How Zurich's Digital Archives Ended Up Full of the Same Image Twice — and What Comes Next

Zurich's municipal digital infrastructure is carrying a problem that built up quietly for years: thousands of duplicate images spread across government databases, cultural archives and public-facing platforms, consuming storage, slowing retrieval systems and complicating the city's push toward unified e-government services. The reckoning is arriving now because Zurich's IT directorate — Stadtinformatik — is midway through a consolidation programme that is forcing different departments to finally map what they actually hold.

The issue matters because it is expensive. Cloud and on-premise storage costs for Swiss public institutions are not trivial. The Swiss Federal Audit Office has previously flagged inefficient IT spending across cantonal administrations as a systemic concern, and Zurich, as the country's largest city, runs some of the most complex digital infrastructure outside Bern. Every redundant file is a small charge that compounds across hundreds of thousands of assets.

A Decade of Competing Platforms

The roots of the duplication problem stretch back to the early 2010s, when individual Zurich city departments digitised their own collections independently and on separate timelines. The Stadtarchiv Zürich on Alfred-Escher-Strasse and the Zentralbibliothek Zürich near the Heimplatz both ran parallel digitisation drives, often capturing the same historical photographs of the Limmatquai, the Grossmünster and wartime Zurich from different donor collections. Neither institution had a shared metadata standard that would flag when two scans were functionally identical.

The tourism and communications arm of Stadt Zürich meanwhile built its own image library for press and marketing use, drawing on the same underlying photographic heritage. By the time cloud migration began in earnest after 2018, three or four copies of certain iconic images — the view from the Üetliberg, the Lindenhügel in winter — existed in separate siloed systems with different filenames, different colour profiles and different rights metadata attached to each.

ETH Zürich's library and archive operations encountered a parallel version of the same dysfunction. The university, consistently ranked among the world's top ten research institutions, digitised laboratory photography, architectural drawings and scientific illustration collections that overlapped substantially with holdings at the Museum für Gestaltung on Ausstellungsstrasse. Cross-institutional deduplication was discussed but never resourced until external pressure made it unavoidable.

The Mechanism That Let It Get This Far

Three structural factors enabled the problem to grow unchecked. First, Swiss data protection conventions made institutions reluctant to share full file inventories with one another, even within the same city administration. Second, procurement rules meant each department bought storage capacity independently rather than through a shared tender, removing any financial incentive to flag duplication early. Third, image deduplication tools that rely on perceptual hashing — software that recognises visually identical images even when file formats differ — were not adopted as a procurement requirement until the cantonal IT guidelines were updated in 2023.

The 2023 guidelines, issued by the Kanton Zürich's Amt für Informatik, set a formal requirement that new public digital asset management systems must include automated duplicate detection at point of ingest. That standard did not apply retroactively, which is why the backlog still exists today.

The scale is not trivial. Stadtinformatik has not published a final audit figure publicly, but the consolidation programme covering the first tranche of city departments — running from January 2025 through December 2026 — covers an estimated 1.4 million digital image assets across six administrative units, according to the programme's published scope documentation.

For residents and journalists who rely on the city's open-data portal at data.stadt-zuerich.ch, the practical consequence has been slower search results and inconsistent licensing information attached to ostensibly the same image. A photograph of the Rathausbrücke might appear three times in search results with three different copyright attributions.

The remediation path forward runs through two parallel tracks. Stadtinformatik plans to complete retrospective deduplication using perceptual hash comparison across the consolidated archive by the end of the first quarter of 2027. Separately, the Zentralbibliothek and Stadtarchiv are negotiating a joint metadata protocol that would let both institutions flag overlapping holdings in real time. Neither track is fully funded yet — budget approval is expected at the Stadtrat level in autumn 2026 — meaning the timeline depends on a political process that has its own pace.