Renewal of coral populations through natural recruitment is declining as oceans warm. One aim of reef restoration is to augment natural recruitment by seeding sexually produced coral larvae and juveniles onto degraded reefs. Yet upscaling restoration with coral propagules will require overcoming inherently low rates of post-settlement survival, particularly on degraded reefs, and improving our understanding of the context-specific drivers of this mortality. To inform coral seeding techniques, we undertook a series of trials on the Great Barrier Reef (GBR) to investigate ecological and methodological questions, including: (i) the effect of conditioning duration, (ii) the role of herbivory, (iii) the effect of microcrevice structure and seeding-device design, (iv) the role of receiving community composition, and (v) the effect of species and genotype on coral survival. The key findings from these trials suggest that protection from herbivory is paramount, that survival is highest in microrefugia within the devices but that many microrefugia designs work, and that receiving community and genotype both significantly influence survival success. These results improve our understanding of context-specific drivers of success or failure in coral seeding, and can be fed into decision support tools to guide future restoration.