Characterizing and Predicting Early Reviewers for Effective Product Marketing on Ecommerce Websites

Abstract

Online  reviews  have  become  an  important  source  of  instruction  for  users before manufacture an informed procure decision. Early reviews of a product tend to have a high  effect  on  the  ensuing  product  sales.  In  this  paper,  we  take  the  initiative  to  study  the behavior  characteristics  of  early  reviewers  through  their  posted  reviews  on  two  real-world large e-commerce  platforms,  i.e.,  Amazon  and Yelp.  In specific,  we  divide  product lifetime into three uninterrupted phase, namely early, majority and straggler. A user who has posted a review  in  the  early  stage  is  contemplating  as  an  untimely  observer.  We  quantitatively characterize early reviewers based on their rating behaviors, the helpfulness scores received from others and the correlation of their reviews with product popularity. We have found that (1) an early observer tends to assign a higher average rating score; and (2) an early observer tends to post more helpful reviews. Our analysis of product reviews also indicates that early reviewers'  ratings  and  their  received  helpfulness  scores  are  likely  to  influence  product popularity. By viewing review posting process as a multiplayer competition game, we present a novel margin-based embedding model for early reviewer divination. Extensive experiments on two different e-commerce datasets have shown that our proposed approach outperforms a number of aggressive baselines.  

Index-Terms: Early reviewer, Early review, Embedding model.