Here, the authors explore the potential of the 16S gene for discriminating bacterial taxa and show that full-length sequencing combined with appropriate clustering of intragenomic sequence variation can provide accurate representation of bacterial species in microbiome datasets.