The Ediacaran fossil record includes some of the earliest known animals (572-550 million years ago (Ma)). However, understanding Ediacaran evolution is hampered by the lack of clear morphological analogues between Ediacaran and extant organisms.  Fortunately, Ediacaran preservation is exceptional, with thousands of soft-bodied, sessile organisms preserved in their in-life positions.  This in-situ preservation means that the position of the specimens on their bedding-planes captures their entire life-history: how they reproduced and interacted with each other and their local environment. As such, spatial analyses can be used to reconstruct Ediacaran eco-evolutionary dynamics. We have systematically mapped Ediacaran paleocommunities using LiDAR, photogrammetry and a laser line probe, and used spatial analyses to infer the underlying processes governing community dynamics.  We found a lack of influence of the local environment on the older deep-water organisms, which suggests that early metazoan diversification may not have been driven by systematic adaptations to the local environment, but instead may have resulted from stochastic demographic differences.  This pattern shifts for younger, shallow-water communities, which were heavily influenced by local environmental patchiness.  Our results provide quantitative support for the “Savannah” hypothesis for early animal diversification – whereby Ediacaran diversification was increasingly driven by patchiness in the local benthic environment.