Minimising biodiversity loss in the face of large scale deforestation relies heavily on protecting selected areas of habitat within reserves. However, limited funding means it is necessary to prioritise areas for protection. Most current prioritisation schemes aim to protect the greatest amount of threatened biodiversity possible within a limited area. Here I describe a complementary approach, based on the systematic identification of areas in which species tend to be inherently extinction-prone. This is a more forward-looking approach to global conservation planning that should allow planners to anticipate and prevent future species declines in many parts of the world that still retain much of their original forest cover.