Understanding influencers and influential content in electronic word of mouth communication: an analytics approach

Abstract

The advent of web 2.0 has brought significant changes in the shopping behaviours of customers. Particularly interesting and important is the rise of internet enabled word of mouth communication, suitably refered to as Electronic Word of Mouth ( E-WOM). With the growing use and populairty of online modes of communication, E-WOM is increasingly complementing the traditional word of mouth communication, in shaping purchase decisions of consumers. The current dissertation is based on online consumer reviews, which arguably constitute the mose effective channel for consumers' voice. The importance of reviews has been highlighted in both academic research and industry reports. It has been reported, for instance, that consumer reviews are the most trusted source of information for consumers, next only to the direct recommendations made by family ans friends. Similarly, based on survey of consumers from USA and Canada, is was reported that 88% of consumers have read online reviews evaluate a local business and about 40% of them do so on a regular basis. Likewise, based on survey, it was reported that reviews impact purchase decisions for 93% of consumers. Furthermore, it was found that two-thirds of consumers were willing to pay more if they were assured of a better experience. It has therefore become imperative for businesses to monitor online reviews written for its product(s). Correspondingly, it is important for academic researchers to understand the mechanism underlying the influence of reviews on readers. Extant literature in the field of online review influence has several important gaps. First, it has not examined the possibility of change in impact of drivers of influence over time, as more reviews are posted. Second, temporal changes in behaviour of online reviews and the corresponding impact on review influence has not been examined. And third, topical composition of reviews and its impact on review influence has not been considered. Research studies conducted as part of this dissertation intend to address these gaps. In first study, using an exploratory paradigm, temporal changes in impact of drivers of influence has been observed. Likewise, in second study, temporal changes in content creation behaviour of online reviews has been noted. Further, using a predictive modelling approach, the importance of content creation behaviour as an early signal predicting influential reviews has been established. And in third study a novel framework, based on Latent Dirichlet Allocation (LDA), a popular topic modelling method has been proposed and tested to explore topical composition of a set of reviews and determine the impact of individual topics on review influence. Online review data from Yelp was used in all three studies. Finding from the studies have important implications for theory and practice, and have been discussed.