Recent interest in animal cultures has been fuelled by high-profile reports of intra- and interpopulation differences in the behavioural repertoires of primates and cetaceans, consistent with the existence of socially learned traditions. Several studies have mapped spatial differences in behaviour, revealing a mosaic of behavioural phenotypes within species. The dominant current approach attempts to determine whether this is cultural variation by excluding asocial learning, ecological or genetic factors. However, claims of animal cultures remain controversial because such comparisons are subject to weaknesses; thus, new approaches to isolating the influence of culture on behaviour are required. Here we suggest that, rather than attributing behaviour to explanatory categories, researchers would often be better advised to partition variance in behaviour to alternative sources.