Building energy efficiency has been a cornerstone of greenhouse gas mitigation strategies for decades. However, impact evaluations have revealed that energy savings typically fall short of engineering model forecasts that currently guide funding decisions. This creates a resource allocation problem that impedes progress on climate change. Using data from the Illinois implementation of the U.S.’s largest energy efficiency program, we demonstrate that a data-driven approach to predicting retrofit impacts based on previously realized outcomes is more accurate than the status quo engineering models. Targeting high-return interventions based on these predictions dramatically increases net social benefits, from $0.93 to $1.23 per dollar invested. •ML-based prediction with household data can increase the accuracy of engineering projections used in the nation’s largest energy efficiency program.•We develop an ex ante targeting function to predict the NPV of benefits from a range of retrofits.•Targeting funds to projects using ex ante projections increases benefits from $0.93 to $1.23.