We used a combination of logistic and multilinear regression models to analyze how the built environment and people's sociodemographic characteristics are related to energy consumption for personal mobility in the Netherlands. This energy consumption was accurately and comprehensively quantified based on distances travelled with a large number of modes, distances travelled with different types of cars, the occupancy of cars, and the probability of staying home. Local density was found to be the most important aspect of the built environment: it strongly reduced both the probability of using a passive mode of transportation and the associated energy consumption. Other relevant spatial variables included the landuse mix entropy, green space, and distances to city centers, supermarkets, and train stations. Sociodemographic characteristics were, however, more important overall. In particular, full-time employment was associated with higher energy consumption. Males, respondents from high-income households, and respondents with a higher education degree also consumed significantly more energy despite owning the most efficient cars. Our results thus highlight the importance of an energy policy mix that goes beyond the stimulation of technological progress.