This paper evaluates the relative efficiencies of a set of economic activities in Mexico through time, with the aim of setting targets for their future performance for the year 2018. These evaluations are based upon parameters obtained by the economic census recorded by the INEGI ('Instituto Nacional de Estadistica y Geografia INEGI' for Mexico); the data have been collected every 5 years, specifically in 1998, 2003, 2008 and 2013. To establish future performance targets, we apply the benchmarking tool Data Envelopment Analysis (DEA) together with the Malmquist index methodology, to measure the relative efficiency of each economic activity. DEA has been widely used in settings where the decision-making units (DMUs) are directly comparable, however for this particular case, the DMUs under study produce widely varying outcomes, hence are non-homogeneous. The paper presents a methodology for measuring efficiency when the DMUs involved exhibit such non-homogeneity. This is accomplished by recognizing that the DMUs fall into different groups, and provide a way of transforming what can be regarded as disadvantaged DMUs, so that those disadvantaged units can be compared to DMUs that have an advantage. To set goals for the future performance of economic activities, we employ an exponential smoothing approach to forecast future Malmquist indexes, thus allowing one to predict future outputs for those activities.