Blood-feeding (sanguivory) has evolved more than two dozen times among birds, fishes, insects, arachnids, molluscs, crustaceans, and annelids; however, among mammals, it is restricted to the vampire bats. Here, the authors revisit the question of how it evolved in that group. Evidence to date suggests that the ancestors of phyllostomids were insectivorous, and that carnivory, omnivory, and nectarivory evolved among phyllostomids after vampire bats diverged. Frugivory likely also evolved after vampire bats diverged, but the phylogeny is ambiguous on that point. However, vampire bats lack any genetic evidence of a frugivorous past, and the behavioural progression from frugivory to sanguivory is difficult to envision. Thus, the most parsimonious scenario is that sanguivory evolved in an insectivorous ancestor to vampire bats via ectoparasite-eating, wound-feeding, or some combination of the two—all feeding habits found among blood-feeding birds today. Comparing vampire bats with other sanguivores, the authors find several remarkable examples of convergence. Further, it was found that blood-feeding has been ca. 50 times more likely to evolve in a vertebrate lineage than in an invertebrate one. The authors hypothesize that this difference exists because vertebrates are more likely than invertebrates to have the biochemical necessities required to assimilate the components of vertebrate blood.