Approximating Purchase Propensities and Reservation Prices from Broad Consumer Tracking

Abstract

A consumer’s web-browsing history, now readily available, may be much more useful than demographics for both behaviorally targeting advertisements and personalizing prices. Using a method that combines economic modeling and powerful machine learning techniques, I find a striking difference. Using demographics yields purchase probabilities at observed prices ranging across individuals from about 8% to about 30%. Adding consumers’ web-browsing histories increases this range to about 5% to 90%, allowing more precise behavioral targeting. I further find that personalizing prices based on web browsing histories increases profits by 12.99% and results in some consumers paying substantially more than others for the same product. Using only demographics to personalize prices raises profits by only 0.25%, suggesting the percent profit gain from personalized pricing has increased 50-fold.

Keywords: First-Degree Price Discrimination, Price Discrimination, Big Data, Behavioral Targeting, Personalized Pricing, Algorithmic Pricing

JEL Classification: D42, L130