One of the basic research activities of genebanks is to partition stocks into groups that facilitate the efficient preservation and evaluation of the full range of useful phenotype diversity. We sought to test the usefulness of making intraspecific groups by replicated rapid visual intuitive impressions of coded plants by multiple uncoached observers. We invented the term “cog” (shorthand for cognate = “born together”) to indicate assumed genetic relatedness of cog members. All of the 16 populations of the wild potato species Solanum okadae in the genebank were thus examined in four separate grow-outs by up to seven genebank staff members, a total of 26 times. They were instructed to place them into two cogs defined only as big and not-big. Four populations were placed in the big cog for 70–90 % of observations, while all remaining populations but one were placed in the big cog for less than 5 % of observations. All populations were then assessed for DNA markers and various empirical traits. AFLP and SNP markers clearly distinguished the two cogs. The big cog populations were also distinguished from the others by virtue of having less foliar late blight resistance, more leaf hairiness, and lower tuber tomatine content. SNP similarity suggests one population of reputed Bolivian origin is really a mislabeled duplicate of another from Argentina. If so, the two cogs also perfectly align with country of natural origin, with big originating exclusively from Bolivia. Using S. okadae as a model, we demonstrated that rapid, simple, and inexpensive visual intuitive cogs reliably predict significant genetic and phenotypic differences. We propose testing the cog technique on other species and applying it as a new potato germplasm descriptor.