The agglomeration of economic activity is undeniable (Krugman, 1991) and everyone can think of examples of specialised districts in cities, or business clusters, for example. The explanation for agglomeration phenomena currently appears to be a theoretical one (Fujita et al., 1999; Fujita and Thisse, 2002), while empirical studies do not seem to have reached the same level of maturity (Rosenthal and Strange, 2004; Ellison et al., 2010; Gibbons et al., 2014). Over the last decade, numerous spatial economics studies have measured geographical concentration. Traditionally, economists used measures based on zoning (such as the Gini index), but recent works have shown that discretising space may introduce biases (Briant et al., 2010). The use of distance-based measures (separating the entities analysed) rather than zoning is recommended nowadays (Combes et al., 2006). Our methodological contribution shows that special attention should still be paid to the definition of spatial concentration in any assessment of agglomeration of economic activity. Focusing on the location of retail outlets in the Lyon urban area, and using three measures of concentration recently introduced into spatial economics (Kd, D and M), we show that the findings do not systematically converge. This difference between estimates stems from the definition used for spatial concentration, as it may be absolute (strong presence of activity), topographical (high density of activity) or relative (overrepresentation of certain activities). We therefore recommend that the choice of concentration measure should be sufficiently informed from a theoretical viewpoint to make a correct appraisal of the phenomenon analysed and thus make a satisfactory evaluation of the distribution under study.