Seven earthworm ecological categories, including anecics, endogeics and epigeics, were defined by Marcel Bouche in 1972 based only on morpho-anatomical characteristics measured on European lumbricids. These categories had an outstanding success, and their use was generalized even outside of Europe. However, most of the time, only the three main categories are used, and over two decades, these categories have been considered functional groups, i.e. to presume how earthworms influence the soil functioning. Moreover, this relationship between ecological categories and functional groups is seldom tested and often uncritically accepted by soil biology researchers. It is then not surprising to observe unexpected trends when earthworm species are gathered in ecological categories to, for instance, analyze burrow systems or cast properties. We believe it is time to acknowledge that ecological categories are not functional, because they were not made for this purpose. We also propose three future directions in order to either improve our knowledge on the functional effects of earthworms or to build new authentically functional groups: (i) going back to the seven initial ecological categories to provide greater accuracy than the usage of the simplified three-level classification; (ii) testing, for a set of selected species, the functionality of the groups, being ecological categories or new tailored ones; and (iii) using trait-based approach in order to study the correlation between some earthworm traits and main soil functions.