This guide is meant for people whose role involves developing services or tools to support research data management (RDM) and digital curation, whether in a Higher Education Institution or a project working across institutions. Your RDM development role might be embedded with the research groups concerned, or at a more centralised level, such as a library or computing service. You will need a methodical approach to plan, elicit, analyse, document and prioritise a range of users’ requirements. The term ‘requirements discovery’ covers these activities, and this guide relates them to the process of developing RDM services. A DCC Guide How to Develop Research Data Management Services describes these services in more detail. Further support is available from the DCC and DataONE websites, and from the resources listed at the end of this guide. In section 2 of this guide we summarise data management roles and responsibilities, and why these make a methodological approach to requirements worthwhile. Effective research data management involves many actors, supporting technologies and organisation, including coordination of human and financial resources. Collectively this research data management infrastructure will call on different support services at different stages of the research data lifecycle. Support services will need to integrate their information systems to some degree. Projects that create research data will also need systems to help manage that data within the project lifetime. And once the data creators’ needs have been fulfilled, other actors and systems may be involved in data archiving and preservation to fulfil the needs of subsequent users and re-users. Non-academic stakeholders will have requirements, for example any external partners, or research participants who use or help to produce the data. Service offerings from other institutions and commercial providers will also have an important role. For example there will be a need to match individual researchers’ needs for long-term storage with the repositories or archives that best serve their discipline, and ensure these also meet the expectations of the funding body or institution. In section 3 we look at the contexts for RDM requirements. We take as the starting point a high-level model of the services that are typically needed. Then we consider why the research context is different from other areas of business process change, and what challenges that presents for identifying requirements.