Unlike many archival organisations, the University of Melbourne's eScholarship Research Centre (ESRC) is not a custodial repository or a teaching facility. This allows the centre to collaborate with a wide range of organisations and individuals, bringing archival thinking and practice to a variety of sectors, many of which are not traditionally associated with information professionals. Central to all the ESRC's work is the importance of effectively documenting things and their context. This paper draws on project examples, the author's PhD research and key concepts from archival and knowledge management theory to explore the idea that effective documentation requires more than a focus on items and collections. Instead, it requires working with individuals, organisations and documentary resources (published and unpublished) to reveal explicit connections and capture implicit knowledge in ways which more accurately reflect the complexity of collections and the entities needed to understand them. These ideas are introduced using two examples: a series of projects carried out over many years with the Victorian Government's Department of Primary Industries and its successors and The Australian Ballet. The paper then uses key concepts from this work to explore the nature of museum documentation and some of the limitations of current practice in museums, including the specific example of the Nordström mining models held by Museum Victoria. Thinking about these issues in the digital world, and applying archival thinking, the author argues for better connections between collection materials, not through convergence but by expanding our concept of collection documentation to include the relationships between things as things in their own right. Arguing for the practical benefits of such a change, the paper concludes by suggesting that testing these ideas in a museum context has the potential to further develop the ideas of the ESRC in ways which will benefit society.