Migration is barely predictable: rapid population movements can happen with hardly any warning yet can bring about momentous challenges, as witnessed for example during the recent conflict in the Syrian Arab Republic. Still, in the context of recent global and regional migration trends and in the light of the UN 2030 Agenda for Sustainable Development, there is a strong policy demand for reliable forecasts of migration. This Data Briefing, produced by IOM’s Global Migration Data Analysis Centre, examines how best such forecasts might be produced and where the challenges lie in such endeavours. It reviews historical examples of forecasting migration and discusses the inherent limitations of predicting the movement of people. There are several reasons why this task is so challenging, including the multiple and unpredictable drivers of migration, the lack of comprehensive and reliable data, and the variations in concepts and definitions for different types of migration. The Data Briefing is a contribution to discussions on how to improve migration forecasting. It includes a consideration of creative new models and adopts a new philosophy in which the author argues that it is more realistic to approach forecasting through probability and use the information about the uncertainty of forecasts to aid decision-making. Using this strategy, the author suggests the “traffic lights approach”, which includes a range of possibilities that are attached to relevant levels of (un)certainty, and aim at supporting prudent policy decisions in migration-related areas.