Concentrating on the case of energy-saving technology, the authors explain the high frequency of implementation failure by examining the reward structure of federal R&D agencies and the limits of current methods of program planning and technology assessment.