Segment Discovery: Enhancing E-commerce Targeting

TL;DR

The paper tackles targeted interventions in e-commerce under practical constraints by proposing a two-stage framework that first estimates individual treatment effects via uplift/CATE models with unconfoundedness, and then solves a constrained optimization to assign treatments through a policy matrix $\\uppi$ that maximizes weighted uplift $\\sum_i\\sum_k\\pi_{ik} w_i \\hat{\\tau}_k(x_i)$. It contributes a generalized approach for custome r targeting, validated across retention and revenue applications using offline policy evaluation and a large online A/B test, with demonstrations that uplift-based policies outperform propensity-threshold baselines. The framework accommodates budget-like constraints and can be extended to continuous treatments and observational studies with sensitivity analyses when unobserved confounding is present. Collectively, the work provides a scalable, principled method for personalizing promotions and interventions in e-commerce, improving value to the business while maintaining customer experience.

Abstract

Modern e-commerce services frequently target customers with incentives or interventions to engage them in their products such as games, shopping, video streaming, etc. This customer engagement increases acquisition of more customers and retention of existing ones, leading to more business for the company while improving customer experience. Often, customers are either randomly targeted or targeted based on the propensity of desirable behavior. However, such policies can be suboptimal as they do not target the set of customers who would benefit the most from the intervention and they may also not take account of any constraints. In this paper, we propose a policy framework based on uplift modeling and constrained optimization that identifies customers to target for a use-case specific intervention so as to maximize the value to the business, while taking account of any given constraints. We demonstrate improvement over state-of-the-art targeting approaches using two large-scale experimental studies and a production implementation.

Figures (1)

• Figure 1: "True" uplift based on simple difference across different retention scores.