Understanding and predicting what influence online product sales? A neural network approach

Abstract Understanding the factors that influence sales is important for online sellers to manage their supply chains. This study aims to examine the roles of online reviews and reviewer characteristics in predicting product sales. With Amazon.com data captured using our big data architecture, this study performs sentiment analysis to measure the sentiment strength and polarity of review content. The predicting powers of sentiment together with other variables are then examined using neural network analysis. The results indicate that all the proposed variables are important predictors of online sales, and among them helpful votes of reviewer and picture of reviewer are the most influential ones. The findings of this study can be helpful for online sellers to manage their businesses, and the big data architecture and methodology can be generalised into other research contexts.