Batch fermentation performances are usually optimized on the basis of an overall criterion, the mean volumetric productivity. For lack of more suitable criteria, a great number of experiments have to be carried out under various conditions, in order to identify the factors acting on product formation rate. With the help of a mathematical model, every batch fermentation is quantitatively described by a set of parameters, so the reason of every improvement observed for the fermentation productivity is easy to recognize. Therefore, such a model appears to be an invaluable tool for finding quickly and at lower expense the optimal conditions. Nitrogen supplementation and inoculum preparation for lactobacilli growing on whey and whey permeate have been assessed with the help of a new mathematical model.