We treated electronic word-of-mouth (eWOM) behavior as a multidimensional construct and investigated the salient predictors for each individual eWOM behavior in this study, with an emphasis on hotel attribute performance. The results show that hotel attribute performance, previous eWOM experience, and platform convenience are the drivers of eWOM behaviors in general. Demographics and lodging preference do not affect eWOM behaviors. Economic incentives also have no effect on eWOM behaviors, which suggests that good reviews cannot be bought. The performance of core attributes can improve eWOM activities and eWOM praise, but the performance of facilitating attributes helps only with eWOM praise.

Introduction

Due to the intangibility and experiential nature of services, consumers seek additional information to reduce the uncertainty and complexity involved in making a purchase decision (Litvin et al., 2008). Electronic word-of-mouth (eWOM) has become a major source of information because of its (1) enhanced volume, (2) dispersion, (3) persistence and observability, (4) anonymity and deception, (5) salience of valence, and (6) community engagement (King et al., 2014). Serra Cantallops and Salvi (2014) conducted an extensive review of the eWOM literature and categorized it into two lines of research: antecedents of eWOM and influences of eWOM. However, the extant literature focuses heavily on the influences of eWOM. King et al. (2014: 172) called for studies to “uncover various antecedents of reviewwriting behavior” to develop feasible eWOM strategies. Specifically, Serra Cantallops and Salvi (2014) pointed out that only a few studies concentrate on factors that lead to eWOM behaviors in the hotel business. In response to this call, we aim to investigate the role of core and facilitating attributes in predicting eWOM behavior in the hotel industry. We focused on hotel attributes because most of the studies on eWOM antecedents concentrate on second-order predictors, such as satisfaction (e.g., Bronner and de Hoog, 2011; Cheung and Lee, 2012). However, these second-order predictors are often a collected reflection of first-order predictors, such as product quality, service performance, previous experience, and individuals’ economic and psychological characteristics (Cheung and Lee, 2012; Liang et al., 2013). Studies have shown that eWOM behaviors may be influenced directly by product or service attributes, instead of by customers’ perceived satisfaction (e.g., Chaniotakis and Lymperopoulos, 2009; Zhang et al., 2014). If firstorder predictors are not included in the model, then the estimation of the second-order predictors could be biased. The results of previous studies may not apply to the hotel industry as Ennew et al. (2000) and Harrison-Walker (2001) found that WOM behaviors could be industry dependent. We delineated hotel attributes into core attributes and facilitating attributes because the experience of a hotel stays consists of tangible products and intangible services (Millar and Baloglu, 2011; Slevitch and Oh, 2010). This approach can yield a complete and accurate understanding of the attribute performance–eWOM behavior relationship. The present study departs from most eWOM studies in that we examined the eWOM predictor–behavior relationship by treating eWOM behavior as a multidimensional construct. Most eWOM behavior studies treat eWOM behavior as a single-dimension construct (e.g., Boo and Kim, 2013; Brown et al., 2005; Tsao and Hsieh, 2012) while the WOM literature suggests that WOM behavior is likely to be a multidimensional construct (Harrison-Walker, 2001). We argue in the hypothesis development section that eWOM behavior, similar to traditional WOM behavior, is a multidimensional construct. We also proposed in this study that eWOM behaviors can be foretold using first order predictors, such as the existence of service attributes, instead of relying on second-order predictors, including satisfaction. Theoretically, the results confirmed eWOM behavior is a multidimensional construct as suggested by previous studies and support the idea that the performance of core attributes and the performance of facilitating attributes have different influences on eWOM behaviors. This study also helps identify critical aspects in hotel operation that managers should focus on to encourage positive eWOM behaviors from customers, which will enhance the property’s online presence.