Biometric Voter Verification: How 88.6% of Voters Were Fingerprinted in 2017

On August 8, 2017, fingerprint scanners verified 6.76 million voters. But over a million had to use manual verification. That tension between technological ambition and on-the-ground reality defined the 2017 election's biometric voter verification process, and it would eventually play a role in the Supreme Court's historic decision to annul the presidential results.

The Verification Funnel: From 7.6 Million to 6.76 Million

The Independent Electoral and Boundaries Commission (IEBC) deployed 10,667 Kenya Integrated Election Management System (KIEMS) kits across the country's polling stations. These electronic devices were designed to verify each voter's identity through fingerprint scanning before they could be issued a ballot.

Here is how the numbers broke down in the verification summary:

• 7,631,705 voters presented themselves for verification
• 6,762,385 were accurately and biometrically verified (88.6%)
• 131,476 were verified through alphanumeric (text) search (1.7%)
• 737,844 remained unaccounted for in the biometric system (9.7%)

The 88.6% biometric verification rate sounds impressive, and it was a significant improvement over 2013. But the 9.7% gap, representing nearly 738,000 voters who could not be biometrically verified through either fingerprint or text search, raised serious questions about the integrity of the voter identification process.

August vs October: Two Very Different Elections

The Supreme Court's annulment of the August 8 presidential results forced a fresh election on October 26, 2017. This gave Kenya a rare natural experiment: the same biometric technology, deployed in the same polling stations, but under very different political conditions.

The comparison is striking:

Several key differences emerge from this data:

1. Turnout collapsed in October. Total electronic verifications dropped from 14.6 million in August to 7.6 million in October, a 48.3% decline. This was largely because Raila Odinga and the NASA coalition boycotted the fresh election, leading to near-zero turnout in opposition strongholds.

2. Presiding Officer overrides dropped dramatically. In August, 1,025,844 voters (7.0% of those verified) were identified through the Presiding Officer's account rather than biometric fingerprinting. In October, this figure fell to 211,446 (2.8%). This suggests that the IEBC tightened verification protocols for the repeat election, or that lower turnout simply reduced the strain on the biometric devices.

3. The biometric system worked better with fewer voters. With roughly half the turnout, the October election saw a much cleaner biometric verification process. This raises an important infrastructure question: were the KIEMS kits overwhelmed by the volume of voters in August?

Why Biometric Verification Matters

Biometric voter verification was introduced to combat two persistent problems in Kenyan elections: impersonation (where someone votes using another person's name) and multiple voting. By requiring a fingerprint match against the voter register, the system was designed to ensure that only registered voters could cast ballots and that each person voted only once.

The technology was central to the Supreme Court's scrutiny of the August election. In the landmark ruling that nullified Uhuru Kenyatta's victory, the court examined whether the KIEMS kits had been properly used and whether the results transmission system had been compromised.

The 7.0% Presiding Officer override rate in August was one data point that drew attention. While the IEBC maintained that overrides were a legitimate fallback for malfunctioning devices, critics argued that the manual verification pathway created an opportunity for voter fraud.

The 10,667 Kits: Scale and Logistics

Deploying 10,667 KIEMS kits across Kenya was a massive logistical undertaking. Each kit needed to be configured with the voter register for its specific polling station, charged, tested, and transported to locations ranging from urban schools in Nairobi to remote pastoralist communities in Turkana and Marsabit.

The kits served a dual purpose: voter identification through fingerprint scanning, and results transmission through an integrated 3G/4G modem. When the results transmission component came under scrutiny after the August election, the entire KIEMS system became a focal point of the political crisis.

Lessons for Future Elections

The 2017 biometric verification data offers several lessons for Kenya and for other countries adopting similar technology:

1. Plan for fallback scenarios. The 7.0% override rate in August shows that no biometric system works perfectly at scale. Electoral commissions need clear protocols for what happens when fingerprints do not match.
2. Lower turnout improves accuracy. The October data demonstrates that biometric systems perform better under less pressure. This argues for more polling stations or extended voting hours to reduce congestion.
3. Transparency builds trust. The public debate over KIEMS kits showed that deploying technology without public understanding of how it works can actually reduce confidence in the process.

For more on the 2017 election's technology infrastructure, see our analysis of Five Years On: Remembering August 8, 2017 by the Numbers and our breakdown of What Makes a Petition Succeed?, which examines how technology failures contributed to successful election challenges.