The influence of eWOM. Analyzing its characteristics and consequences, and future research lines

Abstract

Purpose

This study, first, reviews the existing literature on electronic word-of-mouth (eWOM) and, using communication theory, examines its impact on its readers’ decision-making processes. Second, this paper aims to propose some elements of eWOM communications that might be further researched.

Design/methodology/approach

A literature review of relevant academic articles produced 97 works related to social communication theory, eWOM and new artificial intelligence trends in hospitality. Thereafter, potential avenues for future research were explored.

Findings

The study results showed: valence, relevance, understandability and visual cues are the most important antecedents of message usefulness and the reader’s behavioral intentions; source credibility is the sender characteristic that most affects the reader’s behavioral intentions and consumer susceptibility to interpersonal influence is the receiver characteristic that most influences their attitudes and behavioral intentions. In addition, the study highlights four relevant aspects for future research. First, more research into online fake reviews is needed to better understand sender motivations. Second, companies should actively manage negative reviews. Then, the careful choice of platforms on which companies promote their products/services. Finally, the role of artificial intelligence in increasing the effectiveness of eWOM in the hospitality industry.

Originality/value

This paper integrates – grounded on communication theory – results from previous studies about the central elements of communication (message, sender and receiver) and discusses the main trends in hospitality-related eWOM. In addition, the paper examines the potential of specific eWOM elements as future lines of research, in particular: fake reviews, strategies for dealing with negative reviews, the eWOM platform used and artificial intelligence applications.

Propósito

Este estudio, en primer lugar, revisa la literatura existente sobre el boca-oído electrónico (eWOM) y, utilizando la teoría de la comunicación, examina su impacto en los procesos de toma de decisiones de sus lectores. En segundo lugar, proponemos algunos elementos de las comunicaciones eWOM que podrían investigarse más adelante.

Metodología

Una revisión de la literatura de artículos académicos relevantes produjo 97 trabajos relacionados con la teoría de la comunicación social, eWOM y nuevas tendencias de inteligencia artificial en hotelería. A partir de entonces, se exploraron posibles vías de investigación futura.

Hallazgos

Los resultados del estudio señalaron: 1) la valencia, la relevancia, la comprensibilidad y las señales visuales son los antecedentes más importantes de la utilidad del mensaje y las intenciones de comportamiento del lector; 2) La credibilidad de la fuente es la característica del emisor que más afecta las intenciones de comportamiento del lector; 3) La susceptibilidad del consumidor a la influencia interpersonal es la característica del receptor que más influye en sus actitudes e intenciones de comportamiento. Además, el estudio destaca cuatro aspectos relevantes para futuras investigaciones. Primero, se necesita más investigación sobre las reseñas falsas en línea para comprender mejor las motivaciones del emisor. En segundo lugar, las empresas deben gestionar activamente las críticas negativas. Luego, la cuidadosa elección de las plataformas en las que las empresas promocionan sus productos/servicios. Por último, el papel de la inteligencia artificial en el aumento de la eficacia de eWOM en la industria hotelera.

Originalidad

Este artículo integra –con base en la teoría de la comunicación– resultados de estudios previos sobre los elementos centrales de la comunicación (mensaje, emisor y receptor) y analiza las principales tendencias en eWOM relacionadas con la hostelería. Además, el artículo examina el potencial de elementos específicos de eWOM como líneas futuras de investigación, en particular: revisiones falsas, estrategias para lidiar con críticas negativas, la plataforma eWOM utilizada y aplicaciones de inteligencia artificial.

Palabras clave: eWOM, Teoría de comunicación, Revisiones falsas, Revisiones negativas, Inteligencia artificialTipo de artículo: Revisión de la literatura

目的

本研究, 首先, 回顾了现有的关于电子口碑（eWOM）的文献, 并利用传播理论, 研究了它对读者决策过程的影响。其次, 我们提出了一些可能需要进一步研究的电子口碑传播的要素。

方法。

对相关学术文章的文献回顾产生了97篇与社会传播理论、eWOM和酒店业新的人工智能趋势有关的作品。此后, 对未来研究的潜在途径进行了探讨。

研究结果。

研究结果显示。1）价值、相关性、可理解性和视觉线索是信息有用性和读者行为意图的最重要的前因; 2）来源的可信度是最能影响读者行为意图的发送者特征; 3）消费者对人际影响的易感性是最能影响他们态度和行为意图的接受者特征。此外, 该研究还强调了未来研究的四个相关方面。首先, 需要对网上虚假评论进行更多的研究, 以更好地了解发送者的动机。其次, 公司应该积极管理负面评论。然后, 谨慎选择公司推广其产品/服务的平台。最后, 人工智能在提高酒店业eWOM的有效性方面的作用。

原创性。

本文以传播理论为基础, 整合了以往关于传播中心要素（信息、发送者和接受者）的研究结果, 并讨论了与酒店业相关的电子WOM的主要趋势。此外, 本文还研究了特定的eWOM要素作为未来研究方向的潜力, 特别是：虚假评论、处理负面评论的策略、使用的eWOM平台和人工智能应用。

关键词：eWOM, 传播理论, 虚假评论, 负面评论, 人工智能。

Citation

Akdim, K. (2021), "The influence of eWOM. Analyzing its characteristics and consequences, and future research lines", Spanish Journal of Marketing - ESIC, Vol. 25 No. 2, pp. 239-259. https://doi.org/10.1108/SJME-10-2020-0186

1. Introduction

Technological developments over the past years have changed the communication environment, leading to the emergence of electronic word of mouth (eWOM). eWOM has been defined as all informal communications directed at consumers, through internet-based technologies, related to the usage or characteristics of particular goods and services (Litvin et al., 2008). eWOM has been shown to cause great changes in consumer behaviors in some sectors, principally hospitality (Cantallops and Salvi, 2014). The internet is used for planning trips in more than 75% of all cases (Han et al., 2020). In addition, Oliveira et al. (2020) found that travelers prefer to rely on peers’ online reviews for unbiased information, as other travelers are seen as more reliable than other information sources and create more realistic expectations (Casaló et al., 2011).

Recently, various authors have suggested that travel is one of the sectors most influenced by eWOM (Cantallops and Salvi, 2014). To date, most research carried out in this context has focused on the impact of eWOM on hotel room bookings (Ye et al., 2009), hotel performance (e.g. reputation, overall performance, booking intentions [Cantallops and Salvi, 2014; Yang et al., 2018]), hotel booking intentions (Sparks and Browning, 2011) and hotel choice (Vermeulen and Seegers, 2009). Most of these studies focused on some characteristics of the key elements of eWOM (message and/or sender and/or receiver [Chandler, 1994]), but there is no holistic understanding of the influence of the characteristics of eWOM elements on consumer decision-making processes in the travel sector. Indeed, the only study that has, thus, far reviewed the elements of eWOM communication did so in virtual settings, following a general approach (Cheung and Thadani, 2012).

An updated view of the current status of knowledge in the prolific domain of eWOM would help hospitality managers make better decisions about where to focus their research efforts. Thus, this article aims to map the current state of research in the eWOM field and identify emerging future research lines. To address this need we integrate – grounded on communication theory– results from previous studies into the central elements of communication (message, sender and receiver) and discuss the main eWOM trends in the travel realm. In addition, we propose some elements of eWOM communications that might be further researched; fake and negative reviews as message-related elements, companies’ strategies in dealing with negative reviews as receiver-related elements, artificial intelligence (AI) as a sender-related element and the platform as a medium-related element which determines how senders and receivers relate to each other.

The remainder of the present study is organized as follows. First, we present the methodology used to conduct the literature review and identify the research gaps. Second, we present the literature review, focusing on the determinants of the impact of eWOM from the perspective of communication theory. Then, we use the results of our analysis of the academic literature to identify research gaps that we recommend be addressed in the future. Thereafter, we present the discussion and the main conclusions, the theoretical and the practical implications and, finally, we discuss the study’s limitations.

2. Methodology

The present study is based on an analysis of a collection of academic articles retrieved from Web of Science and SCOPUS. The works were selected based on the following criteria: eWOM is the main focus of the investigation; publications addressing elements of social communication; publications dealing with eWOM in hospitality and publications about new AI trends in the eWOM realm.

The articles were scanned based on words included in their titles, keywords and abstracts; terms searched for included eWOM, social communication, online reviews, online recommendation, communication theory, fake reviews, negative reviews, online travel communities and artificial intelligence. The articles were taken from scientific journals selected based on their importance, research focus, academic rating and the number of related papers they had published. The main journals from which papers were selected were: Tourism management, Computers in Human Behavior, International Journal of Contemporary Hospitality Management and the International Journal of Hospitality Management. Some 97 articles were found that related, in various ways, to the research topic.

3. Literature review: Determinants of the impact of eWOM from the perspective of communication theory

Hovland (1948), one of the founding fathers of social communication theory, defined social communication as “the process by which an individual (the communicator) transmits stimuli (usually verbal symbols) to modify the behavior of other individuals (communicates)” (p. 317). Based on this theory, our work analyzes eWOM communications as a new form of social communication.

In the travel sector eWOM characteristics have been found to influence several consumer perception-related variables (e.g. usefulness [Casaló et al., 2015] and trust [Sparks and Browning, 2011]), evaluations (e.g. attitude [Vermeulen and Seegers, 2009]) and intentions and behaviors (Sparks and Browning, 2011). For instance, both reviewer (e.g. expert vs non-expert) and review characteristics (e.g. review valence) may affect the usefulness of reviews (Casaló et al., 2015), which, in turn, determine intention to follow the review advice/recommendation (Casaló et al., 2011). Similarly, focusing on hotels, review valence (positive vs negative) and/or reviewer expertise may affect the reader’s attitude toward the reviewed service (Vermeulen and Seegers, 2009). Several works have confirmed that review characteristics such as valence and numerical ratings may, in combination, increase booking intentions (Sparks and Browning, 2011). Figure 1 displays the variables analyzed based on the characteristics of the message, the sender and the receiver (Chandler, 1994).

3.1 Online consumer reviews

Online consumer reviews are the most common type of eWOM (Chatterjee, 2001). They serve, first, to provide information about products/services and, second, as recommendations (Park et al., 2007). According to Casaló et al. (2008, 2009), online reviews are an important information source that provides consumers with detailed and reliable information by sharing past consumption experiences. Thus, they are perceived as more credible than information provided by managers. In particular, consumers tend to rely more on other consumers’ reviews when purchasing high involvement products (Park et al., 2007); as travel-related products are high involvement, we expect travel-related reviews will be extensively consulted in travel-related decisions.

Various researchers have examined the impact of online reviews on readers’ behaviors such as booking trips (Xiang and Gretzel, 2010) and hotels and restaurants (Sparks and Browning, 2011) and even on the consumer’s emotional state (Flavián-Blanco et al., 2011). For instance, Dickinger and Mazanec (2008) found that reviews can significantly increase the consumer’s booking intentions and the number of bookings in hotels. In addition, Vermeulen and Seegers (2009) found that online reviews improved consumers’ awareness of hotels and honed their consideration sets. Ye et al. (2009) concluded that review ratings are important elements in the prediction of online room sales. Flavián-Blanco et al. (2011) found that online reviews influence the emotional state of consumers because of the effort they must make in their searches. This can affect their behavioral intentions. In summary, the analysis of online reviews can help improve the quality of products/services, the identification of customer needs and the implementation of new marketing strategies (Yacouel and Fleischer, 2012).

3.2 Electronic word of mouth communication elements

eWOM is a relatively new form of social communication that involves both information seeking and information sharing with customers. Working within this framework and following our review of the previous related works, we conclude that there are three major communication elements (Chandler, 1994): the message is the communication (positive, negative or neutral) transmitted by a sender to a receiver; the sender is the person who transmits the communication and the receiver is the individual who receives the communication. In addition, we identify the different variables related to each element and their interrelationships.

3.2.1 Message characteristics.

3.2.1.1 Intrinsic characteristics.

The intrinsic characteristics of the message are based on the concept that information has quality in its own right (Wang and Strong, 1996). Nelson et al. (2005) defined intrinsic characteristics as properties of information independent from any specific user, task or application. Accuracy, objectivity and valence have been proposed as the dimensions of intrinsic characteristics (Wang and Strong, 1996; Mudambi and Schuff, 2010). Information accuracy is the extent to which information is correct and believable (Wand and Wang, 1996). Information objectivity refers to rational and concrete content and valid argumentation (Park and Lee, 2008). Message valence is the positive, negative or neutral direction of information (Mudambi and Schuff, 2010). In the online context, positive messages highlight the strengths of products/services and encourage people to acquire them, while negative eWOM emphasizes the weaknesses of products/services and, thus, discourages acquisition (Dellarocas et al., 2007; Flavián et al., 2021).

Prior studies into online consumer reviews have suggested that information accuracy has, in general, a positive influence on perceptions of the usefulness of online reviews (Cheung et al., 2008) and on readers’ intentions to adopt the information (Filieri and McLeay, 2014); in particular, in hotel reviews (Zhang et al., 2016). Media richness theory proposes that the more accurate is the message, the higher will be the receiver’s perceptions of its usefulness. In addition, it has been found that perceptions of the objectivity of information positively affect attitudes toward the information (Park and Lee, 2008). Negative reviews are perceived as more useful than positive reviews and generate negative attitudes (Lee et al., 2008). This finding is in line with Cacioppo and Berntson (1994), who suggested that negative input has a greater effect on attitudes and behaviors than has positive input. Nonetheless, positive online reviews have been found to enhance consumers’ assessments of hotels (Vermeulen and Seegers, 2009), booking intentions and sales (Ye et al., 2009).

3.2.1.2 Contextual characteristics.

Contextual characteristics are based on the concept that information must relate to the context of the task at hand; that is, to add value information must be relevant, timely and complete (Wang and Strong, 1996). Filieri and Mcleay (2014) argued that information relevance refers to the extent to which it is applicable and helpful for a task. Information timeliness refers to its status as up to date (Nelson et al., 2005). Completeness refers to the extent to which information has sufficient depth and scope to address the task at hand (Wang and Strong, 1996).

Previous research in the travel sector has demonstrated that information relevance has a significant impact on perceptions of information usefulness and is an important antecedent of behavioral intentions (Filieri and Mcleay, 2014). In addition, as consumers seek quick and effortless information, the timelier is the message, the more it is of use (Cheung et al., 2008). Regarding information completeness, reviews, which include detailed information (e.g. information about the product/service, pictures) can alleviate readers’ uncertainty about a product/service and allow them to develop more confidence in their decision-making processes (Cheung et al., 2008).

3.2.1.3 Representational characteristics.

Representational characteristics are related to how information is presented, that is, is it understandable, easy to read and easy to interpret (Wang and Strong, 1996)? Representational characteristics are language, semantic, lexical expressions and visual cues that increase the understandability and the ease of interpretation of a review (Davis and Khazanchi, 2008).

In the context of online travel communities, some works have shown that a significant relationship exists between ease of reading and readers’ perceptions of the usefulness of reviews (Liu and Park, 2015) and that readers tend to search for reviews that help them obtain the specific information they need. In addition, understandability has been shown to be a positive influence on information adoption in online reviews (Cheung et al., 2008). Visual cues, such as videos, have been found to improve the usefulness of reviews and increase the intention to follow their advice (Casaló et al., 2015; Orús et al., 2017; Flavián et al., 2009, 2017). Moreover, semantic content and style properties such as effective content and figurative language may reinforce the impact of the online reviews on intention to follow their advice (Ludwig et al., 2013).

3.2.2 Sender characteristics.

3.2.2.1 Source credibility.

Various studies have noted that source credibility is the most investigated message sender-related factor (Sussman and Siegal, 2003). Petty and Cacioppo (1981) described source credibility as the extent to which an information source is perceived to be believable and trustworthy. Hovland and Weiss (1953) argued that credibility has two dimensions, expertise and trustworthiness. In online environments users often seek out others who offer trustworthy advice (Boush and Kahle, 2002). In the travel context, credibility has been found to be important due to the intangible nature of travel products and the psychological risks associated with travel-related decision-making (Loda et al., 2009). Casaló et al. (2011) argued that source credibility enhances the usefulness of online reviews and that readers consider that reviews are useful if they provide sufficient, good information that is likely to help them predict how an experience will turn out. Ayeh et al. (2013) showed the positive influence of source credibility on readers’ attitudes toward using user-generated content in their travel planning. In addition, Filieri and Mcleay (2014), Sussman and Siegal (2003) and Ayeh et al. (2013) showed that travel information messages perceived to have higher source credibility are associated with higher levels of adoption.

3.2.2.2 Attribution.

Attribution theory proposes that individuals make causal inferences as to why communicators advocate certain positions or behave in certain ways (Mizerski et al., 1979). Specifically, the theory proposes that the more the reader attributes a review about a product to the product’s actual performance, the more the reader will perceive that the communicator is credible, the stronger will be the reader’s belief that the product is as described in the review and the more the reader will be persuaded by the review (Mizerski et al., 1979). Thus, in the travel industry, for instance, if a reader feels that a reviewer has positively reviewed a hotel because of its actual performance and the review relates to the core services of hotels, then (s)he will attribute the review to the quality of the hotel.

The attribution process has been shown to play a significant role in consumers’ evaluations, attitudes, behaviors (Weber and Sparks, 2010) and decision-making (Mizerski et al., 1979). In the context of online travel communities, Browning et al. (2011) reported that consumer reviews that readers associate with hotels’ services and characteristics are more likely to affect their perceptions and be seen as more useful and generate more positive attitudes than reviews that the readers associate with the characteristics of the reviewer.

3.2.3 Receiver characteristics.

3.2.3.1 Consumer susceptibility to interpersonal influence.

The consumer’s susceptibility to interpersonal influence is his/her general tendency to accept information from others as true (Deutsch and Gerard, 1955). Consumers highly susceptible to interpersonal influence have been shown to be more influenced by the opinions of others when making purchase decisions (Schroeder, 1996). In the eWOM context, an individual with a greater propensity to be influenced by others is likely to attach more weight to eWOM information than one who is less susceptible to interpersonal influence (Sparks and Browning, 2011). It has been found that consumer susceptibility to interpersonal influence positively impacts their attitudes and behavioral intentions (Lee et al., 2011). Park and Lee (2008) found that consumers with greater susceptibility to interpersonal influence perceive reviews as being more useful than do consumers with less susceptibility to interpersonal influence. Sharma and Klein (2020) argued that informing their behavioral intentions, it is likely that individuals more easily influenced by information provided by others will give more weight to their perceptions of the advice offered.

3.2.3.2 Risk aversion.

Risk aversion has been defined as “the extent to which people feel threatened by ambiguous situations and have created beliefs and institutions that try to avoid them” (Hofstede and Bond, 1984, p. 419). In the WOM communication context, previous research has shown that people who perceive higher risk will seek out WOM communication more actively than people who perceive lower risk (Arndt, 1967). Therefore, WOM is a credible source of information to draw on to assess risk and reduce uncertainty about behavioral decisions (Murray, 1991).

Previous research has revealed that the individual’s level of risk aversion can impact his/her decision-making processes and general attitudes and behaviors (Mandrik and Bao, 2005). In the online travel context, consumers’ levels of risk aversion may make them sensitive to safety issues when choosing travel-related products (Mandrik and Bao, 2005). As a result, those with high levels of risk aversion may rely heavily on online travel communities for trustworthy information to use in their purchasing decisions.

Table 1 summarizes some characteristics – not previously discussed – related to the communication elements of eWOM. The table displays, for each characteristic, the definition, the associated eWOM element, the main consequences and some related previous studies.

4. Future research opportunities

Although eWOM has been studied only over the past 15–20 years, the phenomenon has attracted scholars from diverse fields such as marketing (Park and Kim, 2008), communications (Cheung et al., 2009) and psychology (Park and Lee, 2008). However, there are controversies and research gaps in the hospitality-related eWOM literature that remain to be addressed. In this section, we discuss some potential future research avenues related to the aforementioned eWOM elements: fake reviews as message-related elements, strategies for dealing with negative reviews as receiver-related elements, platforms as medium-related elements determining how senders and receivers relate to each other and AI systems as sender-related elements.

4.1 Fake online reviews

Recently, social media and consumer review sites have come to seem quite unreliable and fake (Luca and Zervas, 2016), which is an obvious concern for users. Fake reviews have been described as “the intentional control of information in a technologically mediated message to create a false belief in the receiver of the message” (Hancock, 2007, p. 290). A natural consequence of the existence of fake online reviews is that, in general, the credibility of all reviews will decline (Lappas et al., 2016).

One of the main findings of a study by the UK’s Competition and Markets Authority (2016) was that fake positive reviews are more common than fake negative reviews. The study suggested that this may be because it is easier and less risky for business owners to post positive reviews about their own entities than to post negative reviews. As an example, one study found that around 20% of hospitality industry reviews were potentially fake (Wu et al., 2020). In practice, these percentages are likely to vary across platforms (Lappas et al., 2016). Recently, some researchers have examined reviewers’ motivations to post fake reviews. Choi et al. (2017) and Rixom and Mishra (2014) suggested that powerless individuals are more likely to write fake reviews when offered money than when incentivized by charitable motives.

The growing concern over fake online reviews has motivated online communities to implement various types of defense mechanisms. For instance, reviews that TripAdvisor regard as suspicious can be placed on hold and even be eliminated if the website’s proprietary filtering process finds enough evidence (TripAdvisor Releases Data Detailing Fake Review Volumes In First-Of-Its-Kind Transparency Report, 2019). In addition, Mayzlin et al. (2014), in the context of TripAdvisor, discussed the “verified buyer” badge, which allows only those who have actually purchased a product/service to post reviews. In any case, little is still known about the motivations of those who post fake reviews, their consequences for companies (e.g. reputational loss) and their influence on consumers’ decision-making processes.

4.2 The main companies’ strategies for dealing with negative reviews

While conventional wisdom suggests that any publicity is good publicity, existing research (Zhang et al., 2016) has shown the many downsides of negative reviews such as reputational harm, reduced sales and decrease in consumer trust. Despite the potential harm, only limited research has examined strategies for handling negative reviews. This section examines the main strategies followed by companies for dealing with negative reviews and discusses their advantages and disadvantages. Future research might use this examination as a basis for exploring the effectiveness of each strategy in more detail.

4.2.1 Response.

Response strategy involves listening to, acknowledging and addressing the negative feedback generated by online communities (Rajan and Kundu, 2016). Web specialists have attained near consensus on the most appropriate method of handling unfavorable reviews: respond as positively as possible (Kiesow, 2010). Response strategy recognizes that marketing communication is interactive and consumers’ reactions to messages deployed must be considered in the process. In addition, a response strategy provides an opportunity to quickly and politely correct inaccurate information (Barone, 2009).

Response strategy has the potential to increase customer loyalty and build stronger consumer-company relationships (RightNow.com, 2011). Sparks et al. (2016) showed that, when compared to the no-response baseline, responses from hotels created significantly more trust and positive consumer attitudes. Nieto et al. (2014) found, in the context of Spanish rural lodging establishments, that when companies responded appropriately to negative reviews their average ratings improved. These results are consistent with the findings of Lee and Song (2010), who found that individuals evaluate companies positively when they respond to online complaints. Despite the advantages of the response strategy, it has some significant downsides. For instance, managers must be careful to advocate without angering entire communities of consumers who will see the message and they should understand at what point they need to respond (Thomas et al., 2012).

4.2.2 Delay or ignoring.

Delay strategy is based on the concept that if a company does not respond to a negative review the issue will eventually die down (Vogt, 2009). Although the delay/ignore strategy appears to be a less viable option in today’s socially mediated world, there are reasons why companies may choose to adopt it. For example, by ignoring negative attacks management avoids engaging in a tug-of-war with consumers attacking their brand image (Thomas et al., 2012). However, a delay strategy can create the belief that the company is being unresponsive and is unwilling to listen to its customers. When companies are unresponsive or slow to respond, they are perceived as uncaring and/or guilty of the actions/inactions complained of by their accusers (Thomas et al., 2012).

4.2.3 Partnership.

A partnership strategy engages the consumer in the process, indeed, may involve turning over control of the process to the consumer. The strategy involves the company working in partnership with consumers (fans), thereby creating constructive and committed relationships (Thomas et al., 2012). By collaborating with fans- potential customers- companies can come to understand their preferences and, thereby, respond appropriately to their needs and reduce complaints and negative comments.

For example, Coca-Cola’s Facebook page was created by two fans, helped by the company’s marketing team. Using this strategy, in which the consumer controls its social media reviews, Coca-Cola has built an effective partnership with its fans (Graham, 2011). Consequently, the company is perceived to have a high level of authenticity and transparency. In addition, partnership strategies enable companies to benefit from their fans (Safko and Brake, 2009). In fact, according to Graham (2011), fans are twice as likely to consume and 10 times more likely to purchase, the product than non-fans. However, while partnering with fans can help the company deal with negative online attacks, it does involve giving up control. Simply put, if the company and its partners fall out, the partners have insider knowledge that gives them a great deal of power to do damage.

4.2.4 Censorship.

Censorship is a strategy through which companies attempt to take control over consumer reviews by removing unwanted information (Dekay, 2012). Giving up control of the message is difficult for companies accustomed to operating within the one-way communication model (Thomas et al., 2012) and adopting censorship as a strategy can create negative publicity which can quickly spread among online communities (Jackson, 2008). Although this strategy allows companies to maintain greater control over reviews, the nature of online communities is such that negative perceptions may continue to exist and the companies’ tactics may be seen as aggressive and hostile (Thomas et al., 2012).

4.3 The moderating effects of the platform on electronic word of mouth

The internet has facilitated eWOM communication between customers on a variety of platforms (Cheung and Thadani, 2012). For example, eWOM is exchanged on personal blogs, social networking sites such as Facebook and on online travel communities such as TripAdvisor; academics need to analyze whether these platforms reinforce or diminish the influence of eWOM. Previous studies have mainly tested the effects of eWOM posted on social media (See-To and Ho, 2014) and on online travel communities (Belanche et al., 2019; Casaló et al., 2011; Casaló et al., 2010; Liu et al., 2019) separately. In both cases, eWOM was found to influence consumers’ purchase intentions. However, the effects of eWOM posted on these two different platforms have not hitherto been compared, although there are three major differences between them that may increase or decrease the influence of eWOM.

4.3.1 Tie strength.

Tie strength is the level of intensity of the social relationship between individuals; it varies greatly across consumer social networks (Steffes and Burgee, 2009). There are two types of tie strength, strong ties and weak ties. People with whom one has strong ties are regarded as more credible and reliable than people with whom one has weak ties. Liviatan et al. (2008) proposed that people have more detailed and concrete knowledge about individuals with whom they share close relationships because this closeness involves more intimate interactions and exposure to privileged information about the other person’s personality (Koo, 2016). Information derived from strong ties is perceived as more useful and information derived from weak ties is considered less valuable, even questionable (Wang and Chang, 2013). In the online context, it has been found that the strength of the tie between the sender and the reader positively impacts eWOM adoption (Kim and Bae, 2016).

4.3.2 Social cues.

Social information cues include information about personal identities and spatial and environmental contexts. Social networks usually provide richer social cues than do online travel communities (Baym, 2015). Previous research has shown that a lack of social cues – source information – negatively impacts perceptions of the credibility of online reviews (Dellarocas, 2003). Similarly, Park et al. (2014) argued that personal profile information is an important social cue, which significantly impacts eWOM source credibility. More specifically, Lee and Youn (2009) found that because there are more social cues in personal blogs than there are on other platforms, readers perceive reviews posted on personal blogs as more useful.

4.3.3 Expertise and objectivity.

Travelers engage in information exchange in online travel communities because they provide the opportunity to share knowledge gained from previous trips (Lee and Yang, 2015). Previous research (Liu et al., 2019) has shown that travelers with past experiences who engage in eWOM communication are most likely to be the more preferred sources of information and the most influential, in the pre-trip stage of travel decision-making. In this respect, the distinctiveness of online travel communities, compared to social networking sites, is the travelers’ expertise (Wang and Fesenmaier, 2004). Previous studies have highlighted some unique roles of online travel communities. First, they permit individuals to access other travelers’ knowledge and feelings about specific destinations (Chalkiti and Sigala, 2008; Chang and Chuang, 2011); and they are used as collaborative decision-making platforms that offer unbiased information about emotional experiences regarding given destinations and travel products (Casaló et al., 2011).

4.4 Artificial intelligence in electronic word of mouth (algorithmic word of mouth)

Algorithmic word of mouth (aWOM) is any communication created and shared by non-human AI tools that supports individuals’ decision-making about destinations and/or activities (Young et al., 2018). AI-generated content can be provided via voice assistants such as Alexa, chatbots and virtual assistants (Prasad, 2019). Kozubska (2018) found that recommendations are a common application of AI in the travel industry. For instance, AI is used to make recommendations for flights, hotels, restaurants and clubs based on the user’s preferences, past booking history and search results. This application can reduce the many challenges that travelers face when organizing trips. The following section briefly reviews some examples of AI applications in the travel sector.

4.4.1 Chatbots.

Chatbot is a natural language computer program designed to approximate human speech and interact with people via a digital interface (Thomaz et al., 2020). Chatbots are configured to self-learn in response to users’ requests instead of using pre-programmed answers. When a chatbot gets a new text input, its keywords are saved for future data processing. Hence, the number of questions/situations that it can handle continue to grow and the correctness of each reply it makes may increase (Frankenfield, 2018). Chatbots have multiple applications in various industries, growing particularly strongly in the travel industry (Faggella, 2019). For example, Expedia and Skyscanner took advantage of Facebook’s technologies to launch a basic bot that helps travelers book hotel rooms. Marriott Hotels also features a chatbot on its website that offers basic services such as booking a room.

Chatbots are an alternate way for travelers to search for tips and information; they simply ask questions using keywords and receive appropriate answers (Faggella, 2019). Ambawat and Wadera (2019) found that chatbots improve the customer’s experience; in particular, they can enrich the pre-arrival experience. In addition, this technology can help maintain relationships between consumers and companies.

4.4.2 Voice assistants.

AI-powered voice assistants that engage in natural conversational interactions with humans have been integrated into various devices (e.g. smartphones and cars). Individuals can interact with devices using their voice as the input mechanism to receive oral advice/recommendations about products/services (Kozubska, 2018). Most related studies have demonstrated that voice assistants reduce search effort and increase cross-selling by providing product/services recommendations. Thus, they can affect the consumer’s decision-making processes (Tam and Ho, 2006). Also, it has been argued that voice assistants create a more intimate experience, humanize interactions, simulate social presence and enhance trust (Cherif and Lemoine, 2017). Simms (2019) found that voice assistants learn the consumer’s preferences and consequently increase their influence on his/her behavioral intentions. Nevertheless, some studies have suggested that these technologies raise privacy concerns (Simms, 2019).

5. Discussion and conclusions

The present study differs from previous works that have focused on the effects of eWOM on its readers’ behaviors, the factors that generate eWOM and its impact on hotel performance (Cheung and Thadani, 2012; Cantallops and Salvi, 2014); this work provides- based on social communication theory- a holistic understanding of the influence of the elements of eWOM on the reader’s decision-making processes in the online travel context and identifies emerging hospitality-related eWOM trends for possible future research.

The present study examines three essential components of eWOM communication – message, sender and receiver – and its impacts on the reader’s decision-making processes (perceptions, evaluations, intentions and behaviors). It has been found that message valence is the variable that has been most examined in terms of message usefulness and intention to adopt the message (Lee et al., 2008). Similarly, previous studies have shown that the message’s relevance, understandability and visual cues are important antecedents of the reader’s behavioral intentions (Filieri and Mcleay, 2014; Wang and Strong, 1996). Regarding the sender element, it has been argued that source credibility is the most important feature of readers’ decision-making processes, due to the intangible nature and economic and psychological risks associated with travel-related products. In turn, the readers’ susceptibility to interpersonal influence is –among the other receiver characteristics previously discussed – the variable that has been most examined in terms of influencing attitudes and behavioral intentions.

The present study also identifies future research lines. First, more research is needed into fake online reviews to better understand senders’ motivations. In particular, to establish the underlying psychological mechanisms that cause individuals to post fake reviews without external financial incentives. Second, based on the principle of “bad is stronger than good,” readers tend to value negative eWOM more than they value positive eWOM. Thus, companies should actively manage negative reviews by responding appropriately and/or establishing partnerships with consumers to enhance their outputs (e.g. reputation, performance, sales). Third, the present study highlights, for companies, the importance of choosing the appropriate platform on which to promote their products/services. Consumers who use social media platforms (e.g. blogs) make stronger circumstance attributions than those who use other platforms (e.g. e-commerce websites). Finally, it has been shown that AI increases the effectiveness of eWOM. AI technologies (e.g. voice and virtual assistants) can capture customer preferences, humanize interactions and make experiences more intimate. Thus, they improve the customer experience.

5.1 Theoretical and practical implications

The present study makes several important contributions to the emerging eWOM literature. To the best of the authors’ knowledge, this is the first study built on the social communication framework to classify eWOM research papers in the travel field and to propose specific elements of eWOM for future research. Indeed, the studies that have reviewed the elements of eWOM communication did so in virtual settings following a general approach (Cheung and Thadani, 2012). In addition, most previous studies focused only on one or two consumer response variables and did not examine the interrelationships among eWOM’s key elements. Therefore, the present study explores theoretically how the characteristics of each key element (message, sender and receiver) affect the reader’s response variables (usefulness, trust, attitude, behavioral intentions and actual behaviors) in the travel context.

First, the study makes travel and tourism managers aware that user-generated eWOM messages are a rich source of data that may influence fellow readers’ behavioral intentions. Therefore, improving eWOM message features, by updating the information and making it more visual and attractive, can contribute to the effective management of businesses. Second, lack of source credibility causes psychological discomfort and, consequently, weak purchase intentions. Therefore, managers should take actions to foster credible reviews such as offering awards and/or privileged status to those users who provide pictures and/or videos to support their reviews and those with higher expertise. Third, individuals highly susceptible to interpersonal influence are more likely to purchase products/services that they perceive will improve their reputations in the eyes of others. Thus, practitioners should adopt strategies to use celebrities and/or opinion leaders to promote their products/services and reward loyal consumers by casting them as role models.

The present study makes several proposals as to future research. First, identifying the motivations for posting fake reviews might help managers design suitable defense strategies to improve relationships between companies and consumers. In addition, it was found that it is beneficial for practitioners to manage negative reviews. In this vein, managers should respond effectively to negative reviews and/or collaborate with consumers to understand their needs; in this way, they can turn unsatisfied customers into loyal customers and create committed relationships. Moreover, the important effect of the eWOM platform used on the reader’s behavioral intentions toward the products/services reviewed has been demonstrated. Managers should focus more on social media platforms (e.g. personal blogs) to promote their products/services, due to their extensive use of social cues and greater credibility. Companies should also introduce some social media features (e.g. self-disclosure, commenting, sharing and chatting) onto their websites to encourage consumers to generate eWOM. Finally, this research highlights that eWOM will increasingly be influenced by AI. Hotel managers should use AI to improve their customer experience journeys (pre-stay, during the stay and post-stay). AI can be integrated into the recommendations made by smart speakers such as Alexa and virtual assistants such as Nest Hub. Similarly, due to the COVID-19 outbreak, the demand for contactless technologies has increased. Thus, we are witnessing the massive introduction of AI technology into hospitality services (Table 2).

5.2 Limitations

While the findings of the present study are valuable, they need to be viewed in light of their limitations. The eWOM literature is very extensive and this study has not taken it all into account. For instance, this study focused on eWOM from the perspectives of communication theory and the consumer and did not analyze market-focused studies. Future studies should extend the literature review and increase the number of papers analyzed, based on different analytical perspectives. In addition, the keywords used to undertake the searches might have influenced the findings. However, it is reasonable to believe that the journal papers used are representative of the main eWOM communication-related research efforts in the travel sector. Finally, research into the impact of eWOM communication is continuously developing. It is strongly recommended that a meta-analysis be undertaken to improve our understanding of the impacts of the three elements (message, sender and receiver) on the consumer’s decision-making processes.

Figures

Figure 1.

Holistic framework of the influence of eWOM characteristics on the consumer’s decision-making processes in the travel sector

Table 1.

Other characteristics related to eWOM elements

Characteristics	Definitions	Associated eWOM element	Consequences	Previous works
Identity disclosure	The social identity that an individual establishes in an online community. A way of presenting oneself that helps others find one’s personal profile/ geographic location Kruglanski et al. (2005)	Sender	Perceived usefulness and credibility	Kruglanski et al. (2005), Sussman and Siegal (2003) and Liu and Park (2015)
Enjoyment	The extent to which the reading and understanding of reviews are perceived to be enjoyable in its own right, irrespective of any performance consequences (Davis et al., 1992)	Message	Perceived usefulness	Liu and Park (2015) and Venkatesh et al. (2002)
Review type	The orientation of a review e.g. recommendation vs attribute value information (Park and Lee, 2008)	Message	Purchase intention, perceived informativeness and persuasiveness (moderated by the consumer’s expertise)	Park and Kim (2008), Park and Lee (2008) and Xia and Bechwati (2008)
Review rating	The rating is given by the reviewer to a product/service (Lee and Lee, 2009)	Message	Perceived usefulness and attitude	Lee and Lee (2009)
Recommendation consistency	Whether the eWOM recommendation is consistent with other contributors’ experiences of the same product/service (Cheung et al., 2009)	Message	Perceived usefulness and review credibility (moderated by consumers’ involvement level)	Cheung et al. (2009)
Source type	The information source of a recommendation (e.g. consumer reports, friends and sales assistants) (Huang et al., 2009)	Sender and platform (medium)	Perceived informativeness and perceived usefulness	Huang et al. (2009)
Homophily	The degree to which pairs of individuals are similar in age, education and social status (Steffes and Burgee, 2009)	Sender and receiver	Behavior, trust and attitude	Steffes and Burgee (2009)
Involvement	Degree of psychological identification and affective, emotional ties the consumer has with a message (Park and Lee, 2008)	Receiver	Attitude and purchase intentions	Cheung et al. (2009); Doh and Hwang (2009); Lee et al. (2008), Park and Lee (2008) and Park et al. (2007)
Gender	Gender of the reviewers: male/female (Awad and Ragowsky, 2008)	Sender and receiver	Trust, perceived usefulness and purchase intentions	Awad and Ragowsky (2008) and Dellarocas et al. (2007)

Source: Own source

Table 2.

Main conclusions, theoretical and managerial implications

Conclusions	Theoretical and managerial implications
Valence, relevance, understandability and visual cues are the most important antecedents of message usefulness and the reader’s behavioral intentions	Making travel and tourism managers aware that the eWOM message is a rich source of data that may influence readers’ behavioral intentions Improving eWOM message features by updating information and making it more visual and attractive can contribute to the effective management of businesses
Source credibility is the sender characteristic that most affects the reader’s behavioral intentions	Lack of source credibility leads to psychological discomfort and, consequently, weak consumer purchase intentions Managers should encourage credible reviews by offering awards and/or privileged status to those users who provide pictures and/or videos to support their reviews and/or to those with higher expertise, etc.
Consumer susceptibility to interpersonal influence is the receiver characteristic that most influences their attitudes and behavioral intentions	Individuals susceptible to interpersonal influence are more likely to purchase products that they perceive will improve their reputations in the eyes of others Practitioners should recognize, in their marketing strategies, that consumer susceptibility to interpersonal influence is a key element in understanding different consumer types (e.g. celebrities and opinion leaders promoting products/services and rewards for loyal consumers)
More research into online fake reviews is needed to better understand sender motivations	Understanding senders’ motivations for posting fake reviews can hcelp managers implement appropriate defense strategies and improve customer-company relationships
Companies should actively manage negative reviews	Appropriate management of negative reviews (e.g. responding appropriately, developing a partnership with consumers) helps to: Understand consumers’ needs and reduces negative comments Turn unsatisfied customers into loyal customers Create committed relationships between companies and consumers
Managers should carefully choose the platforms on which their products/services are promoted	Managers should focus more on personal blogs due to their higher credibility and use of social cues
eWOM will be increasingly influenced by AI.	AI technologies can improve the customer experience (pre-stay, during the stay and post-stay) AI can be integrated into the recommendations provided by smart speakers such as Alexa and virtual assistants such as Nest Hub (contactless technology is increasingly in demand as a result of COVID-19)

Source: Own source