The evolution and architecture of a set of retreating Lower Frasnian patch reef outcrops in the Canning Basin of Western Australia were evaluated, and their depositional and stratigraphic contacts spatially recorded using digital surveying tools. The geological data, together with high-resolution digital elevation models, were assembled in three-dimensional visualization and modelling software and subsequently used for building two-dimensional surface models and three-dimensional volumetric models. Numerical data on geometry and shape were extracted from these models and used to quantitatively assess the retrogradation motif of patch reef development. The development of the patch reefs comprises three stages. During stages 1 and 2, the patch reefs exhibited an overall retrogradational escarpment-type configuration displayed by, on average, 60° steep reef-margin walls that lacked the support of coeval slope deposits. The subdivision between stages 1 and 2 is based on minor backstepping reducing less than 10% of the platform-top area. The onset of stage 3 is recognized by stromatolite development fringing reef-margin walls. During stage 3 an aggrading accretionary reef-margin developed, comprising allochthonous and autochthonous slope deposits. Both types of slope deposit onlap the previous stages and are distributed unevenly with allochthonous slope deposits being noticeably absent around the smaller and more elongate patch reefs. The variation in distribution of slope sediment type can be explained by the amount, linked to platform size, of platform-top shedding. Small patch reefs were unable to fill the available accommodation adjacent to escarpments with allochthonous slope sediments and were thus encroached by autochthonous slope sediments. The variation, which cannot be explained by the size difference in the platform-top factory, has been related to the difference in perimeter length. For patch reefs with similar platform-top production areas, a more elongate patch reef inherits a longer perimeter and a proportionally smaller volume of allochthonous slope sediment per margin length will be transported to the flanks. Thus, the more elongate patch reef intrinsically contained more sites within which autochthonous slope sediments developed. Digital outcrop modelling and numerical evaluation of the evolution of the patch reefs revealed the major differences in retrogradation motif. The quantified variations in progressive decline of platform-top area with height were confirmed by hypothetical decline curves for ellipse-shaped carbonate systems for which aspect ratio (ratio between length and width) varied. This mathematical model demonstrates that the progressive decline of the production area is highly sensitive to shape and can be used to numerically assess and predict the relative timing of drowning, i.e. when the platform-top production area becomes nil, of retrogradational isolated carbonate platforms that are controlled by high accommodation. Wider implications can be surmised for highstand systems tracts and prograding carbonate systems. For example, for equally sized platforms with hypothetically similar carbonate factories and identical external forces, the potential to prograde by platform-top shedding is higher with a smaller aspect ratio because the shorter perimeter implies less accommodation space needing to be filled up to commence slope progradation. Clearly, there are intrinsic effects of shape on the development of carbonate platform systems.