The American baseball play Yogi Berra once observed that “The future ain’t what it used to be”. Berra's colourful expression states a fundamental truth: the future as we perceive it today is, more often than not, not how it was perceived in the past. The proof is in the chequered history of forecasting in every field from politics to technology. e-Government and e-democracy are no exceptions. In this article we examine the problem of inaccurate forecasting of the impact of ICT in the public sector. To do this, we use a typology of forecast failure and then use this typology to examine three case of inaccurate forecasts of the impact of e-government and e-democracy over the past 25 years and ask why these predictions were wrong. Drawing on the lessons from these examples and others, we propose a series of steps/actions that can help to reduce the probability of such errors occurring in the future.