The perpetuation of testing

This chapter makes several interventions to map the internal inconsistencies of the pay-for-success model. It argues that measuring changes in the lives of service users is far more complex and provisional than typically assumed in the impact bond discourse. Beyond demonstrating the effects of an intervention, actors must therefore also protect reported ‘results’ from contestation and carefully manage external scrutiny. The chapter further problematises the experimental logic underpinning impact bonds, showing that even when interventions are tested successfully—according to predefined metrics—they are never integrated into standard provision. Instead, success leads to further rounds of outcome-based contracting. Against this backdrop, the chapter argues that the mechanism is less about discovering ‘what works’ through experimentation and more about advancing a mode of governance defined by data-driven performance management and the disciplining of service providers.