Kenya's Digital Archives Are Drowning in Duplicate Images — and the Numbers Tell a Damaging Story

Kenya's public institutions collectively store an estimated three to five duplicate copies of every digital image file they hold, according to data governance assessments circulated among technology officers in Nairobi this year. That redundancy is not a minor housekeeping problem. It translates directly into ballooning storage costs, slower service delivery, and — in the case of national identity systems — genuine security risks when outdated photographs persist alongside current biometric records.

The timing matters. The Ruto administration is operating under a tight IMF austerity framework, with every discretionary shilling under scrutiny. Government IT departments that cannot demonstrate lean, auditable data infrastructure are increasingly vulnerable to budget cuts. The pressure to clean up digital records is no longer a back-office concern — it sits in the same fiscal conversation as payroll reform and subsidy rationalisation.

The Scale of the Problem in Nairobi's Key Systems

Two institutions illustrate the challenge most sharply. The National Registration Bureau, which operates enrolment centres across Nairobi including offices along Harambee Avenue and in Kibera's Kambi Muru ward, has processed more than 38 million national identity cards since digitisation began in the early 2000s. Each enrolment generates a frontal photograph, sometimes a secondary scan, and in recent upgrade cycles a biometric image — yet legacy system architecture at many centres never automatically flagged or retired earlier versions of the same person's record. The result is layered image stacks that inflate database sizes and, periodically, cause verification systems to return conflicting identity matches.

The Nairobi Metropolitan Services data hub on Haile Selassie Avenue, which coordinates city permit and land registry records, faces a parallel issue on the property documentation side. Scanned title deed photographs, surveyor site images, and planning approval attachments routinely exist in three or four versions across different departmental folders — original scan, emailed copy, cloud backup, and a printed-then-rescanned version — with no automated deduplication layer in place. Storage administrators who spoke generally about the problem at a Kenya ICT Authority workshop in March 2026 described the situation as structurally inefficient rather than the result of deliberate mismanagement.

What Deduplication Actually Costs — and Saves

The economics of fixing this are more compelling than most budget holders realise. Cloud object storage in the AWS Africa (Cape Town) region — the closest enterprise-grade facility used by several Nairobi-based fintechs and government contractors — costs approximately $0.023 per gigabyte per month. A mid-sized government ministry storing 200 terabytes of image data, of which conservative audits suggest 40 percent is duplicate content, is effectively paying for 80 terabytes of redundant files every single month. At current exchange rates hovering around Ksh 129 to the dollar, that redundancy alone can represent annual costs exceeding Ksh 28 million for a single institution.

Safaricom's M-Pesa platform, one of the Silicon Savannah's most data-intensive local operations, has publicly discussed hash-based deduplication as part of its Know Your Customer image management since at least 2023. The principle is straightforward: every image file is assigned a unique cryptographic fingerprint; if two files share the same fingerprint, only one is retained. Applied across a government database, pilot exercises in comparable East African administrations — including Tanzania's National Identification Authority, which completed a deduplication exercise in late 2024 — have recovered between 30 and 45 percent of consumed storage within six months of deployment.

For Kenya's Gen Z-era accountability movement, the data resonates differently. The 2024 tax protest legacy left a public deeply sceptical about government waste. Duplicate image bloat is precisely the kind of quantifiable, preventable inefficiency that erodes trust — especially when citizens queue at Huduma Centre branches on Mama Ngina Street or in Westlands only to be told the system cannot resolve their biometric record.

The Kenya ICT Authority has flagged data quality standardisation as a priority under its 2023–2028 digital economy blueprint. The practical next step for ministries is a time-bound image audit — cataloguing file counts, storage volumes, and duplication rates by department — before any procurement of deduplication software begins. Without that baseline count, no ministry can honestly report to Treasury how much it is wasting, or how much it could save.