Paths on a website: research on customers’ online search and purchase behaviours using clickstream data

Abstract

This dissertation is comprised of three chapters studying customers’ online search and purchase behaviours using clickstream data collected from the context of air travel.

The first chapter explores how choice overload effect influences customers’ purchase decisions under different time pressure conditions. Two studies yield two findings. The first finding is that larger choice sets result in purchase deferral, which is consistent with choice overload effects. Time pressure significantly moderates this effect, because more deferrals occur when the purchase deadline is further away. The second finding is that it is not the real, physical time passing per se that creates a sense of time pressure; instead, time pressure appears to be defined by customers’ perception of time limit, which moderates the choice overload effect by shifting customers’ regulatory focus.

The second chapter develops the modelling approaches of using path data to predict purchases. We develop the concepts of two types of sequence of browsing behaviours: the sequence of search strategies and the sequence of viewing behaviours. We find that viewing behaviour is a better indicator of purchase tendencies. We develop the modelling approaches of predicting the next viewing behaviour and using this predicted viewing behaviour to predict the purchase probability. Our approach improves current methods of predicting purchases by overcoming two disadvantages: inflexibility in adaption to different websites and missing detailed information of customers’ behaviours. We provide managerial insights on customising information shown to customers according to predicted viewing behaviours in order to improve purchase conversion rates.

The third chapter reveals the relationship between customers’ online search strategies and decision strategies. Our first finding is that the search strategy of filtering can be viewed as a decision strategy characterised by Elimination by aspects (EBA) strategy, while flexible searches can be viewed as a decision strategy featured by Satisficing strategy. Our second finding is that the goal-related variable has the predominate effect on choice of decision strategies in the studied context. Customers choose the decision strategy that can enable them to fulfil the goal of finding a good price, instead of the strategy that simplify the choice task.