Some cavity-dwelling vertebrates have adopted buildings in response to the loss of natural cavities in human-disturbed landscapes. Unfortunately, public health concerns surrounding human–bat interactions often lead to the exclusion of bats from structures. Provisioning evicted bats with roost boxes is a common management strategy; however, factors influencing bats' use of roost boxes are poorly understood. Using data from roost boxes monitored by conservation groups in New Jersey, we tested the importance of exclusion history in predicting use of roost boxes by bats relative to roost-box characteristics and installation practices thought to be important. We found that boxes installed after a bat exclusion were most likely to be used by bats (effect size = 2.07 [CI: 0.76, 3.39]), particularly if installed on a building rather than a non-building (effect size = 4.17 (CI: 2.53, 6.88]). Our results support the use of roost boxes to ameliorate potential negative impacts to bats that have been excluded from structures and to reduce human–bat conflicts.