The inclusion of a conservation genomic approach promises to be indispensable for detecting genetic divergence, particularly for relatively poorly studied species with divergent phenotypes. We are using genotyping-by-sequencing (GBS) data to determine whether ‘summer’ and ‘winter’ breeding populations of New Zealand’s threatened Kermadec petrel (Pterodroma neglecta neglecta) represent genetically divergent lineages. The Kermadec Island group, ~1,000 km northeast of New Zealand, currently hosts >10,000 breeding pairs of Kermadec petrel in which two divergent breeding behaviours are represented. Most are ‘winter’ breeders (laying February-April), however a small number ‘summer’ breeders (~250 breeding pairs, laying October-November) have been identified on the Meyer Islands, 4 km from Raoul Island (the largest of the Kermadec Island group). These ‘summer’ breeders represent the last survivors of a great population once hosted by Raoul Island that was effectively extirpated during the mid-20th century through predation. Raoul Island became predator-free in 2004 and later expeditions have confirmed that ‘winter breeders’ have recolonised the island, while ‘summer’ breeders remain absent. Preliminary genetic evidence based on a small number of known ‘winter’ (n=22) and ‘summer’ (n=6) breeders indicates no shared mitochondrial cytochrome oxidase 1 haplotypes. Should genomic data, which will represent both putatively adaptive and non-adaptive variation, indicate that ‘winter’ and ‘summer’ breeders are genetically distinct, conservation action for ‘summer’ breeders will be warranted - particularly if this is further supported by non-genomic data. To this end, our findings will inform a larger interdisciplinary collaboration investigating the genetic, ecological and behavioural distinctiveness of Kermadec petrels across the Indo-Pacific.