Title: How to fairly share streams: Indices characterizations for music streaming platforms problems
Speaker: Di Feng（Dongbei University of Finance and Economics）
Time: 1.10 (Saturday), 14:00-15:00
Venue: Room 315, Gewu Building
Abstract:
We study revenue sharing in music streaming platforms through the axiomatic approach. Motivated by the shift from per-unit pricing to subscription-based access, we examine how platforms should distribute total subscription revenue among artists based on user streaming activity. Particularly, we focus on distribution in a fair way, due to concerns that have arisen in reality.
We propose a broad class of floored proportional indices (FP indices) that provides an appealing compromise between proportionality and equality, a feature absent in previous work and strongly demanded by the artist community. Notably this indices maintain the standard fairness notion, equal treatment of equals.
Then, by introducing two additional normative axioms: invariant index-summation under redistribution, which reflects proportional fairness, and pairwise redistribution invariance, which reflects fairness from a dynamic perspective, we characterize our FP indices through these three fairness axioms.
We also introduce and characterize a sub-class of it, weighted proportional indices (WP indices), by replacing equal treatment of equals with null artist. Finally, by imposing additional axioms -- anonymity for users and canceling-off across redistributions -- we identify two widely used indices in practice, pro-rata index and user-centric index, respectively.
Our framework provides a normative foundation for designing fair revenue-sharing methods in streaming platforms.
