This is a process for clustering shared matches into subgroups at AncestryDNA. These subgroups are often more helpful than large shared match groups.