Debates about ‘land-sparing’ and ‘land-sharing’ strategies for conserving biodiversity in cities provide an overly simplistic characterization of alternate planning options. Increased urbanization manifests in a number of ways and sophisticated analyses of how species respond to urban environments are required before generalizations about the relative merits of either planning strategy should be made. We investigated how insectivorous bats respond to housing density (a measure of urbanization intensity) and a range of habitat variables by modelling the occupancy and activity of 12 species in Melbourne, Australia. In addition to species-level analyses, species were grouped into guilds (‘matrix’, ‘patch’ or ‘edge’ species) depending on their ecomorphology and level of patch-dependency. Housing density (dwellings per hectare: dph) was the dominant explanatory variable of occupancy and activity of most species studied. Site occupancy probabilities of patch and edge species were reduced by half at housing densities of just 1·1–4·5 dph depending on species, while their activity was halved at housing densities of 2·6–10·3 dph. The activity of two matrix species also declined, but at higher housing densities (reducing by half at 4·5–7·9 dph). Patch and edge species responded nonlinearly to tree cover at low housing densities, with peaks in occupancy and activity at 20–60%. This response broke down at medium housing densities, however, as most species were no longer present. Synthesis and applications. Occupancy probability and activity of almost all bat species studied were substantially reduced even at very low housing densities. Increasing tree cover had no positive effect on the occupancy and activity of patch and edge species once medium housing densities were reached. In order for land-sharing strategies to provide good habitat for patch and edge bat species housing densities would need to be so low as to be impractical in most urban planning contexts. By contrast, land-sparing with forested habitat remaining in 20% or more of the landscape could provide positive conservation outcomes for the species modelled here. We discuss the relevance of our findings in planning for conservation of bats and their habitats in other recently established and expanding cities.