Modeling consumer ratings: The trade-off between quality and consumer preference

Abstract

• An overview of different methods of ratings analysis with a focus on the quality and preference components of these methods. • A measurement framework to analyze the relative quality and preference components in consumer ratings. • A Monte Carlo simulation to test the measurement framework. • An empirical case study analysis of real-world online reviews. • Recommendations for managers on how to utilize the framework to improve online review platforms and marketing strategy Is there such a thing as a good quality product or service or is quality purely a result of the heterogeneity in preferences of different consumers? This article looks into the trade-offs between quality and preference aspects of consumer ratings from a measurement perspective. First, we give an overview of different methods of ratings analysis with a focus on the quality and preference components of these methods. We develop a measurement framework and associated methods to analyze the relative quality and preference components in consumer ratings. This framework is tested using a Monte Carlo simulation on generated data and using an empirical case study analysis of real-world online reviews. The framework shows a good ability to distinguish between quality-based and preference-based ratings and the case study results show good face validity. In particular, an entropy-based measure gives a consistent quality-preference scale on a bounded [0, 1] range and an information-theoretic interpretation. We give suggestions for future academic work on analyzing quality and preference trade-offs in ratings analysis. We give recommendations for managers on how to utilize the framework to improve online review platforms, on whether to analyze consumer ratings using traditional preference-based methods or more quality- or knowledge-based methods, and on how to improve STP (segmentation, targeting, and positioning) strategies.