Warnings issued by meteorological or oceanographic agencies are a common means of allowing people to prepare for likely impactful events. Quantifying the relationships between ocean conditions and coastal impacts, such as shoreline change or flooding of coastal assets (e.g. flooded access points, overtopping of sea walls) is crucial for developing operational coastal hazard warnings. Existing studies have largely omitted empirical data, relying on modelling to estimate total water levels and impact potentials. It is well documented that site-specific conditions influence local morphodynamics and as such, detailed data related to the physical environment is a necessary component of these existing approaches. The capacity to collect these data is not always available, however, and so an alternative approach that does no rely on detailed modelling may be necessary in some instances to identify the conditions that lead to coastal impacts. We propose an alternative empirically based approach for isolating oceanic conditions that are conducive to impact along open coasts, using two case studies from Victoria, southeast Australia: Port Fairy and Inverloch. Oceanic conditions were defined using data obtained from a WAVEWATCH III (WW3) model hindcast, assessed against newly installed wave buoys, which evidenced variation in mean conditions between the two sites. We coupled impact-based data derived from citizen-science and social media to modelled and observational data, to identify the oceanic conditions that led to impacts. We found heterogeneity in the response of the case study locations to deviations from the local mean wave characteristics and still water levels. This paper demonstrates an approach through which impact-based thresholds for erosion could be developed for management applications and early warning systems.