A Daily-level Purchasing Model at an E-commerce Site

Abstract

Recently, the e-commerce sites market in Japan is growing. Many companies in e-commerce sites compete for customers each other. Therefore, e-commerce managers need to find good customers who may benefit before these customers purchase at rivals. Some preceding studies discuss which visits lead to purchase. However, it is unclear when a customer visits. In our approach, we predict the day a customer will purchase. As a result, we found some customers who tend to purchase on the day after they viewed many product pages. We also found that product page viewing before purchasing may be one of good predictors.DOI:http://dx.doi.org/10.11591/ijece.v2i6.1816