Institutional repositories (IRs) are said to be 'essential infrastructure for scholarship in the digital age', lessening the access barriers to scholarly communication, providing a measure of institutional research prestige and visibility, and facilitating formal research assessment. Why, then, with such large potential benefits, have IRs been less frequently implemented, harder to fill, and less visible than their proponents would have hoped? While there has been a lot of research time poured into technical platforms such as DSpace, ePrints, and the VITAL interface to Fedora used by the ARROW Project, very little is known about the users of IRs, nor how benefits of IRs to users can be maximised. An examination of the usability aspects of IRs from the perspectives of three msjor user groups (authors, information seekers, and data creators and maintainers) shows that IRs are not well investigated from a usability standpoint, and that this is likely to be impacting on uptake.