In recent years there has been an upsurge of studies on ecosystem multifunctionality (EMF), or the ability of ecosystems to simultaneously provide multiple functions and/or services. The concept of EMF itself, the analytical approaches used to calculate it, and its implications depending on the spatial scale and field of study have been discussed in detail. However, to date there has been little dialogue concerning the basis of EMF studies: what should or should not be considered appropriate measures for ecosystem functions. To begin this discussion, we performed an in-depth review of EMF studies across four major terrestrial ecosystems (agroecosystems, drylands, forests and grasslands) by analysing 82 studies, which together have assessed 775 ecosystem functions from a variety of field and greenhouse experiments across the globe. The number of ecosystem functions analysed varied from two to 82 per study and we found large differences in the distribution of functions across ecosystem types and ecosystem service categories. Furthermore, there was little explanation of why certain variables were included in the EMF calculation or how they relate to ecosystem functioning. Synthesis. Based on the literature analysis, it is clear that there is no general agreement regarding which measurements should or should not be considered functions in the field of ecology. To address this issue, we propose a general guideline for determining and measuring appropriate functions.