One of the most important objectives in marketing field is to get a better understanding about consumer. In this paper, we have analysed the concept of heterogeneity and its treatment in consumer choice models. Following a review of the theoretical foundations and formulations about heterogeneity in consumer choice models, we proposed a classification of different kinds of heterogeneity. After this, and using scanner data we describe an empirical application in the context of consumer choice. Firstly, we developed a non heterogeneity MNL model. After that, we compare it with a random effects model and with a latent segments ones. Results show that models which considered heterogeneity explain consumer choice behavior more eficiently. Finally, main contributions of the study are discussed.