The Influence of Context on Online Shopping Behavior: The Case of Concurrent TV Consumption

Abstract

It is widely agreed that people use digital devices while watching TV, a phenomenon called Screening. While previous research has shown that the diffusion of mobile devices can increase online sales, work on multitasking suggests that second screening may lead to lower sales. In this paper, we study the impact of concurrent TV viewership on online shopping behavior and try to understand the conflicting results. Our first study analyzes the impact of TV consumption on online shopping behavior at the aggregate level using a panel of 100,000 U.S. customers over a period of 24 months. We address potential endogeneity problems with an instrumental variable approach. Our second study proposes a novel approach to realizing individual-level analysis over thousands of consumers and products. We identify a sample of consumers who watched a certain TV show and analyze their Amazon shopping carts in the context of TV consumption. We then compare these results to the shopping behavior of the same sample of consumers one week before they watched the TV show. Both studies reveal that while TV consumption is correlated with a higher likelihood of online shopping in general, the causal effect of TV watching differs with respect to the complexity of the products purchased. If a TV program appeals to a large number of viewers, this results in fewer sales of high-complexity products (1% increase in TV consumption leads to -1% sales) and more sales of low-complexity products (1% increase in TV consumption leads to 6.5% sales). New insights on this effect and information about its magnitude could help marketers better allocate their TV ad spending and make better online shopping recommendations for online shoppers who concurrently consume TV.