Anthropogenic environmental modification is placing as many as 1 million species at risk of extinction. One management action for reducing extinction risk is translocation of individuals to locations from which they have disappeared or to new locations where biologists hypothesize they have a good chance of surviving. To maximize this survival probability, the standard practice is to move animals from the closest possible populations that contain presumably related individuals. In an empirical test of this conventional wisdom, we analyzed a genomic dataset for 166 translocated desert tortoises (Gopherus agassizii) that either survived or died over a period of two decades. We used genomic data to infer the geographic origin of translocated tortoises and found that individual heterozygosity predicted tortoise survival, whereas translocation distance or geographic unit of origin did not. Our results suggest a relatively simple indicator of the likelihood of a translocated individual's survival: heterozygosity.