Behavioral policy often aims at influencing behavior by mitigating biases due to, e.g., imperfect information or inattention. We study how this is affected by the simultaneous presence of multiple biases arising from different sources, through a field experiment on resource conservation in an energy- and water-intensive everyday activity (showering). One intervention, shower energy reports, primarily targeted knowledge about environmental impacts; another intervention, real-time feedback, primarily targeted salience of resource use. We find a striking complementarity. While only the latter induced significant conservation effects when implemented in isolation, each intervention became more effective when implemented jointly. This is consistent with predictions from a theoretical framework that highlights the importance of targeting all relevant sources of bias to achieve behavioral change.