The most commonly used methods for detecting and characterising regional-scale changes in cliff morphology involve differencing high resolution Digital Elevation Models. An inherent assumption of this 2D method is that the cliff can be reduced to a planar surface, which becomes invalid where cliffs change aspect. In this paper we examine the relative benefits of 2D and 3D methods of change detection, the latter of which draw on raw point cloud data, for deriving inventories of change. In our analysis we test both methods of change detection on two high resolution point clouds derived from Terrestrial Laser Scanning of the coastal cliffs at Staithes, North Yorkshire (UK). The analysis highlights the importance of the chosen method for accurately constraining the size distributions of rockslope failures, as well as the geometry of the failures themselves. We conclude by considering the implications of 3D techniques for defining rockfall geometry and inferring different processes of change.