Marketing & Policy Studies-Faculty Candidate Seminar

Abstract

A Flop or a Hit?: Modeling the Impact of  the Evolution Of Consumer Reviews on Movie Success

Consumer reviews play an increasingly important role in the success of new products. Although extant literature has advanced our understanding of this recent phenomenon, researchers are divided on the impact of consumer reviews on sales.  We propose that these differences are caused by the failure to fully account for the evolution of the reviews. We quantify this evolution in consumer reviews on three dimensions: assessment of quality (average valence), interest (volume of reviews), and disagreement (dispersion of valence across reviews).  In this paper, using 511,972 online movie reviews from IMDB on 405 movies released from 2005 to. 2010 and using functional data analysis, we seek answers to the following questions: How do consumer reviews evolve? More specifically, does the overall assessment of quality (valence) change over time? Does the consumer interest (volume) change over time? And does the uncertainty surrounding the quality (dispersion) change over time? How do valence, volume, and dispersion affect one another, and how do marketing actions affect the evolution of these metrics? Does the evolution of reviews affect sales, and which metric can help predict market success? We conclude by discussing how managers can influence the impact of consumer reviews on product success.

Keywords: Consumer reviews; Valence; Volume; Dispersion; Functional Data Analysis; Functional Principal Components; Functional Clustering; Spline Regression