Think Twice Before You Buy! How Recommendations Affect Three-Stage Purchase Decision Processes

Consumer decision-making is usually modeled as a two-stage process of initial screening and subsequent in-depth consideration of attractive alternatives. Recent evidence indicates, however, that consideration is not necessarily the direct precursor of choice: consumers may narrow their consideration sets further to the choice set. We examine how choices in a three-stage purchase decision process evolve by observing consumer behavior in an online shopping experiment. Specifically, we examine the effects of system- and user-generated recommendations (SGR and UGR) moderated by gender. Our contribution to information systems research is threefold. First, we suggest a new experimental design for observing the stages in purchasing processes. Second, we show that effects of SGR and UGR indeed vary between stages. UGR reduce consideration set size and increase females’ choice probability while SGR reduce males’ transition probabilities. Third, our results suggest that omitting choice set formation can lead to incorrect estimates of choice probabilities.