G protein-coupled receptors (GPCRs) account for approx. 2% of the human genome and represent the major targets for around 30% of all medicines on the market. Traditionally, optimising the interaction of lead molecules with the binding site for the endogenous agonist (“orthosteric” site) has been viewed as the best means for obtaining selectivity of receptor action at GPCRs. It is now recognised, however, that GPCRs possess allosteric binding sites, and that ligands can utilise these sites to modulate receptor activity through conformational changes transmitted from the allosteric to the orthosteric site and/or to effector coupling sites. Allosteric modulators may offer therapeutic advantages over orthosteric ligands for some receptors, including a greater potential for receptor selectivity and a higher safety in overdose due to a “ceiling level” to the allosteric effect. With the advent of high throughput screening, the study of GPCR allosterism represents an important new paradigm that has the potential to significantly open up the chemical space associated with novel GPCR ligands. In order to capitalise on the promise of GPCR allosteric modulators, however, a number of challenges need to be addressed. The first challenge is conceptual. The word “allosteric” has been used in a number of ways since it was first coined four decades ago. Much of the confusion arises from the fact that the binding of an allosteric modulator to a GPCR causes a conformational change in the receptor that can manifest a variety of behaviours. In order to describe allosteric drug actions in a quantitative manner that can provide useful information for drug discovery programmes, therefore, a basic (minimal) mechanistic framework is required. Analytically, this has been approached via the application of an allosteric ternary complex model (ATCM) of interaction, and more recent extensions, to quantify not only modulator binding affinity, but also the cooperativity manifested between orthosteric and allosteric sites. At the molecular level, recent studies are beginning to dissect the “topography” of allosteric sites. For example, we have found that the dynamics of the second extracellular loop of the M2 muscarinic acetylcholine are vital for the actions of prototypical muscarinic receptor allosteric modulators. A second challenge to the study of GPCR allosterism is practical. The perceived paucity of allosteric modulators in the known population of biologically active molecules is likely due to the fact that classic high-throughput screens have traditionally been biased towards the detection of orthosteric ligands. Because functional assays have now overtaken radioligand-binding assays as the high-throughput method of choice, allosteric ligands that have minimal effects on orthosteric binding have been discovered through their effects on receptor signalling. However, many allosteric ligands are still likely to appear quiescent in functional assays when tested in the absence of orthosteric probe. In addition, it remains more difficult to validate an allosteric mechanism/site of action in a functional assay than in a binding assay, and thus the optimal detection of novel allosteric ligands requires the combination of both standard functional and modulator-optimised binding assays. A third challenge for allosteric modulator-based drug discovery arises as a consequence of a combination of the first two challenges and, for want of a better term, can be described as cultural. From a “Chemistry” perspective, the structure-activity relationships (SARs) that govern orthosteric effects do not apply to allosteric binding sites. Although this can lead to a greater scope for ligand fishing in the chemical space encompassing biologically active molecules, it can also impose additional complexities/constraints in the application of traditional approaches to rational drug design. From a “Biology” perspective, the ability to not only detect, but to validate and quantify allosteric phenomena in terms of plausible models remains of paramount importance, as it is the parameters derived from such models that can be used to inform the design of refined SAR studies. In turn, the SAR studies can yield more efficient chemical tools with which to further probe the validity of the allosteric GPCR models. There is now at least one GPCR allosteric modulator on the market, and a number in clinical trials. Provided that drug discovery programmes recognise and accommodate the nuances involved in detecting allosteric effects, the search for allosteric GPCR modulators could yield a significant number of novel therapeutics in the new millennium.