Repeat Purchasing in a Social Network Setting

• Activity Number: 145
• Type: Invited
• Date/Time: Monday, August 1, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics and Marketing
• Author(s): Peter Fader*+ and Eric Schwartz and Renana Peres
• Companies: University of Pennsylvania and University of Pennsylvania and Hebrew University
• Keywords: social network ; agent-based modeling ; stochastic models ; repeat purchasing ; customer attrition ; diffusion models

Abstract

We study this link between repeat purchasing, customer attrition, and social network structure/activity. Using an agent-based model, we simulate the initial diffusion, ongoing repeat usage, and ultimate customer attrition for a new product in a social network setting. The adoption, repeat purchasing and attrition decisions of individuals in the network depend on the communication dynamics within the social network, and are described using a well-established "buy till you die" probability model that has been augmented to account for network characteristics. To obtain the parameters used in the agent-based model (and to establish the overall validity of our proposed approach), we first estimate this type of model using actual data from a new application on Facebook. We then expand upon this empirical model using agent-based techniques, in order to explore how the overall long term profitability of the network depends on the repeat-purchasing/attrition parameters, and the sensitivity of repeat purchasing/attrition to the network characteristics. We also demonstrate the biases created when using a plain repeat-purchase model that ignores the social network characteristics.