The Customer Journey as a Source of Information

Marketing

Columbia Business School Research Paper

Padilla N;Ascarza E;Netzer O

Nicolas Padilla

2019

In high involvement purchases such as flights, insurance, and hotel stays, firms often observe at most only a handful of purchases during a customer lifetime. The lack of multiple past purchases presents a challenge for firms interested in inferring individual preferences. Moreover, customers in these industries often search for products that satisfy different needs depending on the purchase context (e.g., flights for a family vacation vs. flights for a business trip), further complicating the task of understanding what a customer might prefer on the next purchase occasion. Fortunately, these settings also collect other pieces of information; prior to a purchase, firms often have access to rich information on the customer journey, over the course of which customers reveal their specific preferences as they search and click on products prior to making a purchase. The objective of this paper is to study how firms can combine the information collected through the customer journey together with historical data to infer the customer’s preferences and likelihood of buying. We build a non-parametric Bayesian model that links the customer’s search queries and clicks over the course of a journey, as well as across journeys, with the customer’s history of purchases. The model accounts for what we call context heterogeneity, defined as journey-specific preferences that depend on the context in which the journey is undertaken. We apply our model in the context of airline ticket purchases using data from one of the largest travel search websites. We show that our model is able to accurately infer preferences and predict choice in an environment characterized by very thin historical data. We find strong context heterogeneity across journeys, reinforcing the idea that treating historical journeys as reflecting the same set of preferences may lead to erroneous inferences.

Customer Journey; Bayesian Non-parametrics; Clickstream Data; Customer Search

Columbia Business School Research Paper

No

View document