The EU-funded SHAMAN (Sustaining Heritage Access through Multivalent ArchiviNg, http://www.shaman-ip.eu/) project is investigating the long-term preservation of large volumes of digital data in a distributed environment, by developing a preservation framework that is verifiable, open and extensible. During the initial stages of the project, a detailed user requirements analysis led by HATII at the University of Glasgow was conducted across three domains: memory institutions, industrial design and engineering, and e-science. This research pinpointed the needs and expectations that end-users and service providers feel should be met by such a preservation framework. This paper gives an overview of the requirements that were gathered, formulated and adopted by this project. It then discusses the outcomes of this empirical research and indicates both how these outcomes are being implemented within SHAMAN and how external parties may also benefit from the findings.Approaches to digital preservation are often still ad hoc and based on a single institution focus. They frequently do not take into consideration the needs of the variety of actors who will come into contact with a system throughout the preservation lifecycle. This paper provides an insight into the preservation practices that a broad range of real-world organisations would like to follow and provides a discussion of how SHAMAN intends to meet the needs of the identified users.