Fraud-Proof Revenue Division on Subscription Platforms

We study a model of subscription-based platforms where users pay a fixed fee for unlimited access to content, and creators receive a share of the revenue. Existing approaches to detecting fraud predominantly rely on machine learning methods, engaging in an ongoing arms race with bad actors. We explore revenue division mechanisms that inherently disincentivize manipulation. We formalize three types of manipulation-resistance axioms and examine which existing rules satisfy these. We show that a mechanism widely used by streaming platforms, not only fails to prevent fraud, but also makes detecting manipulation computationally intractable. We also introduce a novel rule, ScaledUserProp, that satisfies all three manipulation-resistance axioms. Finally, experiments with both real-world and synthetic streaming data support ScaledUserProp as a fairer alternative compared to existing rules.

Online streaming platforms such as Spotify or Netflix collect monthly fees from users and distribute revenue to content creators—like musicians or filmmakers—based on user activity. But this setup is vulnerable: some content creators can try to game the system to unfairly earn more money, for example by using bots or fake accounts. Current methods to detect such fraudulent behavior rely heavily on machine learning, which often turns into a cat-and-mouse game with bad actors. Our research asks a different question: can we design payment rules that make cheating unprofitable in the first place?We propose three principles (or “axioms”) that any fair and manipulation-resistant revenue distribution system should follow. We then study which current methods meet these standards—and find that some commonly used approaches not only fail to prevent fraud, but also make it extremely hard to detect. To address this, we introduce a new method called ScaledUserProp, which discourages manipulation by design. We test our method using both synthetically-generated and real data from streaming platforms. The results show that ScaledUserProp distributes money more fairly and resists fraud better than existing systems.