Vers une mesure plus juste de la performance commerciale: une approche par la méthode DEA

Retailers need to determine the performance of their stores and beyond of the store managers in charge of running them. But this benchmarking is mostly based on variables which do not fill in these objectives. This paper proposes a fairer measure of the store performance in a French supermarket chain. Based on a DEA two-step model, this method enables the neutralization of environmental effects on sales. This modelling process leads to a ranking of store performances which is compared to this of the retailer. Comparing the DEA ranking and the retailer's ranking highlights no real relationship between the two rankings. Thus the retailer develop an evaluation process which favours stores which benefit from a larger surface with a higher number of employees located in more populated trade area, in other words variables dealing with size. Everyone knows that such a store is likely to obtain a higher level of sales. But is it optimal? The ranking stemming from a two-step DEA model links the store performance to the geo-demographic profile of the trade area and specifically the share of household with several vehicles or the proportion of individual houses. These variables explain also the performance and even makes it relative as well as the ability of the store and hence of its manager to adapt to the environment. We tend to favour a geomarketing approach which is both fairer to the manager and better adapted to the local markets.