Creators of resource discovery systems are at a turning point in development: their systems must satisfy the users’ needs for the delivery of high-quality, academically-rigorous content, while also providing the most straightforward, Google-like interfaces possible. The role of metadata in this scenario becomes paramount as users enter a world which seems serendipitous, but which is still subject to academic rigour, relevance and comprehensiveness. Ironically, such a system, necessarily supported by high-quality metadata, increasingly conceals them, thus lowering their profile in the user’s mind, and their importance becomes even more camouflaged to the user community. Without them, however, there is no system. This paper explores this dichotomy, describing a vision for resource discovery which creates a seamless journey between related, yet disparate, resources. It moves on to look at a case study of a single tool: the HASSET thesaurus, developed at the UK Data Archive. This tool’s journey from an inhouse-developed thesaurus to the lynch-pin of many online services is described in detail. Lastly, the paper takes the theory of achieving focused results from a simple interface and, using HASSET, applies it to an existing service: the Economic and Social Data Service (ESDS). Metadata is shown to be the essential – and yet increasingly invisible – force at work in effective resource discovery systems.