Essays on modeling consumer behavior in online shopping environments.

Abstract

With the proliferation of online retailing, marketers are presented with both unprecedented challenges and opportunities. On the one hand, consumer behavior in the online environment is drastically different from its offline counterpart. On the other hand, technology development has provided marketers with more innovative tools to study online consumer behavior: In this dissertation I investigate two aspects of consumer behavior in online shopping environments. First, I explore how online retailers can make accurate product recommendations to customers. Product recommendation systems are backbones of the Internet economy, leveraging customers' prior product ratings to generate subsequent suggestions. A key feature of recommendation data is that few customers rate more than a handful of items. Extant models take missing recommendation rating data to be missing completely at random, implicitly presuming that they lack meaningful patterns or utility in improving ratings accuracy. For the widely-studied EachMovie data, I find that missing data are strongly non-ignorable. Recommendation quality is improved substantially by jointly modeling selection and ratings, both whether and how an item is rated. Second, I examine online purchase behavior across multiple shopping sessions. Shopping cart abandonment is the bane of many e-commerce websites. Compared to the drop-off rate of 2-3% in brick-and-mortar stores, online shopping cart abandonment rates range from 25% to 75%. I investigate abandoned shopping carts in an online grocery shopping setting. Specifically, I develop a joint model for the cart, order, and purchase quantity decisions. The interdependence between the three decisions is captured by the correlations between the error terms. Empirical analysis shows that not all abandoned shopping carts result in lost sales. Customers routinely pick up abandoned carts and complete the final orders. Among the factors that propel customers to continue with aborted shopping are the time of shopping, time elapsed since the previous visit, the number of items left in the abandoned cart, and promotion intensity. The study offers marketers important managerial implications on how to mitigate the shopping cart abandonment problem.