Some development practices transform people’s lives; some do not. In the literature on women/gender and development, arguments are often made for or against a particular strategic approach: projects for women, gender focused projects or gender mainstreaming. The paper uses case studies of women-specific projects and of the institutionalisation of gender mainstreaming to argue that the way development is practiced is a more significant determinant of effective development outcomes than the choice of one strategy over another. It further argues that a concentration or range of strategies may be required. The conditions in which development may be transformational are explored. The need to start from the specificity and variability of each setting and to work in particular ways with all concerned is the starting point. It is argued that the constituent elements of transformative development are processual, interactive and emergent.