Genetic markers are widely used to define and manage populations of threatened species based on the notion that populations with unique lineages of mtDNA and well-differentiated nuclear marker frequencies should be treated separately. However a danger of this approach is that genetic uniqueness might be emphasized at the cost of genetic diversity, which is essential for adaptation and is potentially boosted by mixing geographically separate populations. Here we re-explore the issue of defining management units, focussing on a detailed study of Galaxiella pusilla, a small freshwater fish of national conservation significance in Australia. Using a combination of microsatellite and mitochondrial markers, 51 populations across the species range were surveyed for genetic structure and diversity. We found an inverse relationship between genetic differentiation and genetic diversity, highlighting a long-term risk of deliberate isolation of G. pusilla populations based on protection of unique lineages. Instead we adopt a method for identifying genetic management units that takes into consideration both uniqueness and genetic variation. This produced a management framework to guide future translocation and re-introduction efforts for G. pusilla which contrasted to the framework based on a more traditional approach that may overlook important genetic variation in populations.