Recent initiatives towards the reduction of domestic energy consumption largely depend upon the availability of thorough information to enable focalised interventions and a posteriori impact assessment. This is a significant issue in developing countries, as informality of fuel markets and lack of intelligent technologies can undermine the capacity of policy makers to effectively target consumer behaviour. This study explores an alternative approach to inform such measures using energy forecasting before occupancy. Accordingly, a secondary analysis of publicly available datasets was conducted to assess general patterns of operational energy consumption in Chilean social housing and to develop a set of forecasting models which accuracy was later evaluated with the results of an on-site survey. Although the forecasting capabilities of the proposed models is not yet conclusive, the results of this study suggest that discrete socio-demographic factors can predict general patterns of operational energy consumption and therefore increase the accuracy of future energy efficiency measures.