Considerable progress has been realized for the collection of consumed behavior data (store level scanner data). So, competition analysis models using this type of imput are becoming more numerous. This research comprises two parts. In the first part, we present the main concepts, directly or indirectly linked to the two brands relative measurement problem. We present a exhaustive view of competition models proposed in the last two decades, models using judgement or behavior data. The second part is focused on the measure of competition using panel data and is implemented in the SCAN 5000 panel of theNielsen company. Our idea of the "competition" is an unobservable and theoretical concept. So the proposed measures are only operationalizations of this concept. We use theory and technics developed in psychometry (measurement theory) for studying competition concept validity (reliability, traits validity, face validity). We propose a measure which accounts for the non symmetrical property of competition relationships and which has properties of distance. Finally, we study the influence of some important variables (Income, professional activity, consumed quantities, size of households, point of purchase, observation period) on the measures of competition of interest in this research
