exchange information, express themselves, and create value (Brodie et al. 2013;Ek Styvén, Mariani, and Strandberg 2020;Mudambi and Schuff 2010). In the context of eWOM-giving, shared experiences can be positive or negative. Consumers engage with social media and online communications because social media is a hedonic system, meaning that its content is pleasurable for its own sake (Amaro, Duarte, and Henriques 2016;van der Heijden 2004). In online communication, regardless of the opinion valence (i.e., positive and negative), being supported by others and achieving opinion consensus with others enhances one's pleasurable experience, which facilitates further engagement on social media (Liu et al. 2020;Wei, Miao, and Huang 2013;Wu et al. 2016). However, while reaching a negative consensus with others facilitates the focal customer's self-enhancement by reinforcing his/her opinion, it may also minimize his/her interest in the service provider, resulting in low engagement (Kim 2011). Conversely, sharing positive eWOM is more likely to elicit further customer engagement (Eigenraam et al. 2018). Thus, compared with negative eWOM, positive eWOM, when supported by a peer customer, induces greater pleasure and facilitates continued customer engagement. Therefore, we hypothesize the following: H1a. eWOM-triggered C2C interactions influence the focal customer's engagement. Specifically, compared with positive eWOM, when negative eWOM is supported by another customer, the focal customer will show i) lower passive engagement and ii) lower active engagement.
Through the theoretical lens of cognitive dissonance, when there are opinion inconsistencies between customers, a customer is more likely to engage in collecting more information  and interacting further with the brand/peer customers (i.e., active and passive engagement) to address the inconsistencies between his/her own beliefs and those of others. Opinions that are inconsistent with one's existing beliefs result either in an asymmetrical quantity of cognitive processing or altered processing aimed at solving the dissonance (Wei, Miao, and Huang 2013). Such effects are more significant when the focal customer's experience is positive. This is because positive experience triggers customers' interests in the service firm, while customers in general are more likely to spend additional effort on products or services that they are interested in (Siegrist, Stampfli, and Kastenholz 2009;Werth and Foerster 2007). However, if the focal customer's experience is negative, he/she is more likely to ignore the conflicting thoughts from other customers to solve the dissonance. As would be expected, positive experiences are more likely to endear customers to a firm than negative experiences (Serra-Cantallops, Ramon-Cardona, and Salvi 2018). Thus, we hypothesize the following: H1b. eWOM-triggered C2C interactions influence the focal customer's engagement. Specifically, compared with negative eWOM, when positive eWOM is criticized by another customer, the focal customer will show i) higher passive engagement and ii) higher active engagement.
According to cognitive dissonance theory, when one's belief is challenged, cognitive dissonance occurs, which triggers more cognitive evaluations that are aimed at understanding the challenge. Consumers are likely to give a more positive evaluation when exposed to information that is evaluatively consistent versus evaluatively inconsistent (Ek Styvén, Mariani, and Strandberg 2020;Sen and Lerman 2007). Accordingly, Liu et al. (2020) have suggested that positive opinion consistency strengthens consumers' confidence in the hotel and facilitates their further engagement. Conversely, in the post-purchase phase of eWOM-giving, the eWOM-giver has experienced the product or service. If people have already developed a positive attitude toward a product or service, exposure to contradictory information can diverge their positive attitude. However, compared with negative beliefs, consumers tend to jettison contrasting information when their initial evaluation is positive, which results in a lower level of behavioral engagement (Lord, Ross, and Lepper 1979). In the context of travel purchases influenced by online reviews, Tanford and Montgomery (2015) argued that cognitive dissonance can be decreased by changing one of the inconsistent elements such as disregarding other customers' reviews when one has had a positive experience, thereby lowering consumers' interests in the service provider and level of engagement. Following this reasoning, a customer who shares a positive travel experience on social media will perceive a supportive peer customer response on that post as reinforcing his/her positive experience, while a critical peer customer response may dissuade him/her from further dealings with the firm. Therefore, we expect reinforcement of a positive evaluation of a brand to have an overwhelming impact on customer engagement with that brand, and the following is hypothesized: H1c. eWOM-triggered C2C interactions influence the focal customer's engagement. Specifically, compared with positive eWOM supported by other customers, when positive eWOM is criticized by other customers, the focal customer will show i) lower passive engagement and ii) lower active engagement.
Furthermore, when a customer has a negative experience with a hotel and shares negative eWOM about the hotel on social media, how peer customers respond to the negative eWOM affects the focal customer's engagement with the brand and peer customers . Although a negative opinion consensus may allow the customer to achieve self-enhancement, it also minimizes the customer's interest in the hotel (Liu et al. 2020). By contrast, when there is a critical response to a customer's negative eWOM, the focal customer is more likely to seek further information and reevaluate their opinion, and there are two potential reasons for this. First, from the perspective of cognitive dissonance, when the focal customer's negative eWOM is challenged, such opinion inconsistency arouses the focal customer more than a negative opinion consensus would (Chen 2018;Festinger 1957). Their additional interest is drawn from the possibility that the negative experience may have been a special case. Therefore, by engaging further with the hotel and other customers on social media, the focal customer expects to learn whether the negative experience was a special case and thereby understand the dissonance caused by the opinion inconsistency. Second, when a customer's negative view is challenged, his/her self-esteem may be challenged. Searching for additional information and interacting further with the brand and customers is a process for regaining self-esteem (Vostanis et al. 1996). Further engagement not only allows the customer to address the dissonance caused by opinion inconsistencies but also helps restore self-esteem. A piece of information that appears more frequently (e.g., same valenced information) triggers less cognitive processing because less effort is required to process it compared with when opinions are inconsistent (Block and Keller 1995;Kim 2011). Therefore, less effort is required to process congruent opinions, leading to less cognitive evaluations and resulting in fewer further actions compared with divergent views that trigger higher elaboration (Ahluwalia 2002;Shen and Dillard 2009). Thus, we hypothesize the following: H1d. eWOM-triggered C2C interactions influence the focal customer's engagement. Specifically, compared with negative eWOM supported by another customer, when negative eWOM is criticized by another customer, the focal customer will show i) higher passive engagement and ii) higher active engagement.

eWOM-Triggered C2C Interactions and Repurchase Intention
In the post-purchase phase, repurchase intention is an essential customer behavioral response because it captures a customer's judgment about rebuying a designated service from the same company based on the current situation and likely circumstances (Hellier et al. 2003). Previous research has shown that repurchase intention in the hospitality sector relies on service satisfaction and service quality (Hellier et al. 2003;Hsu et al. 2014;Shin et al. 2013), which attenuates the consumer's perceived risk of making a purchase (Azer and Alexander 2020;Fang et al. 2016;Herjanto and Amin 2020). Pioneering research has also suggested that interactions with other parties in the post-purchase stage may influence customers' repurchase intentions (Fang et al. 2016;Herjanto and Amin 2020) and that eWOM shared by consumers who have previously experienced a service can be used to influence consumers' product evaluations (Sen and Lerman 2007). Additionally, Wei, Miao, and Huang (2013) demonstrated that firms' responses to eWOM enhance consumer trust and perceived communication quality evaluations. In this study, we postulate that a customer's repurchase intention could also be influenced by other customers' responses to his/her eWOM.
When a customer shares his/her hotel experience through social media eWOM, other customers' support/criticism either reinforces or undermines the customer's existing beliefs about the hotel. According to Wei, Miao, and Huang (2013), the presumptions that consumers have about hotels can bias their perception of eWOM shared by peer consumers such that eWOM that is inconsistent with existing beliefs is distorted to attenuate one's dissonance. Such reevaluation of the hotel weakens the focal customer's repurchase intention. The differences in repurchase intention resulting from supportive and critical responses can be explained by the reinforcement effects-a supplementary view of cognitive dissonance theory (Johnson-Cartee and Copeland 1997; Surlin and Gordon 1976). Reinforcement effects focus on the consistency and consensus between different parities and posit that information that is consistent with a customer's preexisting beliefs reinforces such beliefs and drives the customer's behaviors (Surlin and Gordon 1976). Therefore, when a customer's eWOM about a service brand receives supportive responses from other customers, his/her original evaluation of the brand is reinforced and drives further behavioral intention (e.g., repurchase intention). By contrast, when a customer's eWOM is challenged by other customers, cognitive dissonance occurs. Critical responses weaken the focal customer's original beliefs and neutralize their evaluations and behavioral tendencies. More precisely, positive experiences and eWOM sharing often lead to higher repurchase intention (Williams and Soutar 2009). Supportive comments from peer customers will reinforce the focal customer's positive evaluation of the hotel and further increase their repurchase intention. Critical responses neutralize the focal customer's positive evaluation and lower their repurchase intention. Similarly, supportive responses to negative eWOM validate the focal customer's negative evaluation and further lower their repurchase intention, while critical responses to negative eWOM result in higher repurchase intention as an outcome of the focal customer's coping with cognitive dissonance. Thus, we hypothesize the following: H2. eWOM-triggered C2C interactions influence the focal customer's repurchase intention. H2a. Compared with positive eWOM supported by another customer, when positive eWOM is criticized by another customer, the focal customer will show lower repurchase intention. H2b. Compared with negative eWOM supported by another customer, when negative eWOM is criticized by another customer, the focal customer will show higher repurchase intention.

Cross-Cultural Differences: Collectivism Versus Individualism
Differences in cultural orientation, especially in terms of collectivism/individualism, will influence how dissonance occurs. This is because, in travel research, cultural backgrounds shape behavioral adjustments and affect the acceptance of alien cultural tenets (Gao et al. 2018). The effects of cognitive dissonance theory vary across cultures, particularly in terms of the collectivism/individualism cultural dimensions (Hoshino-Browne et al. 2005;Triandis 2001). While individualistic consumers believe more in the self as an independent entity, collectivistic consumers are more oriented toward the group (Frank, Enkawa, and Schvaneveldt 2015). It has equally been argued that consumers from collectivistic cultures tend to subscribe to opinion conformity while their individualistic counterparts are more likely to hold inconsistent views in their reviews compared with the views held by other reviewers (Filieri and Mariani 2021). The perception of collectivistic consumers is governed by interpersonal treatment because they tend to gravitate toward obedience, and they desire group recognition and interaction (Hofstede 2001;Patterson and Mattila 2008). Conversely, individualistic consumers value self-interest over group interest and are less likely to be influenced by marketing stimuli (Frank, Enkawa, and Schvaneveldt 2015;Triandis 2002). In the tourism context, tourists from collectivistic cultures (e.g., Japan) are more likely to seek others' recognition than those from countries that are more individualistic (e.g., the U.S.) (Mariani, Borghi, and Okumus 2020). Social media platforms, such as Facebook, afford individuals the opportunity to share interactive experiences with like-minded people (Pan and Chiou 2011); therefore, collectivists are more likely to be influenced by information received on social media than are individualists. More importantly, prior research on eWOM has highlighted the impact of collectivism (vs. individualism) on customers' eWOM engagement (e.g., Kitirattarkarn, Araujo, and Neijens 2019;Levy and Gvili 2020;Luo et al. 2014). Thus, we specifically examine the impact of collectivism/individualism on customers' processing of others' responses to their eWOM sharing.
Additionally, the collectivism/individualism cultural dimension often shapes individuals' social identity construction because collectivistic and individualistic consumers will vary in how they construct meanings and beliefs (Hogg and Vaughan 2002;Tajfel and Turner 1979). Specifically, given that collectivistic consumers are more receptive to other people's views and opinions (Frank, Enkawa, and Schvaneveldt 2015;Hofstede 2001;Patterson and Mattila 2008), they tend to appreciate and value shared beliefs (e.g., supportive responses from other customers) more than individualistic consumers (Petrova, Cialdini, and Sills 2007). Thus, when there is an opinion inconsistency on social media (e.g., critical responses from other customers), collectivistic consumers will make an additional effort to understand the inconsistency via further engagement with the brand and reevaluating products and services (e.g., customer engagement and repurchase) compared with individualistic consumers. This is because individualistic consumers are less sensitive to other people's views (Kim 2011). Individualists are not only less appreciative of the opinion consensus but are also less mindful of inconsistencies (Hoshino-Browne 2012). Previous cross-cultural research (see especially Han 2017 andKrishen andHu 2018) suggested that customers' perceptions and behaviors are shaped by the collectivism/ individualism cultural dimensions. In the eWOM research in the travel context, for example, Gao et al. (2018) found that culture influences online rating behaviors. Therefore, we expect that a customer's cultural orientation shapes the way he/she responds to supportive and critical comments from other customers to his/her eWOM. Specifically, opinion consistency and inconsistency between the focal customer and other customers have a stronger impact in collectivistic cultures compared with individualistic cultures. Thus, we propose the following hypotheses: H3a. In a collectivistic culture (vs. an individualistic culture), the supportive response to an online review from other customers will have a greater impact on the reviewer's i) passive engagement, ii) active engagement, and iii) repurchase intention. H3b. In a collectivistic culture (vs. an individualistic culture), the critical responses to an online review from other customers will have a greater impact on the reviewer's i) passive engagement, ii) active engagement, and iii) repurchase intention.

Design and Context
To test the research hypotheses, we devised a scenario-based experiment-a popular design in advertising (e.g., Zhang et al. 2014) and social media, especially Facebook-and a discussion forum in the eWOM literature (e.g., Lo and Yao 2019;Pan and Chiou 2011;Shin, Perdue, and Pandelaere 2020). Scenario-based experimental design overcomes recall bias, to which self-reporting techniques are prone, and aids proper manipulation of experimental variables (Smith, Bolton, and Wagner 1999).
We followed a 2 (eWOM-giving: positive vs. negative) × 2 (peer customer response: supportive vs. critical) betweensubject design. Consistent with previous studies that utilized nationality as a proxy for national culture (Hwang and Mattila 2018; Nath, Devlin, and Reid 2016; Song, Noone, and Mattila 2018), we selected South Africa and Ghana, as the most polar cases in terms of collectivism/individualism in Africa for stimuli presentation and data collection which lasted from January to March 2021. Facebook was selected as the eWOM platform because it is one of the most popular social media platforms in Africa; nearly 10% of Facebook users are based in Africa (Bitter and Grabner-Kräuter 2016; Guttmann 2019; Internet World Stats 2021). Additionally, Facebook facilitates C2C interactive experiences (Dedeoğlu et al. 2020) on a scale greater than that of other social media platforms, such as Twitter.

Stimuli
Consistent with previous manipulations of message statements (see Pan and Chiou 2011; Shin, Perdue, and Pandelaere 2020), we manipulated eWOM-giving into positive and negative and manipulated accompanying peer customer response into supportive and critical to realize four scenarios (see Supplemental Table 1). The participants who were randomly assigned to one of the four scenarios were asked to imagine that they went on a vacation to Marrakech, Morocco, with their friends and stayed in the Diamond Palace Hotel. They were further asked to imagine that they shared their hotelstaying experiences on Facebook; after that, a fellow Facebook user who had also stayed at the hotel responded to their Facebook post to support/criticize the participants' eWOM. We chose Marrakech, Morocco, because it was voted the most popular tourist destination in Africa and one of the 25 most popular tourist destinations in the world by TripAdvisor in 2020 (Tubei 2020). Consistent with recent research (Shin, Perdue, and Pandelaere 2020), a fictitious hotel name was used in the experimental scenarios to rule out biases that could arise from prior experience, attitude, or knowledge. After reading the scenarios, the participants completed the accompanying survey.

Measures
The survey included questions regarding manipulation checks, realism checks, and dependent variables. For manipulation checks, the participants were asked to rate how positive/negative their hotel stay experience and Facebook post were in the scenario (i.e., valence of eWOM-giving) (Liu et al. 2020) and to rate to what extent the other customer's response was critical/supportive of their own Facebook post (i.e., response type) (Ranaweera and Jayawardhena 2014). The participants were then asked to rate their feelings on experimental realism (how realistic the scenario was) and mundane realism (how likely it was that the described situation could happen in real life) of the depicted scenario (Liao 2007;Roschk and Kaiser 2013). Measures of the dependent variables were adapted from established scales for (1) passive engagement, (2) active engagement (Dijkmans, Kerkhof, and Beukeboom 2015), and (3) repurchase intention (Hellier et al. 2003;Wang, Cao, and Park 2019). Additionally, following previous research that documented the influence of credibility on social media information consumption motives of travelers (see Hur et al. 2017;Nath, Devlin, and Reid 2016), we deployed information credibility as a covariate and adapted its four-item measures from Freeman and Spyridakis (2004). The measures of all variables were internally consistent: passive engagement (α South Africa = .850; α Ghana = .858); active engagement (α South Africa = .904; α Ghana = .918); repurchase intention (α South Africa = .930; α Ghana = .936); and information credibility (α South Africa = .884; α Ghana = .765). Additionally, to ensure data validity, a series of comprehension and attention checking questions were inserted throughout the survey. The response was only valid if the participant correctly answered all the comprehension and attention checking questions. All measurement scales are shown in Supplemental Table 2.

Pretest
We conducted a pre-test to ensure the validity of the experimental design and survey. To ensure the content validity of the survey measures, we consulted four marketing professors experienced in eWOM research (Liu et al. 2021c). To ensure face validity, we conducted a pilot study with 14 postgraduate university students who were recruited using convenience sampling. The participants were invited to participate in the experiment online and make notes regarding potential issues and observations. They were then approached individually to provide feedback and make suggestions for improving the understandability and readability of the scenarios and the survey (Liu et al. 2021b). Based on their feedback, we made minor changes and finalized the experimental design and scenarios.

Sampling
We collected data concurrently in South Africa and Ghana through a professional marketing company. A sample of 461 participants was gathered from both countries (240 from South Africa and 221 from Ghana). This sample resonates with Iacobucci (1994), who stated that a sample of 150 subjects is adequate for a 2 × 2 factorial design. The experimental approach requires that participants have adequate knowledge of the task involved in the experiment (Nath, Devlin, and Reid 2016). Thus, data quality was ensured through two filter questions: recent intracontinental travel experience (yes/no) and Facebook use experience (5 = always to 1 = never). Participants who lacked travel experience to any other African country in the last three years or who did not use Facebook were excluded. Therefore, qualified participants were expected to relate themselves to the experimental scenarios.

Analysis and Results
To test the research hypotheses, a multivariate analysis of covariance (MANCOVA) was used. Consistent with previous research that: demographic factors motivate tourists of African descent differentially (Kara and Mkwizu 2020); social media usage varies across African countries (Varrella 2021); credibility influences eWOM (Dedeoğlu 2019;Hur et al. 2017), we initially considered demographic factors, such as gender, age, social media use, social media hotel experience sharing, and information credibility, as covariates. Only covariates that significantly influenced the results were included in the final analysis. Table 2 shows the characteristics of the two samples. Overall, 56.8% of the participants were male, 43% were between 21 and 29 years old, 47.5% had a bachelor's degree, 58.4% were employed and working full-time, and all had intracontinental tourism experience. Furthermore, on average, their social media usage was M = 4.527 on a 5-point scale.

Demographics
The distribution of the two samples was as follows: gender (Male South Africa = 48.8%; Male Ghana = 65.6%); age (the most represented age group in the South African sample was 30-39 years of age with 33.8%; the most represented age group in the Ghana sample was 20-29 years of age with 59.3%); education (bachelor's degree South Africa = 43.3%; bachelor's degree Ghana = 52%); employed and working fulltime (South Africa = 67.5%; Ghana = 48.4%). On average, social media usage was M South Africa = 4.454; M Ghana = 4.606 on a 5-point scale.

Hypotheses Testing
MANCOVA. Outputs from a MANCOVA with the covariates influenced the results significantly compared with a preliminary MANOVA outputs. Thus, the covariates were included in the analysis. The MANCOVA results and the descriptive means are shown in the Appendix and Table 3, respectively. The results show that the main effect of cultural orientation (collectivism vs. individualism) on passive engagement,   supported. Table 4 summarizes the results of hypotheses testing.

Discussion
Beyond the existing research of eWOM-seeking and -giving on online opinion sites, this study focuses on the interactivity of the social media sites and examines the impact of C2C interactions elicited by eWOM-giving on the focal customer's post-eWOM behaviors. Understanding these C2C interactions can help hospitality service providers gain better insight into how subsequent engagement actions and future purchase intention can be triggered. Drawing upon the theory of cognitive dissonance, this study demonstrates the effects of opinion (in)consistency in eWOM-triggered social media interactions on focal customers' post-eWOM repurchase intention and social media engagement with a hotel in collectivistic and individualistic cultures. Though the data upon which our analysis is based did not account for the impact of COVID-19 pandemic on social media C2C interactions in Africa, our findings provide novel insights into how tourism firms might manage eWOM-triggered C2C interactions that transpire on social media. Specifically, our findings suggest that the valence of a focal customer's eWOM-giving (i.e., positive vs. negative) and peer customers' responses (i.e., supportive vs. critical) influence the focal customer's engagement with the hotel. For example, we found that in eWOM-triggered C2C interactions, where customers commonly share positive opinions about a service provider, the focal customer is more likely to further engage with the service provider on social media compared with when there is a negative consensus between customers. Such findings suggest that positive opinion consistency drives customer engagement more than negative opinion consistency does. This reinforces that customers tend to seek pleasant experiences in the customer engagement process (Verhagen et al. 2015), while the positive consistency between their own and others' experiences brings more pleasantness and leads to further engagement (Wu et al. 2016).
Additionally, when there is an inconsistency in C2C eWOM interactions, the focal customer is more likely to further engage with the brand when his/her positive opinion (vs. negative opinion) is being challenged by other customers. This is because a customer with a positive service experience is more likely to develop loyalty to the service provider. However, when others challenge the focal customer's positive view of the service provider, the focal customer is more willing to engage in finding more information and support through brand engagement to understand the cognitive dissonance. By contrast, when other customers challenge the focal customer's negative eWOM of a service provider, the focal customer is more likely to cope with such dissonance by ignoring different opinions (Mao and Oppewal 2010). Customers who have negative experiences with a brand usually have a minimal interest in the brand and are less likely to make any additional effort to understand dissonance.
Our findings further suggest that a customer who shares positive eWOM about a brand tends to engage with the brand regardless of other customers' responses, while customers who share negative eWOM about a brand are more likely to engage with it when other customers criticize their view. Thus, customer engagement is voluntary when a customer has a positive experience with a service provider, but it is subject to other customers' responses when the focal customer has a negative experience and shares negative eWOM about the provider. The focal customer is motivated to make additional effort (e.g., engagement) to solve the dissonance because social media users are more cautious when sharing negative information (Suler 2004). When negative information is criticized by others, they engage further with the brand and learn more information in order to justify their negative eWOM shared.
Furthermore, our findings show that the valence of the focal customer's eWOM-giving and other customers' responses interactively influence the focal customer's repurchase intention. More precisely, the repurchase intention of customers who share positive eWOM is not influenced by other customers' responses. However, the repurchase intention of customers who share negative eWOM about the service provider is affected by other customers' responses. Specifically, when the focal customer's negative eWOM is challenged by other customers, his/her repurchase intention is higher than when other customers support his/her eWOM. This implies that customers' repurchase intentions rely on their own consumption experience when that experience is positive. However, customers who have a negative experience with a service provider tend to consider other customers' opposing opinions and are more likely to give the service provider another opportunity to provide good service. Finally, our findings also suggest that the effects of eWOM-triggered C2C interactions on social media engagement and repurchase intention vary between collectivistic and individualistic cultures. Interestingly, our results found no significant difference between these two cultures in terms of appreciating opinion consensus. Yet, when there is opinion inconsistency in an eWOM-triggered C2C interaction, collectivistic customers are more likely to readjust their post-eWOM behaviors to cope with the inconsistency compared with customers in individualistic cultures. Therefore, the cultural difference in opinion (in)consistency is rooted in 'inconsistency avoidance' rather than 'consistency preference'.

Theoretical Implications
This study makes several important theoretical contributions. First, we have pushed the boundaries of cognitive dissonance theory by identifying that dissonance coping is more likely to be activated and post-eWOM behaviors are more likely to be readjusted when a customer's negative beliefs encounter others' positive thoughts about the same brand. Conversely, customers with positive beliefs about a brand tend to stay true to their own experience and beliefs and are less likely to be influenced by other negative thoughts. This advances the theoretical understanding of dissonance effects in the post-eWOM stage. Previous eWOM research has examined the precursors and consequences of eWOM (see Filieri and McLeay 2014;Mariani, Mura, and Di Felice 2018;Wei, Miao, and Huang 2013). By revealing the inconsistency avoidance tendency in collectivistic cultures (vs. individualistic cultures), this research responds to urgent calls from prior research and contributes to a greater understanding of the cross-cultural nuances in cognitive dissonance (Kastenmüller et al. 2010;Triandis 2001).
Second, by illustrating often-neglected C2C interactions in eWOM communication, we have advanced existing understanding of customers' eWOM journeys and revealed the significance of peer customers in shaping focal customers' post-eWOM behaviors . eWOM research stresses the inimical repercussions of negative eWOM for tourism brands (Dijkmans, Kerkhof, and Beukeboom 2015). We have offered fresh insights to this research stream by demonstrating that the detrimental effect of negative eWOM is more pronounced when negative eWOM is challenged by a critical peer customer response than when it is consolidated by a supportive peer customer response. By focusing on such a unique touchpoint of eWOM on social media, we take eWOM research in the hotel sector beyond online opinion sites (i.e., review sites) and highlight the interactivity in social media eWOM. Thus, our research adds to the body of literature that stresses when managers should be the most concerned about negative eWOM (see Mudambi and Schuff 2010;Sen and Lerman 2007).
Third, this work contributes significant empirical evidence to understanding modern consumer behaviors in African countries in the age of digital transition (Ledikwe, Stiehler-Mulder, and Roberts-Lombard 2020;Ukpabi et al. 2018). Previous research has not comprehensively mapped how consumer intentions can be influenced through the interactions of C2C eWOM and the collectivism (vs. individualism) cultural orientation. Although extant research has illuminated how cultural differences influence eWOM enactment and evaluation (see Stamolampros et al. 2019), it has not provided insight into how the dissonance that emerges from eWOM-triggered C2C interactions influences intentions among collectivistic (vs. individualistic) tourists. Due to differences in cultural orientation, consumers are known to differ in how much they value self-interest over group interests, with individualistic consumers placing more emphasis on self-interest than on group interest (Frank, Enkawa, and Schvaneveldt 2015;Triandis 2002). Our unique cultural angle adds to the emerging cross-cultural insights of cognitive dissonance and advances the understanding of customers' social media eWOM engagement. Furthermore, the intracontinental cultural differences highlight the cultural diversity of Africa and reinforce the importance of culturespecific consumer behaviors and marketing practices (van Pinxteren 2020). Thus, the current study adds fresh insights to previous studies that limited the study of eWOM evaluation across cultures (Hwang and Mattila 2018; Shin, Perdue, and Pandelaere 2020).

Managerial Implications
Our study has several important managerial implications. First, managers of hospitality and travel firms, especially hoteliers targeting customers, must strive to support the development of a positive consensus in social media eWOMtriggered C2C interactions by closely monitoring social media and providing intervention when needed. This is because satisfied tourists are likely to provide constructive feedback or make helpful comments (Shin, Perdue, and Pandelaere 2020). A positive opinion consensus is important because it facilitates customer engagement and maintains interactivity in the social media community. Our findings also suggest that consumers who share positive experiences do not consider other peoples' views when determining their future engagement and repurchase behaviors. Consumers who share positive eWOM are more likely to be actively and positively engaged and to show an intention to buy more compared with those who do not. Thus, given that engaging in eWOM is voluntary for customers who have had positive experiences, hoteliers are encouraged to emulate e-retailers, such as Amazon, by requesting eWOM from satisfied tourists. With this knowledge, managers can better manage social media C2C brand-related interactions and improve customer engagement and repurchase behaviors in different geographic locations in Africa to fit the collectivistic and individualistic tendencies of tourists.
Also, while it is useful to ensure that consumers share positive eWOM based on positive customer experiences, some industry observations (e.g., Lo and Yao 2019) have indicated that firms or brands with only positive reviews often lead to customers assuming that the reviews are fake or that negative customer comments have been removed. Such assumptions may affect the brand's trustworthiness and reputation in the long run. The findings of this research suggest that hospitality service providers should strategically accentuate negative information in C2C interactions on social media. More precisely, managers should not worry excessively about negative peer customer responses that challenge shared positive eWOM because such comments do not deter consumers from future positive engagement and repurchase behaviors. However, they should pay careful attention to peer customer responses that criticize a focal customer's shared negative eWOM because such comments can engender future helpful engagement and repurchase behaviors. To dilute the detrimental effects of a focal customer's negative eWOM, hospitality service providers should encourage satisfied customers to act as brand ambassadors and post comments that portray what is good about the service. This allows the service provider to restore the failed services provided and gain another opportunity to satisfy the focal customer.
Finally, our examination of culture at the national level aimed to understand the dominant values that govern the eWOM behaviors of African tourists. This enabled us to distill eWOM enactment and consequent effects in a manner that is relevant to firms. Based on our findings, we suggest that providers of hospitality services consider the collectivistic and individualistic orientations of tourists when analyzing social media brand-related C2C interactions. Specifically, managers must understand that it makes little to no sense to segment tourists into collectivists and individualists when analyzing an eWOM consensus. Rather, managerial actions targeted at improving engagement and repurchase behaviors will realize benefits when there is inconsistency in the social media eWOM-triggered C2C interactions because the coping strategies for such dissonance vary between collectivistic and individualistic tourists. Therefore, it is more effective to target collectivistic tourists, such as Ghanaians, with inconsistent social media C2C interactions because they are better able to readjust their post-eWOM behaviors than are individualistic consumers, such as South Africans. For instance, when dealing with collectivistic tourists, hotel management could develop intervention strategies to strengthen the tourists' favorable attitude by highlighting the positive consensus among them and take recovery actions to restore the inconsistencies between different tourists. Meanwhile, from the social media management's perspective, the social media platforms should better play the mediating role that encourages customers to not only share their own experience but also openly discuss with others, thereby helping other customers and maintaining the interactivity of the platform. Similarly, social media managers should also encourage service firms' engagement on social media, particularly in terms of providing additional support for consumer interactions (Xu, Liu, and Gursoy 2019). After all, the interactivity of social media platforms is vital in their sustainable development in the long term ).

Limitations and Suggestions for Future Research
As with every piece of research, this study is limited in some respects. Therefore, our findings should be interpreted with caution given the following limitations, and the suggested future research directions should be considered. First, we employed a scenario-based experimental design. This enabled us to rule out alternative explanations and focus on the variables of interest and cause-effect relationships as precisely and accurately as possible (Calder, Phillips, and Tybout 1981). To ensure that we mimicked reality as much as possible, we presented stimuli to the participants in a manner typical of user experiences while also checking the realism of the experimental scenarios. However, an often-accepted weakness of scenario-based experimental studies is that they can be threatened by external validity, thereby posing challenges in generalizing the findings (Song, Noone, and Mattila 2018). More precisely, although our findings revealed the significant impact of opinion consensus between the focal customer and his/her peers, the focal customer's repurchase decision in practice is often subject to many other considerations (e.g., hotel location and convenience, budget constraints, fellow travelers' preference, and competitive alternative choices). Therefore, to overcome such a shortage, further research is expected to identify other factors that affect the customer's repurchase decision using other research methods (e.g., field experiment, large-scale survey, and in-depth interviews). Additionally, a longitudinal approach will allow future research to illustrate the customer journey of repurchase decision-making. Such an approach will complement the experimental design, capture other influencing factors, and enhance the generalizability of our findings.
Second, our study used Facebook as an example of a social media eWOM platform and identified C2C interactions elicited by the focal customer's positive and negative eWOM-giving on their post-eWOM behaviors (Ben-Shaul and Reichel 2018). This approach suffers from several limitations. More specifically, although our focus on Facebook contributes to understanding the often-neglected social media eWOM platform (vs. online opinion sites), it also causes difficulties in generalizing the findings to other social media platforms because systems and interfaces of social media sites (e.g., Facebook, Twitter, and Instagram) vary significantly, and C2C interactions may be platform-specific (Liu et al. 2021a). Additionally, submission devices (desktop vs. mobile) and the popularity of social media sites also influence the mechanism of the interactions (Mariani, Borghi, and Gretzel 2019;Wang, Kirillova and Lehto 2017). Therefore, future research should replicate this study on other social media sites using various devices and identify the impact of the media penetration rate and device differences.
Additionally, following prior similar research on eWOM information (in)consistency (Liu et al. 2020;Wu et al. 2016) and cognitive dissonance theory, we focused on positive/ negative eWOM-giving and supportive/critical responses in this study. However, neutral eWOM may also exist in both eWOM-giving and eWOM-responding in practice (Mudambi and Schuff 2010). Thus, future research could consider alternative theoretical angle and examine the effects of neutral eWOM-giving and -responding. Meanwhile, high-end hotels (e.g., 5-star hotels) attract more positively valenced comments than budget hotels (e.g., 1-star hotels) (Mariani and Borghi 2018). Therefore, we suggest that future research should examine the effects of eWOM-triggered C2C interactions across different levels of hotels.
Moreover, this study focused on the focal customer's perspective and examined the C2C interactions on the focal customer's post-eWOM evaluations and behaviors. However, eWOM-triggered interactions in practice are complicated. For example, whether a peer customer responds to the eWOM-giver's post depends on the signals revealed by the post (e.g., interest level, helpfulness, and captivating style), while whether the eWOM-giver values the peer customer's response is often subject to the characteristics of the peer customer (e.g., credibility, expertise, and homophily) (Martínez, Herrero, and García-de los Salmones 2020;Mehraliyev, Choi, and King 2021;Zhang, Liang, and Qi 2021) and their own eWOM experiences as a reviewer (Mariani and Predvoditeleva 2019). Therefore, we suggest that future research should further identify the dynamics in C2C interactions by exploring which factors influence the quality of C2C interactions from the perspectives of both eWOM-givers and peer customers. Meanwhile, the concept of customer engagement includes both online and offline dimensions, while the online activities may also influence customers' offline engagement behaviors (Taheri, Jafari, and O'Gorman 2014;So, Wei, and Martin 2020). Therefore, future research should also examine the effects of online C2C interactions on customers' offline engagement with a brand, and vice versa.
Third, this study also highlighted the significance of culture in shaping customer behaviors. However, the cultural dimensions were limited to collectivism vs. individualism. Although our findings provide pioneering evidence for understanding eWOM-triggered C2C interactions based on national culture, which aligns with previous studies that utilized a single cultural dimension (e.g., Izogo, Mpinganjira, and Ogba 2020;Song, Noone, and Mattila 2018), only the two most polar opposite African countries along the collectivism/individualism cultural dimension (Ghana and South Africa) were studied. Thus, the sampling procedure utilized in this research limits the extent to which the findings are generalizable. This warrants replication across other African countries. Meanwhile, although most eWOM research highlights the importance of collectivism/individualism in shaping customers' eWOM engagement, Hofstede's (2011) other cultural dimensions may also shape customer behaviors. Therefore, we encourage future research to identify polar cases for other cultural dimensions and develop a holistic view of eWOM communication according to Hofstede's (2011) cultural dimensions. For instance, could indulgence versus restraints influence social media eWOM-triggered C2C interactions in a way that leads to dramatically different behavioral outcomes in polar countries, such as Nigeria and Egypt? Extending this line of research along other cultural dimensions that differ substantially across African countries and thus have the potential to impact consumers' behavioral responses toward social media C2C interactions would strengthen the framework within which hoteliers can analyze eWOM-triggered C2C interactions on social media.