Reminds me of when we removed the “Previous Track” button on the Rdio player controls. The data showed that users skipped to previous only a tiny fraction of th

Reminds me of when we removed the “Previous Track” button on the Rdio player controls. The data showed that users skipped to previous only a tiny fraction of the time. Lasted about 3 days in production. Underlying this fallacy is the assumption that usage should somehow be uniformly distributed across all features. In reality, a power law distribution is more natural.