Policymakers and firms use behavioral interventions to promote sustainable development in various domains. A correct impact evaluation requires looking beyond the targeted domain and assessing its interactions with similar interventions. Existing evidence in this area is limited, leading to potential misestimation of behavioural interventions and poor guidance on their design. Here, we test the impact of a two-year social information campaign to nudge water conservation through a large-scale randomized controlled trial implemented with a multi-resource company,. We find that the water nudge significantly decreases water and electricity usage, but not that of gas. Spillovers arise for customers who do not receive nudges targeting the other resources. Customers receiving the water report are also significantly less likely to deactivate their gas and electricity contracts, regardless of whether they receive other reports. Our results suggest that multiple nudges strain users’ limited attention and ability to enact conservation efforts. Users’ constraints in attending to multiple stimuli need to be accounted in designing policy interventions to foster sustainable practices.