Tractable Time Slot Assortment Optimization in Attended Home Delivery under Consider-Then-Choose Customer Choice

Abstract

In attended home delivery, the delivery time slot offering problem has a significant impact on the efficiency of the retailer providing this service. Since the decision on which time slots to offer is typically made online during the customer booking process, this problem requires a real-time solution.

However, solving this problem is complex since most companies offer multiple time slots over multiple days and must account for customer preferences. To adequately reflect customer choice, an appropriate customer choice model must be adopted. We employ the consider-then-choose choice model, which in the empirical literature has been shown to reflect the general underlying customer choice behavior well.

We address the time slot offering problem from a customer choice perspective. In particular, we propose a time slot assortment optimization model that exploits a customer cluster structure underlying the consider-then-choose choice model to solve the time slot offering problem in real-time. In addition, we propose a method for estimating the consider-then-choose choice model with such a cluster structure from historical transaction data. We evaluate this estimation method and the proposed time slot offering model in a numerical study and demonstrate that the estimation procedure can extract the underlying choice behavior and that the time slot offering problem can be solved in real-time for realistic problem sizes of an e-grocer.

Keywords: Consider-Then-Choose, Assortment Optimization, E-Grocery, Attended Home Delivery