A Model for Estimation of Anonymous Visits on Websites

• Activity Number: 129
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Marketing
• Author(s): Julie Novak*+
• Keywords: Bayesian Statistics ; Missing Data ; MCMC

Abstract

Tailoring content to users has become a hallmark of digital media and marketing, however doing so requires that users identify themselves, or have the potential to be digitally watched, across usage or purchase occasions. However, across a wide variety of contexts, companies find that users or customers do not consistently identify themselves, leaving a substantial fraction of anonymous visits. In this paper, we explore the question of how these anonymous visits should be handled by companies wishing to practice targeted marketing. Specifically, we develop a model that allows us to probabilistically assign anonymous sessions to users. We demonstrate the importance of avoiding the popular practice of ignoring (case-deleting) the anonymous visits both because it decreases efficiency and because of the potential for bias due to the non-ignorability of which types of events are more likely to be anonymous. Our model is implemented using Markov Chain Monte Carlo simulation.