The general principles of scale and coarse and fine filters have been widely accepted, but management agencies and industry are still grappling with the question of what to monitor to detect changes in forest biodiversity following forest management. Part of this problem can be attributed to the lack of focused questions for monitoring including absence of null models and predicted effects, a certain level of disconnect between research and management, and recognition that monitoring can be designed as a research question. Considerable research from the past decade has not been adequately synthesized to answer important questions, such as which species or forest attributes might be the best indicators of change. A disproportionate research emphasis has been placed on community ecology, and mostly on a few groups of organisms including arthropods, amphibians, migratory songbirds, and small mammals, while other species, including soil organisms, lichens, bats, raptors, some carnivores, and larger mammals remain less well-known. In most studies of community ecology, the question of what is the importance, if any, of the regularly observed subtle changes in community structures, and causes of observed changes is usually not answered. Hence, our ability to deal with questions of persistence is limited, and demographic research on regionally-defined key species (such as species linked to processes, species whose persistence may be affected, species with large home ranges, species already selected as indicators, and rare and threatened species) is urgently needed. Monitoring programs need to be designed to enable managers to respond to unexpected changes caused by forest management. To do this, management agencies need to articulate null models for monitoring that predict effects, focus fine-scale monitoring on key species (defined by local and regional research) in key habitats (rare, declining, important) across landscapes, and have a protocol in place to adapt management strategies to changes observed. Finally, agencies must have some way to determine and define when a significant change has occurred and to predict the persistence of species; this too should flow from a well-designed null model.