Online footsteps to purchase: Exploring consumer behaviors on online shopping sites

As an important part of the Internet economy, online markets have gained much interest in research community as well as industry. Researchers have studied various aspects of online markets including motivations of consumer behaviors on online markets. However, due to the lack of log data of consumers' online behaviors including their purchase, it has not been thoroughly investigated or validated on what drives consumers to purchase products on online markets. Our research moves forward from prior studies by analyzing consumers' actual online behaviors that lead to actual purchases, and using datasets from multiple online shopping sites that can provide comparisons across different types of online shopping sites. We analyzed consumers' buying process and constructed consumers' behavior trajectory to gain deeper understanding of consumer behaviors on online markets. We find that a substantial portion (24%) of consumers in a general-purpose marketplace (like eBay) discover items from external sources (e.g., price comparison sites), while most (>95%) of consumers in a special-purpose shopping site directly access items from the site itself. We also reveal that item browsing patterns and cart usage patterns are the important predictors of the actual purchases. Using behavioral features identified by our analysis, we developed a prediction model to infer whether a consumer purchases item(s). Our prediction model of purchases achieved over 80% accuracy across four different online shopping sites.