This paper critically reviews the range of energy efficiency indicators that can be used, particularly at the policy level. Traditional thermodynamic indicators of energy efficiency were found to be of limited use, as they give insufficient attention to required end use services. The specific limitations and appropriate uses of physical-thermodynamic, economic-thermodynamic and pure economic indicators of energy efficiency are also considered. The paper concludes with a discussion of the persistent methodological problems and issues which are encountered when attempting to operationalize all of the energy efficiency indicators. These include the role of value judgements in the construction of energy efficiency indicators, the energy quality problem, the boundary problem, the joint production problem and the question of isolating the underlying technical energy efficiency trend from the aggregate indicator.