Best of Three Worlds: Adaptive Experimentation for Digital Marketing in Practice
This work tackles the problem of improving experimentation efficiency and reliability for digital marketers, but it appears incremental as it builds on existing AED methods by adapting them to real-world non-stationary conditions.
The paper addresses the challenges of using adaptive experimental design (AED) methods in non-stationary industrial settings, such as digital marketing, and presents a framework for counterfactual inference that was tested in a commercial environment.
Adaptive experimental design (AED) methods are increasingly being used in industry as a tool to boost testing throughput or reduce experimentation cost relative to traditional A/B/N testing methods. However, the behavior and guarantees of such methods are not well-understood beyond idealized stationary settings. This paper shares lessons learned regarding the challenges of naively using AED systems in industrial settings where non-stationarity is prevalent, while also providing perspectives on the proper objectives and system specifications in such settings. We developed an AED framework for counterfactual inference based on these experiences, and tested it in a commercial environment.