B1111
Title: Modelling customer journeys with transformers
Authors: Luuk van Maasakkers - Erasmus University Rotterdam (Netherlands) [presenting]
Dennis Fok - Erasmus University Rotterdam (The Netherlands)
Bas Donkers - Erasmus University Rotterdam (Netherlands)
Abstract: A customer journey is a sequence of customer- and firm-initiated actions that may or may not end in a particular target event, such as a purchase being made by the customer. Firms collect data on these customer journeys by tracking online clicks and monitoring other customer behavior. This data can be used to learn how firm-initiated actions affect customer behavior, which enables a firm to optimize its marketing efforts. We propose a neural network-based method to model the dynamics in individual customer journeys over time. More specifically, we model the next action of a customer as a function of all prior touchpoints, both firm- and customer-initiated. Not only the type of the next touchpoint is modelled, but also the time until the next customer-initiated action. While much of the previous literature has modeled the sequence of all touchpoints, both firm- and customer-initiated, our model solely aims at predicting customer-initiated touchpoints; firm-initiated touchpoints only serve as explanatory input. In this way, we can estimate the effects of the type and timing of firm-initiated touchpoints on the customer's behavior. This enables us to optimize the firm actions, as we can simulate how customers would respond to new firm policies.
