Online Product Ratings: Dynamic Trends and Diagnosticity Assessment Explanation

We investigate the dynamic trends of online product ratings and explanations by examining a sample of 2,595 online product ratings of 14 randomly drawn apparel products from landsend.com. Contrary to the predominant declining sequential and/or temporal trends reported in previous studies, we find a predominant increasing sequential trend. We extend the diagnosticity assessment explanation offered by Godes and Silva (2012) by considering how changes of reviewer similarity and review information may affect the diagnostic value of prior reviews in customer decision making, thus explain both the increasing trend in our sample and the declining trend in previous studies. We find support to the idea that diagnostic value of prior reviews is high, leading customers to make good purchase decisions and give higher reviews over time, when reviewer dissimilarity and information noise do not increase; When reviewer dissimilarity and information noise increase overtime, prior reviews may mislead and confuse customers, leading bad purchase decisions and thus a declining review trend.