In a new environment marked by the growing importance of Green House Gas emissions, fuel poverty, and energy efficiency in the different national agendas, the comprehension of energy demand factors appears to be crucial for the effectiveness of energy policies. We consider the latter could be improved by targeting specific household groups rather than looking to follow a single energy consumption level target. This article explores the scope of having a disaggregated energy consumption market to design policies aimed at curbing residential energy consumption or lowering its carbon intensity. Using a clustering method based on the CHAID (Chi Square Automatic Interaction Detection) methodology, we find that the different levels of energy consumption in the French residential sector are related to socio-economic, dwelling and regional characteristics. Then, we build a typology of energy-consuming households where targeted groups (fuel poor, high income and high consuming households) are clearly and separately identified through a simple and transparent set of characteristics. This classification represents an efficient tool for energy efficiency programs and energy poverty policies, but also for potential investors, which could provide specific and tailor made financial tools for the different consumer groups. Furthermore, our approach helps designing some energy efficiency score that could reduce the rebound effect uncertainty for each identified household group.