We study ethnic discrimination in Europe’s largest carpooling marketplace. Using a unique dataset with more than 17,000 rides, we estimate the effects of drivers’ perceived name origins on the demand for rides. Carpooling is a novel application for studying ethnic discrimination where consumer choice entails social interaction with the service provider (i.e. driver). We find large discrimination effects for drivers with Arab, Turkish or Persian sounding names. Further analyses support assumptions consistent with statistical discrimination. Our findings broaden the perspective of ethnic discrimination by shedding light on subtle, everyday forms of discrimination in social markets and fuel ongoing discussions about anti-discrimination efforts in an era in which markets increasingly move online.