Social networks are known to influence migration decisions, but connections between individuals can hardly be observed. We rely on individual-level surveys conducted by Gallup in 147 countries that provide information on migration intentions and on the existence of distance-one connections for all respondents in each of the potential countries of intended destination. The origin-specific distribution of distance-one connections from Gallup closely mirrors the actual distribution of migrant stocks across countries, and bilateral migration intentions appear to be significantly correlated with actual flows. This unique data source allows estimating origin-specific conditional logit models that shed light on the value of having a friend in a given country on the attractiveness of that destination. The validity of the distributional assumptions that underpin the estimation is tested, and concerns about the threats to identification posed by unobservables are substantially mitigated.