Given high-level commitments to reducing the rate of biodiversity loss by 2010, there is a pressing need to develop simple and practical indicators to monitor progress. In this context, a biodiversity intactness index (BII) was recently proposed, which provides an overall indicator suitable for policy makers. The index links data on land use with expert assessments of how this impacts the population densities of well-understood taxonomic groups to estimate current population sizes relative to premodern times. However, when calculated for southern Africa, the resulting BII of 84% suggests a far more positive picture of the state of wild nature than do other large-scale estimates. Here, we argue that this discrepancy is in part an artefact of the coarseness of the land degradation data used to calculate the BII, and that the overall BII for southern Africa is probably much lower than 84%. In particular, based on two relatively inexpensive, ground-truthed studies of areas not generally regarded as exceptional in terms of their degradation status, we demonstrate that Scholes and Biggs might have seriously underestimated the extent of land degradation. These differences have substantial bearing on BII scores. Urgent attention should be given to the further development of cost-effective ground-truthing methods for quantifying the extent of land degradation in order to provide reliable estimates of biodiversity loss, both in southern Africa and more widely.