Studies linking the functional diversity of a biota to ecosystem functioning typically employ a priori classifications of species into hypothetically complementary groups. However, multiple alternate classifications exist in which the number of functional groups, the number of species per functional group, and the grouping of species differ from the a priori scheme. Without assessing the relative precision, or ability of an a priori scheme to accurately predict ecosystem functioning relative to its many alternatives, the validity and utility of analyses based on a single a priori classification scheme remains unclear. We examine the precision of a priori classifications used in 10 experimental grassland systems in Europe and the United States that have found evidence for a significant role of functional plant diversity in governing ecosystem function. The predictive precision of the a priori classifications employed in these studies was seldom significantly higher than the precision of random classifications. Post-hoc classification schemes that performed well in predicting ecosystem function resembled each other more with regard to species composition than average classifications, but there was still considerable variability in the manner in which these classification schemes grouped species. These results suggest that we need a more nuanced understanding of how the diversity of functional traits of species in an assemblage affects ecosystem functioning.