A new method for the comparative analysis of productivity of tropical domestic flocks is presented here. This method used discrete-time demographic matrix models and the steady-state approach. Conventional steady-state demographic models use an annual step-time, which is poorly adapted to species with relatively short breeding cycles and with parturition occurring throughout the year. In the new model, we divided the year into two-week periods. The three major operational advantages were as follows: first, the short step-time that could decrease the bias generated by competing risks in the estimated demographic parameters (fecundity, mortality, animal outflow or inflow); second, the periodic model that could represent both the intraannual and interannual variations in demographic parameters and in other parameters like the live weight or the sale price of animals; finally, a proposed inference method, which used non parametric bootstrap, that helped calculate confidence intervals and build tests to compare the productivity of various flocks. The method was tested on field data obtained from sheep flocks in Senegal. It can also be applied to other domestic species and to wild species in various zootechnical or ecological environments