Assemblages are nested if species present at species-poor sites are subsets of those present at species-rich sites. In fragmented landscapes, nestedness analyses have been suggested as a means of assessing which patches are most important for biodiversity conservation. In the theoretical situation of perfect nestedness in relation to patch size, the single largest patch is disproportionally more important compared to smaller patches and will capture all species of conservation concern. However, real ecosystems are rarely perfectly nested. Here, we examined how different the implications for conservation management would be for an assemblage of birds that was highly significantly, but imperfectly nested in relation to patch size. The study focused on a fragmented landscape in southeastern Australia. Across 43 patches, 76 species of birds were recorded and classified as generalist, intermediate and sensitive species. The dataset was highly significantly nested by patch size (p=0.002). Under perfect nestedness by patch size, the single largest patch would have captured all species, and all sensitive species would have co-occurred in the largest patch. In our imperfectly nested dataset, co-occurrence patterns were substantially weaker. Usually, less than half of the sensitive species co-occurred in any given patch, and using the largest patches only, over a quarter of the study area would have been required to capture 80% of sensitive species at least once. These findings highlight there can be large qualitative differences between theoretical perfectly nested assemblages, and real imperfectly nested assemblages. Despite the outcomes of our study which showed highly significant nestedness by area, smaller patches in the system were important to complement large patches. We therefore argue that nestedness analyses need to be interpreted carefully, especially in an applied conservation context. Alternative conservation planning tools which consider the complementarity of various different patches are likely to be more informative for conservation management than nestedness analyses.