This article presents an original method for capturing household social position, a configurational approach likely to improve sociological description and explanation of a number of points in this research area. The method consecutively uses geometric data analysis and classification, making it possible to account at once for several household members and several social stratification dimensions. We applied it to two French nationwide representative surveys: Trajectories and Origins (Trajectoires et origines or TeO, Ined/Insee 2008) and a panel survey of students starting secondary school in 2007 (Panel d'élèves entrant dans le secondaire en 2007, DEPP-MEN). In addition to classic variables like occupation or educational attainment, the configurations thus constructed bring to light the impact of residential variables on social differentiation of households together with the role played by family situation and migration origin. The article concludes by showing that configurations have stronger explanatory power than classic approaches to social position.