You are on page 1of 5

A BRIEF REVIEW OF LITERATURE AND IDENTIFICATION OF SOME KEY RESEARCH QUESTIONS

RELATED TO ONLINE REVIEW DYNAMICS

Word-of-mouth (WOM) is often considered to be an important driver of product sales because of its perceived reliability as a source of information. With the advent of internet, word-of-mouth has transcended the boundaries of closely knit peer groups to much larger consumer networks. This trend has been facilitated by the emergence of online technologies that enables the individual consumers to share their experiences and opinions to the marketplace through discussion groups, product reviews, and blogs. The exponential growth of online opinion and consumer-review sites has resulted in a deluge of freely and easily accessible information. This new form of WOM has brought with it mixed blessings to the market place. On one hand, consumers are getting directly involved in the generation of word-of-mouth and the potential consumers are getting the benefit of learning from others experiences; while on the other hand, the marketers are increasingly facing a diminishing level of control over this communication process. Given its increasing popularity and enormous impact on consumer decision making process, online review dynamics have attracted the attention of marketing scholars. Consequently, this field of study has produced a significant body of research. A number of papers (Chevalier and Mayzlin 2006, Dellarocas et al. 2007, Chintagunta et al. 2010, Liu 2006, Duan et al. 2008, Moe and Trusov 2011, Zhu and Zhang 2010) have investigated the impact of online reviews (both in terms of valence and volume) on consumer choice. On the other hand, a number of researchers have examined the evolutionary dynamics of online reviews (Godes and Mayzlin 2004, Moul 2007, Hu et al. 2008, Duan et al. 2008, Zhu and Zhang 2010, Moe and Schweidel 2011, Godes and Silva 2011).

Despite this rapid growth of scholarly work, a number of questions remain open in this field. One such issue that deserves further attention is how pre-purchase expectation and post-purchase evaluations relate to subsequent rating decisions of an individual in terms of incidence (whether to contribute) and evaluation (what to contribute). Expectations-confirmation theory posits that expectations, coupled with perceived performance, lead to post-purchase satisfaction. If a product falls short of expectations (negative disconfirmation) the consumer is likely to be dissatisfied (Oliver, 1980; Spreng et al. 1996). The extant literature suggests that the prepurchase evaluation provides a benchmark against which the actual product experience is compared (Anderson and Sullivan 1993). A similar sentiment is echoed by expectationconfirmation theory that hypothesize that expectations serve as the comparison standard that consumers use to evaluate performance and form a disconfirmation judgment (Halstead, 1999). Disconfirmation, in turn, affects satisfaction, with positive disconfirmation leading to satisfaction and negative disconfirmation leading to dissatisfaction (Churchill and Surprenant 1982, Woodruff et al 1985, Swan 1988, Tse and Wilton 1988, Oliver and DeSarbo 1988, Cooper et al 1989, Vezina and Nicosia 1990, Bolton and Drew 1991). The service quality model (Berry et al 1985) is also based on the view that service quality results from customers comparing their expectations prior to receiving service to their perceptions of the service experience itself (Parasuraman et al 1985, Brogowicz et al 1990, Grnroos 1990, Berry and Parasuraman 1991, Haywood-Farmer and Nollet 1991). Given this widely accepted view that the quantum of difference between pre-purchase expectations and post-purchase experience has a definitive impact on the satisfaction / dissatisfaction and hence the future course of actions of individual consumers (Churchill and Surprenant, 1982), it would be logical to expect that expectationdisconfirmation would play an important role in dictating an individuals posting behavior in

terms of whether to contribute (incidence) and what to contribute (evaluation). This proposition is also supported by Moe and Schweidel (2011). In the present work we investigate this issue by developing an integrated model of individual posting behavior that takes into account the prior (pre-purchase) expectations of consumers. Another issue that we address in this paper is the impact of subjective (textual) content on individual rating behavior. A review of present literature reveals that most of the existing work in this area has focused on how objective aspects of online rating environment (product ratings) individual posting behavior. However, a large amount of information in this environment is disseminated through subjective medium (textual content). There is strong need to generalize the existing frameworks to incorporate the effect of the textual content (Moe and Schweidel 2011). This is especially important for experiential products which consumers choose, buy and use solely to experience and enjoy. The consumption experience, especially its hedonic and aesthetic aspects, is often better communicated in words rather than numbers. Marketing scholars have long recognized that consumers perceptions of the subjective characteristics exert an important influence on their product evaluations (e.g. Srinivasan, Lovejoy, and Beach 1997, Yamamoto and Lambert 1994).Consequently, ignoring the textual component of the online content would result in an incomplete and biased understanding of the online rating dynamics. A number of recent studies have emerged documenting the presence of noticeable opinion dynamics in online product ratings (Godes and Silva 2011, Li and Hitt 2008, Moe and Trusov 2011). These studies suggest that an individuals publicly expressed opinion can be influenced by the opinions of others and does not necessarily mirror the individuals socially unbiased and independent product evaluation. In a controlled experimental setting, Schlosser (2005) showed that an individual poster has a tendency to adjust his or her posted product evaluation after

viewing what others have posted. The present body of literature captures this adjustment effect in terms of a set of metrics that best describe previously posted ratings (Dellarocas and Narayan 2006). These metrics have focused on the valence, variance, and volume of posted product ratings. Valence is typically represented by average rating, variance has been measured using statistical variance measures as well as other dispersion measures such as entropy, and volume is simply captured as the number of posted product ratings. However, the interpretation of these metrics can be problematic when each metric is considered separately. In order to address this issue Moe and Schweidel (2011) consider the two-way interactions among these metrics in addition to their main effects. This approach, while definitely offering a more refined way of representing the influence of the characteristics of rating environment on individual posting behavior, does not completely capture the complexity of the process. We argue that the aggregative nature of these three metrics (valence, variance and volume) results in the loss of fine grained distributional information about the rating environment. In specific we posit that multiple distinct distributional patterns are possible for a given set of valence, volume and variance values. However, each of these distributional patterns might lead to drastically different adjustment effects. In order to handle this problem we propose a more dynamic framework that retains the fundamental distributional characteristics of the rating environment and tries to capture the process through which the individual consumer adjusts his rating so as to derive maximum utility from the his rating decision. This framework also enables us to explore a key issue that is yet to be examined by the extant literature: does a sustained bias in the direction of the conversation encourage (discourage) selective participation and consequently becomes less representative of the general customer base?

Figure 2 Proposed Conceptual Model of the Consumer Rating Process


Relevant Prior Experience
Experience Related To Diagnostic Cue 1

Rating Environment
Objective Evaluation

Experience related to Diagnostic Cue i

Subjective Evaluation Experience related to Diagnostic Cue n

Pre-Purchase Evaluation Incidence Decision Evaluation Decision

Purchase and Experience

Disconfirmation

Post-Purchase Evaluation

You might also like