Conservation planning is sensitive to a number of scale-related issues, such as the spatial extent of the planning area, or the size of units of planning. An extensive literature has eported a decline in efficiency of conservation outputs when planning at small spatial scales or when using large planning units. However, other key issues remain, such as the grain size used to represent the spatial distribution of conservation features. Here, we evaluate the effect of grain size of species distribution data versus size of planning units on a set of performance measures describing efficiency (ratio of area where species are represented/total area needed), rate of commission errors (species erroneously expected to occur), representativeness (proportion of species achieving the target) and a novel easure of overall conservation uncertainty (integrating commission errors and uncertainty in the actual locations where species occur). We compared priority areas for the conservation of freshwater fish in the Daly River basin (northern Australia). Our study demonstrates that the effect of grain size of species distribution data was more important than planning unit size on conservation planning performance, with an increase in ommission errors up to 80% and conservation uncertainty over 90% when coarse data were used. This was more pronounced for rare than common species, where the mismatch between coarse representations of biodiversity patterns and the smaller areas of actual occupancy of species was more evident. Special attention should be paid to the high risk of misallocation of limited budgets when planning in heterogeneous or disturbed nvironments, where biodiversity is patchily distributed, or when planning for conservation of rare species.