Since the first stage models of e-government were proposed around 2000, there have been at least 15 variants published in the academic and professional literature. This paper is a critical examination of these models which places them in the wider historical context of information systems stage and maturity modeling. It is argued that, with a small number of exceptions, most e-government stage models are theoretically weak being descriptive, not well grounded in empirical evidence and/or normative. If such models are to be useful, a different approach is required. A number of ideas for doing this are proposed.