Doubly Adaptive Cascading Bandits with User Abandonment

A central task of digital marketing is to promote user engagement. However, overexposure to marketing activities, especially those with irrelevant content, can result in customer dissatisfaction and ultimately can lead to user abandonment (e.g., unsubscribing from mailing lists or deleting an app). Motivated by this phenomenon, we focus on an online learning problem where a platform interacts with its users by sending them a list of messages over time. The platform earns a reward whenever a user accepts a message and is penalized when a user abandons after being targeted with irrelevant content. The setting extends the popular “cascading bandits” by allowing user abandonment that is triggered by low-quality or irrelevant recommendations. Moreover, instead of a single instantaneous interaction, we explicitly model multiple interactions that an individual has with the platform over a period of time. Thus, we focus on a type of doubly adaptive algorithm that not only updates its learning across users but also is capable of dynamically adjusting the sequential content for a given user. We refer to the online learning task as doubly adaptive cascading bandits (DAC-Bandit). For the offline combinatorial problem, we prove a polynomial-time algorithm. For the online setting, we investigate both the noncontextual and contextual problems and quantify the performance of the proposed algorithms through regret analysis. Moreover, we discuss other factors that can trigger user abandonment behavior, such as boredom from similar messages. The numerical performance of our algorithms is evaluated using both synthetic and real-world data sets. Our algorithms demonstrate strong theoretical performance guarantees and promising empirical results when compared with benchmarks.

Supplemental Material: All supplemental materials, including the code, data, and files required to reproduce the results, are available at https://doi.org/10.1287/opre.2024.1449.