Bike-sharing systems are changing the urban transportation landscape; we have been working with New York City Bike Share (aka Citi Bike), using analytics and optimization to improve the management of their system. Huge rush-hour usage imbalances the system, and in this talk we focus on methods used to mitigate the imbalances that develop. In particular, we will focus on the use of incentives; we have helped guide the development of Bike Angels, which enlists users to make “rebalancing rides”, and we will describe tradeoffs among a number of policies for determining when and where rides should be incentivized, all of which are based on a user dissatisfaction function model of the performance of the system. Motivate, the operator of Citi Bike, has thus far exported this incentive program to Ford GoBike in San Francisco, Bluebikes in Boston, and Capital Bikeshare in DC. The more recent incentive results are joint work with Hangil Chung and Daniel Freund, but the basis for much of the research presented is also joint with Shane Henderson and Eoin O’Mahony.