Machine learning based prediction of consumer purchasing decisions : the evidence and its significance

Abstract

Every day consumers make decisions on whether or not to buy a product. In some cases the decision is based solely on price but in many instances the purchasing decision is more complex, and many more factors might be considered before the final commitment is made. In an effort to make purchasing more likely, in addition to considering the asking price, companies frequently introduce additional elements to the offer which are aimed at increasing the perceived value of the purchase. The goal of the present work is to examine using data driven machine learning, whether specific objective and readily measurable factors influence customers’ decisions. These factors inevitably vary to a degree from consumer to consumer so a combination of external factors, combined with the details processed at the time the price of a product is learnt, form a set of independent variables that contextualize purchasing behaviour. Using a large real world data set (which will be made public following the publication of this work), we present a series of experiments, analyse and compare the performances of different machine learning techniques, and discuss the significance of the findings in the context of public policy and consumer education.