We evaluate a large, randomized UKaid Girls' Education Challenge project in rural Zimbabwe, where the staggered roll-out of the project allows us to isolate the impact of intervention components specifically intended to improve knowledge of and support for girls' education. Such information and engagement campaigns are often included as components of major development projects, but previous work has been unable to isolate their importance and effectiveness separately from the broader projects. We show that information and engagement campaigns increased mathematics performance and school enrolment. In later periods, the program was expanded to provide resources such as bicycles and books and to update curriculum. During these later periods, we observed improvements in literacy, but no additional improvements in mathematics and enrolment beyond what was observed following the information and engagement campaign alone.