The cleanup focused on three primary data triggers. First, administrators ran a cross-check against the National Death Registry. This flagged entries where the date of death preceded the registration date or occurred years prior. These individuals were categorized as "ghosts" because they remained on the rolls despite being deceased.
Next, the team targeted duplicates using Unique Identification Numbers (UIN). Any entry sharing a UIN with another person was flagged immediately. If the names differed slightly—due to spelling errors or aliases—but the ID matched, the system marked the record for manual review by clerks.
Biometric discrepancies provided the hardest evidence. The software compared fingerprints and facial scans. If two separate profiles shared identical biometric markers, the system flagged them as clones. These cases often indicated fraudulent multiple registrations.
Finally, the commission audited physical addresses. Many entries listed nonexistent streets or single residences containing thousands of voters. Local officials verified these sites during field audits. Those without a verifiable physical residence were removed. This systematic approach cut 130 million entries from the rolls by relying on raw data mismatches and ground-level verification.
