Many Group Decision Support Systems are designed to work within a complex organizational environment and to tackle complex strategic issues. Many have called for these systems to be evaluated in relation to clear success criteria. This paper addresses the issues that this request raises and seeks to encourage a debate about alternative research methods appropriate to such a task. Specifically the paper discusses: the purpose of the systems, the need to attend to the past and future of a group, GDSS to support negotiation towards decision-making, the significance of the consultant client relationship, the role of decision models, the duration of the intervention, and the role of a theoretical framework. A number of research methods are discussed and some of the problems in their use noted Ultimately the paper argues for more ‘action research’ rather than less of any other sort of research.