How Zurich's Digital Archives Ended Up Full of the Same Image Twice — and What the City Is Doing About It

Zurich's public institutions are sitting on thousands of duplicate digital images — the same photograph stored two, five, sometimes a dozen times across disconnected servers — and the problem has grown expensive enough that the city's IT governance unit has made deduplication a formal budget line for 2026. The Stadtarchiv Zürich confirmed earlier this year that a preliminary audit of its digital holdings identified redundancy rates that were straining both storage infrastructure and staff time spent cataloguing assets that already existed elsewhere in the system.

The issue sounds mundane. It is not. Digital storage in public institutions is not free, and the cumulative cost of holding redundant image files across departments — each with its own procurement cycle, its own naming conventions, its own backup regime — adds up quickly. More importantly, duplicate images create version-control problems: an edited or watermarked copy sits alongside an unedited original, and archivists cannot always tell which is authoritative. For a city that prides itself on precision in public administration, that is an uncomfortable situation.

How the Duplication Happened

The roots of the problem go back to roughly 2008, when Zurich's departments began digitising physical archives in earnest. The push accelerated after 2015, when the city adopted a broader open-data strategy that encouraged departments to make holdings publicly accessible. Each department digitised independently. The Stadtarchiv on Neumarkt worked from one roadmap; the communications office at the Stadthaus on Stadthausquai worked from another; ETH Zürich's image library, though technically a federal institution, shared assets with city bodies through informal agreements that were never formally governed.

By the time anyone ran a systematic cross-departmental check, the same aerial photograph of the Limmatquai had been ingested seventeen times under different filenames. Multiply that by thousands of images and you have a structural problem, not an isolated clerical error. The issue was compounded by successive software migrations — from early content management systems to SharePoint environments to cloud-based digital asset management platforms — each of which imported whatever existed without first deduplicating.

The UBS-Credit Suisse merger fallout has an oblique relevance here. Several Zurich-based financial institutions, renegotiating their office footprints since 2023, handed large tranches of corporate photography archives to cantonal bodies for preservation. Those collections arrived with their own internal duplications already baked in, and integrating them cleanly has proven harder than anticipated.

What's Being Done Now

The Stadtarchiv launched a pilot deduplication project in the first quarter of 2026, using hash-based file comparison tools to flag identical binary files before moving to perceptual hashing, which catches near-identical images that differ only in compression or minor cropping. The pilot ran against approximately 400,000 image files held across three departments and found a redundancy rate that, according to the project documentation circulated to the city council's committee on digital governance, exceeded 22 percent.

That figure — one in five images effectively stored more than once — translates into real money. Enterprise cold-storage pricing in Swiss data centres currently runs at roughly CHF 0.018 per gigabyte per month for archival tiers. Across the full municipal image estate, which Zurich's IT planners estimate at several hundred terabytes, the annual overhead attributable to duplication is not trivial. The city has not published a final figure, pending completion of the broader audit expected by September 2026.

ETH Zürich's library science group, based on the Hönggerberg campus, has been advising the Stadtarchiv on metadata standards that would prevent future duplication at ingestion — essentially, a check-before-you-store protocol that flags potential matches before a file is written to the archive. Similar approaches have been adopted by major European municipal archives in Amsterdam and Vienna.

For Zurich residents and researchers who rely on the city's digital holdings — whether through the online portal at stadtarchiv.ch or through direct institutional access — the practical upshot is a cleaner, faster, more reliable catalogue. The deduplication project is scheduled to complete its first full departmental pass by the end of 2026. After that, the harder work begins: agreeing on which copy is canonical when duplicates differ, and building the cross-departmental governance to make sure the problem does not simply accumulate again over the next decade.