Experiments on ethnic or racial discrimination often use names to signal originin market situations under the assumption that non-ethnic name signals do notsystematically affect the outcome. Using survey data, we demonstrate significantvariation in perceived trustworthiness and professionalism of names within origins.Linking this to a large-scale name-matched field experiment, we show thatthe variation in name perception predicts discriminatory behavior: Prospectivetenants in Switzerland with names exogenously rated as less trustworthy or lessprofessional receive fewer invitations to an apartment viewing. These marketrelevantcharacteristics explain, to a large extent, group differences in measureddiscrimination. Ethnic minority names that score high on either trustworthinessor professionalism are not discriminated against, while names scoring low on theseare discriminated against (relative to the majority name). In line with intergroupcontact theory, we also show that, regardless of their origin, more common namesface less discrimination. Our results reveal that discriminatory behavior is muchmore nuanced than a simple in-group/out-group dichotomy would suggest.