Conserving migratory species is difficult because wide-ranging animals are challenging to study and aspects of their annual cycles occur in geographically distant areas. This challenge is illustrated by the hundreds of thousands of migratory bats killed annually during autumn migration by industrial wind turbines. It is unknown why bats are killed at turbines because they are so difficult to observe during migration. A conservation behaviour approach has potential to explain what is, arguably, the most significant environmental impact of wind energy. We tested predictions of two hypotheses to explain the presence of bats at turbines: (1) bats are attracted to visually conspicuous tall structures during autumn migration; (2) attraction is linked to social behaviour rather than foraging. We compared acoustic activity of migratory tree bats at conspicuous tall structures (telecommunication towers) to activity at two types of control sites (woodlot edges: consistently attractive to bats; open fields: consistently unattractive) before and during migration. Activity of migratory bats increased dramatically at towers during migration, from low premigration levels, to exceed that at open fields and match that at woodlots. Moreover, the proportion of feeding calls at towers remained low while the proportion of events during which multiple bats visited the towers simultaneously increased during migration to surpass that at woodlots and open fields. This suggests that migratory bats actively visit tall structures during migration for reasons other than foraging. Exploiting this behaviour, and the abundance of communication towers on the landscape, could reveal critical information about migratory behaviour of bats and help guide site selection for wind energy projects. Our findings demonstrate the importance of understanding the behaviour of migratory species at all stages of their annual cycle for effective conservation.