DNA barcoding, a method for species identification using short DNA sequences, has gained attention for its potential to aid in species discovery and identification. However, few well-sampled datasets are available to test its performance. This study examines the performance of DNA barcoding in a comprehensively sampled and diverse group, the cypraeid marine gastropods (cowries). The authors use a novel phylogenetic approach to calculate intraspecific variation and interspecific divergence and estimate error rates for identifying samples against a well-characterized phylogeny and assisting in species discovery for partially known groups. They find that the lowest overall error for species identification is 4%. In contrast, barcoding performs poorly in incompletely sampled groups, with species delineation relying on the use of thresholds to differentiate between intraspecific variation and interspecific divergence. The authors find substantial overlap between intraspecific and interspecific variation, leading to minimal error rates of ~17% in cowries. Error rates double if only traditionally recognized species are analyzed. Thus, DNA barcoding holds promise for identification in taxonomically well-understood and thoroughly sampled clades but is less effective in understudied groups. The authors conclude that the promise of barcoding will be realized only if it is based on solid taxonomic foundations.DNA barcoding, a method for species identification using short DNA sequences, has gained attention for its potential to aid in species discovery and identification. However, few well-sampled datasets are available to test its performance. This study examines the performance of DNA barcoding in a comprehensively sampled and diverse group, the cypraeid marine gastropods (cowries). The authors use a novel phylogenetic approach to calculate intraspecific variation and interspecific divergence and estimate error rates for identifying samples against a well-characterized phylogeny and assisting in species discovery for partially known groups. They find that the lowest overall error for species identification is 4%. In contrast, barcoding performs poorly in incompletely sampled groups, with species delineation relying on the use of thresholds to differentiate between intraspecific variation and interspecific divergence. The authors find substantial overlap between intraspecific and interspecific variation, leading to minimal error rates of ~17% in cowries. Error rates double if only traditionally recognized species are analyzed. Thus, DNA barcoding holds promise for identification in taxonomically well-understood and thoroughly sampled clades but is less effective in understudied groups. The authors conclude that the promise of barcoding will be realized only if it is based on solid taxonomic foundations.