Bioequivalence of two drug formulations is currently defined by drug regulatory authorities in terms of the mean responses following administration of the test and reference formulations (average bioequivalence). However, the various potential shortcomings of average bioequivalence are now understood, and switchability, and thus individual bioequivalence, has become a reasonable expectation when changing from one pharmaceutically equivalent drug product to another. Progress has been made in developing criteria for individual bioequivalence, and an overview and classification of most of the different approaches to the assessment of individual bioequivalence have been achieved. As a consequence of this classification, the different character of scaled and unscaled bioequivalence measures has been recognized and, in turn, this leads to the proposal, made in this paper, of using both scaled and unscaled criteria for bioequivalence assessment of different classes of drugs, depending on their within-subject variability and therapeutic range. This strategy addresses the shortcomings of average bioequivalence, and, when applied to data sets from bioequivalence studies with four-period replicate crossover designs, turns out to have some satisfactory properties. Open questions and areas for further research are discussed.