Index numbers have been used to decompose aggregate trends in energy intensity, i.e., the ratio of energy use to activity. By making a direct appeal to the theory underlying price index numbers used by the energy decomposition literature, this note proposes the chain weighted Fisher Ideal Index as a formula that solves the 'residual problem.' The connection to index number theory also allows us to illustrate that the measures of activity used to define energy intensity need not be additive across the sectors that are involved in the decomposition. We give an empirical example using recent U.S. manufacturing data of the Fisher Ideal Index, compared to the Tornqvist Divisia index, a popular index in the energy literature.