This paper examines whether the chances of job placements improve if unemployed persons are counselled by caseworkers who belong to the same social group, defined by gender, age, education, and nationality. Based on an unusually informative dataset, which links Swiss unemployed to their caseworkers, we find positive employment effects of about 4 percentage points if caseworker and unemployed belong to the same social group. Coincidence in a single characteristic, e.g. same gender of caseworker and unemployed, does not lead to detectable effects on employment. These results, obtained by statistical matching methods, are confirmed by several robustness checks.