Demographic, socioeconomic, and housing characteristics influence variation in household energy consumption. By combining household-level utility, public, and proprietary data, we examine predictors of household energy consumption in a Texas urban area. Using quantile regression, this analysis assesses the relationship between energy consumption and predictors at the middle and both ends of the distribution (10th and 90th percentiles). Results indicate potential opportunities to lower consumption among the highest energy-consuming households including those with pools, with non-central cooling, with people working from home, those built on pier/post foundation, and those that are renter-occupied. These findings suggest significant opportunities to reduce consumption and demand as in the study area, almost 10% of housing units are renter-occupied, 18% percent are without central cooling, and 7% have pools. Capturing a significant portion of these homes for retrofit conservation efforts through marketing has potential to produce substantial results. Producing a better understanding of determinants of household energy consumption using the methods presented has potential to assist development and implementation of strategies to reduce consumption and increase efficiency. (C) 2014 Elsevier Ltd. All rights reserved.