Within the context of the increasing application of combinatorial methodology, the term 'diversity' has gained significant importance. The general understanding of this term is that diversity describes the degree of dissimilarity within a set of chemical structures. This Opinion article proposes that this understanding is superficial at best and irrelevant at worst. It is argued that relevant diversity can only be measured by the application of external criteria (such as a biological assay), which can discriminate the different structures by their different behaviour within this external context. According to this understanding, the diversity of a collection is highly dependent on the applied criteria. Therefore, a relevant diversity of chemical structures, per se, does not exist.