We undertake a randomized evaluation of “Pathways to Choice”, which provides mentored girls’ clubs, life skills, and vocational training to empower adolescent girls to delay marriage and pursue education in Northern Nigeria. Two years post-intervention, adolescent girls in treated communities are 65 percentage points less likely to be married, estimates an order of magnitude larger than comparable interventions. An important channel is the program’s effects on educational uptake, which increases by 69 percentage points among the treated. However, the effects on education are themselves insufficient to fully explain the effects on child marriage, suggesting the bundled nature of the program is essential to its success. We argue that the whole community focus of the program reduces the likelihood of social backlash, allowing Pathways to produce large effects on entrenched, normative behavior.