Technological innovations can be divided into two types. 'Incremental' innovation is the development of existing technologies, and dominates industrial R&D. 'Radical' innovation is more difficult to manage and evaluate because the uncertainties are much higher, preventing the quantification of likely impacts and success rates. However, radical innovations are at the core of many research programmes in the public sector, and their financial consequences can be enormous. Therefore, effective techniques are needed for their management, not based on conventional methods such as cost-benefit analysis.