Over the last decade, the idea of societal impact resulting from publicly funded research has changed from being a relatively fringe concern related to high-technology entrepreneurship and spin-off companies to becoming an increasingly important public policy concern. This has stimulated academic science policy research to investigate the impact of research and conceptualize how knowledge created in academic contexts can be coupled to real-world problems or needs. Most of the work in this stream of research focuses on "ex post" impacts, that is, the impacts created by individual research activities or research strands after their completion and likewise there has been much research on "ex post" impact evaluation. However, "ex ante" impact evaluations have become increasingly important for funding decisions, but little is known about how to evaluate impact when considering research project proposals. In this article, we propose a conceptual framework to evaluate "ex ante" impact of research based on the idea that a research proposal is a 'promise' to achieve impact. We suggest that evaluators could assess social impact promises by considering two elements, namely the activities that couple their knowledge with non-academic users and the interdependency and consistency of such activities throughout the overall project proposal. We ultimately propose an analytical framework for refining our 'openness' conceptual framework in future empirical research.