Accretionary orogens are the sites of long-lived convergent margin tectonics, both compressional and extensional. They are also the hosts to the majority of the world's important gold deposits. A very diverse range of deposit types occurs within accretionary orogens, commonly in close proximity in space and time to each other. These include porphyry and associated high-sulphidation Au-Cu-Ag deposits, classic low-sulphidation Au-Ag deposits, low-sulphidation Au deposits centred on alkalic intrusive complexes, Carlin-type Au deposits, Au-rich volcanic-hosted massive sulphide deposits, orogenic Au deposits, intrusion-related Au deposits and iron oxide Cu-Au deposits. Empirical patterns of spatial distribution of these deposits suggest there must be fundamental generic controls on gold metallogeny. Various lines of evidence lead to the proposal that the underlying key generic factor controlling accretionary orogen gold metallogeny is regional-scale, long-term, pre- and syn-subduction heterogeneous fertilisation of the lithospheric mantle that becomes a source of mineralisation-associated arc magma or hydrothermal fluid components. This process provides a gold-enriched reservoir that can be accessed later in a diverse range of tectonomagmatic settings. Based on this concept, a unified model is proposed in which the formation of a major gold deposit of any type requires the conjunction in time and space of three essential factors: a fertile upper-mantle source region, a favourable transient remobilisation event, and favourable lithospheric-scale plumbing structure. This framework provides the basis for a practical regional-scale targeting methodology that is applicable to data-poor regions.