Product Comparison Networks for Competitive Analysis of Online Word-of-Mouth

Abstract

Enabled by Web 2.0 technologies social media provide an unparalleled platform for consumers to share their product experiences and opinions---through word-of-mouth (WOM) or consumer reviews. It has become increasingly important to understand how WOM content and metrics thereof are related to consumer purchases and product sales. By integrating network analysis with text sentiment mining techniques, we propose product comparison networks as a novel construct, computed from consumer product reviews. To test the validity of these product ranking measures, we conduct an empirical study based on a digital camera dataset from Amazon.com. The results demonstrate significant linkage between network-based measures and product sales, which is not fully captured by existing review measures such as numerical ratings. The findings provide important insights into the business impact of social media and user-generated content, an emerging problem in business intelligence research. From a managerial perspective, our results suggest that WOM in social media also constitutes a competitive landscape for firms to understand and manipulate.

An influence framework on product word-of-mouth (WoM) measurement

“…The polarity value of the sentence is calculated by the algorithm shown in Table 1. As suggested by Zhang et al [22], a positive score implies that p1 is superior to p2 and vice versa. A direct link from p2 to p1 is produced if the linguistic context conveys a considerable amount of dominating positivity:…”

“…The product-comparative network is used to identify the comparative opinions between products. It proposes an intuitive illustration about product-comparative relations for competitive intelligence [22]. Different from other studies that emphasize structural-level findings, Zhang et al [19] used the comparative network as product WoM indicators and predicted the sales rank.…”

“…In the product network, following the research of Zhang et al [22], each node represents a product and a direct link indicates a comparative relationship between two products. At the same time, the corresponding weight on each link indicates the sentiment strength of the comparison relationship.…”

Crowd intelligence: Analyzing online product reviews for preference measurement

“…First, this study contributes to online product review research in information systems (IS) field. Most of IS literature on product reviews investigate how online product reviews affect product sales (Chen & Xie, 2008;Forman, M a n u s c r i p t 27 Ghose, & Wiesenfeld, 2008;Kuksov & Xie, 2010;Zhang, Guo, & Goes, 2013) or examine the antecedents of the helpfulness of product reviews (Mudambi & Schuff, 2010;Korfiatis, García-Bariocanal, & Sánchez-Alonso, 2012;Yin, Bond, & Zhang, 2014). Our current study extends the scope of IS research on online product reviews and propose effective and viable methods for converting a vast amount of data (i.e., online product reviews) into useful business intelligence (i.e., preference measurement).…”

A Convolution Kernel Approach to Identifying Comparisons in Text

“…Others are interested in yet another task to identify the direction of the comparisons (Ganapathibhotla and Liu, 2008;Tkachenko and Lauw, 2014), or the aggregated ranking (Kurashima et al, 2008;Zhang et al, 2013;Li et al, 2011). Our task precedes these tasks in the pipeline.…”