Since its introduction in the late 2000s, there has been growing interest in sharing economy platforms. To explain outcomes, scholars have taken two main approaches—institutionalism, which focuses on employment classification and precarious labor, and technological control via algorithms. Both predict relatively similar outcomes for workers. On the basis of 111 in-depth interviews with workers on seven platforms (Airbnb, TaskRabbit, RelayRides, Uber, Lyft, Postmates and Favor) we find that because platform labor is weakly institutionalized, worker satisfaction, autonomy and earnings vary significantly across and within platforms, suggesting dominant interpretations are insufficient. The extent to which workers are dependent on platform earnings to pay basic expenses rather for supplemental income explains the variation in outcomes, with supplemental earners being more satisfied and higher-earning. This suggests platforms are free-riding on conventional employers. We also find that platforms are hierarchically ordered, in terms of what providers can earn, conditions of work, and their ability to produce satisfied workers. Our findings suggests the need for a new analytic approach to platforms, which emphasizes labor force diversity and connections to conventional labor markets.