postal companies around the world have been seeking methods to support their operational practices in order to keep or increase their market share. the brazilian mail and telegraph company (ect), established since 1969, offer a portfolio of products and services to its customers through several similar postal offices all over brazil. the units use several resources in different intensities, currently having their performance evaluated by a single economical factor: revenue over expense. we propose a new method of evaluation, using data envelopment analysis (dea), a linear programming technique that makes possible the inclusion of other factors beyond the simplistic and purely economical approach. we make our point using a set of units in the state of rio grande do sul. the objectives of this paper are: 1. to define which factors can be used to evaluate the units; 2. to define sets of similar units that develop the same functions, but differ only in the intensity of resource usage; 3. to generate, through the technique, the results of the evaluation process, obtained from the defined quantitative factors, providing a sensitivity analysis; and 4. to provide useful information to help managers in their decision making process. using the technique made possible: 1. to identify factors used in different intensities in different units; 2. to identify practices used by 100% efficient units, that can contribute to improve non efficient units; and 3. to allow and to encourage the participation of local managers in the modification and extension of the model observing particular features of their units.