Functional diversity is the component of biodiversity that concerns the diversity of roles that organisms play in ecosystems. It is increasingly used to investigate the processes that structure species assemblages, the functional consequences of extinctions, and as a measure of biodiversity that relates to ecosystem services and processes. However, measuring multivariate trait diversity is not a simple matter. A commonly used measure, functional group richness, has several limitations. Trait based multivariate measures escape these limitations and are being applied to an increasing range of ecosystems, taxa, and ecological questions. Recently, we have found that the choice of traits and null model used to answer a particular ecological question are more influential than details of the functional diversity measure. For example, whether local communities appear more or less diverse than expected is largely determined by the chosen null model. Such findings suggest that rhetorical arguments about the most appropriate functional diversity measure may have limited value. Focusing on trait data and analyses appropriate for answering real ecological questions may be a more profitable and interesting approach.