Ligand-selective receptor conformations introduce the concept of ‘texture' to drug effects, with respect to ligands possessing quality in addition to quantity of efficacy. This cell-dependent phenotypic efficacy extends to ligand properties beyond G-protein signaling and, in terms of drug development, presents a two-edged sword to pharmacologists. On the one hand, such efficacy promises more selective agonism but on the other hand it predicts problems associated with the use of recombinant or natural lead optimization assays as predictors of therapeutic value in humans. In this article, the evidence to suggest that not all agonists produce the same receptor active state is reviewed.