Interspecific competition is an important driver of community assembly in plants and animals, but phylogenetic evidence for interspecific competition in bacterial communities has been elusive. This could indicate that other processes such as habitat filtering or neutral processes are more important in bacterial community assembly. Alternatively, this could be a consequence of the lack of a consistent and meaningful species definition in bacteria. We hypothesize that competition in bacterial community assembly has gone undetected at least partly because overly broad measures of bacterial diversity units were used in previous studies. First, we tested our hypothesis in a simulation where we showed that how species are defined can dramatically affect whether phylogenetic overdispersion (a signal consistent with competitive exclusion) will be detected. Second, we demonstrated that using finer-scale Operational Taxonomic Units (OTUs) (with more stringent 16S rRNA sequence identity cutoffs or based on fast-evolving protein coding genes) in natural populations revealed previously undetected overdispersion. Finally, we argue that bacterial ecotypes, diversity units incorporating ecological and evolutionary theory, are superior to OTUs for the purpose of studying community assembly.