The underrepresentation of women in electoral politics and parliaments remains one of the biggest challenges to equal representation in Western Democracies. In light of cumulating evidence that voters do not discriminate against female candidates, parties’ recruitment and selection strategies have come under increasing scrutiny. Given the increasing pressure on parties to promote women, the question arises when, where and why they choose (not) to do so. Focusing on female candidate nominations in UK general elections from 1992 to 2017, we argue that party selectorates are more likely to nominate women where local public opinion is in favor of gender equality, and most consistently so where electoral incentives to follow public opinion are strongest: in competitive races for marginal seats. Using a novel combination of candidate, constituency, census and survey data, we put these arguments to the test and show that the interplay of public opinion and local power constellations in electoral districts indeed systematically predicts the nomination of female candidacies. Our findings uncover notable indirect effects of public preferences on female candidacies that operate through the strategic nomination choices of party selectorates and thereby yield important implications for the literatures on representation and candidate selection.