The megadiverse subfamily Staphylininae traditionally belonged to the best-defined rove beetle taxa, but the advent of molecular phylogenetics in the last decade has brought turbulent changes to the group’s classification. Here, we reevaluate the internal relationships among the tribes of Staphylininae by implementing tree inference methods that suppress common sources of systematic error. In congruence with morphological data, and in contrast to some previous phylogenetic studies, we unambiguously recover Staphylininae and Paederinae as monophyletic in the traditional sense. We show that the recently proposed subfamily Platyprosopinae (Arrowinus and Platyprosopus) is a phylogenetic artefact and reinstate Arrowinus as a member of Arrowinini stat. res. and Platyprosopus as a member of Platyprosopini stat. res. We show that several recent changes to the internal classification of the subfamily are phylogenetically unjustified and systematically unnecessary. We, therefore, reestablish Platyprosopini, Staphylinini, and Xantholinini as tribes within Staphylininae (all stat. res.) and recognize Coomaniini as a tribe (stat. nov.) rather than subfamily. Consequently, the traditional ranks of the subtribes Acylophorina, Afroquediina, Amblyopinina, Antimerina, †Baltognathina, Cyrtoquediina, Erichsoniina, Hyptiomina, Indoquediina, Quediina, and Tanygnathinina are restored (all stat. res.). We review the current classification of Staphylininae and discuss sources of incongruence in multigene phylogenies.