Kinship plays a fundamental role in the evolution of social systems and is considered a key driver of group living. To understand the role of kinship in the formation and maintenance of social bonds, accurate measures of genetic relatedness are critical. Genotype-by-sequencing technologies are rapidly advancing the accuracy and precision of genetic relatedness estimates for wild populations. The ability to assign kinship from genetic data varies depending on a species’ or population’s mating system and pattern of dispersal, and empirical data from longitudinal studies are crucial to validate these methods. We use data from a long-term behavioral study of a polygynandrous, bisexually philopatric marine mammal to measure accuracy and precision of parentage and genetic relatedness estimation against a known partial pedigree. We show that with moderate but obtainable sample sizes of approximately 4235 SNPs and 272 individuals, highly accurate parentage assignments and genetic relatedness coefficients can be obtained. Additionally, we subsample our data to quantify how data availability affects relatedness estimation and kinship assignment. Lastly, we conduct a social network analysis to investigate the extent to which accuracy and precision of relatedness estimation improve statistical power to detect an effect of relatedness on social structure. Our results provide practical guidance for minimum sample sizes and sequencing depth for future studies, as well as thresholds for post hoc interpretation of previous analyses.