Because of limitations on the availability and quality of data, energy analysts and policy makers are forced to use indicators of industrial energy intensity (energy input/industrial output) that are less than ideal, In this paper, we identify and discuss several issues and problems that arise in the construction of commonly-used industrial energy intensity indicators; in particular, we focus on issues that arise due to the use of alternative measures of output in intensity indicators, Several commonly-used indicators of industry-level energy intensity were constructed for several US manufacturing industries in order to illustrate these issues and problems, We find that trends in energy intensity based on value of output (eg value of production) can diverge quite sharply from trends in energy intensity based on volume of output (eg tons of output), Discrepancies between value-and volume-based indicators of energy intensity in large part appear to be due to the way in which industrial output statistics are constructed under the US Standard Industrial Classification system, Discrepancies may also be the result of measurement errors in price deflators and simple definitional differences between various value-based measures of output, We conclude that policy makers should carefully examine and understand the limitations of commonly-used energy intensity indicators before such indicators serve as a basis for policy decisions.