We study mediated many-to-many matching in markets in which valuations evolve over time as the result of shocks, learning through experimentation, or a preference for variety. The analysis uncovers the key tradeoffs that platforms face in the design of their matching protocols. It shows that the...