Zurich Archivists and Planners Tackle a Growing Duplicate-Image Crisis This Week

Zurich's major public institutions are dealing with a surprisingly stubborn administrative headache this week: tens of thousands of duplicate digital images clogging archives, slowing workflows, and inflating storage costs at a moment when budgets are already under pressure. The problem is not new, but a combination of converging deadlines has pushed it to the top of the agenda for several organisations simultaneously.

The trigger is partly procedural. ETH Zurich, which consistently ranks among the world's top ten technical universities, is mid-cycle on a rolling data-governance review that runs through summer 2026. As part of that process, its library and research-data teams have been auditing shared image repositories used across departments — from architecture to earth sciences — and finding significant duplication. Multiple scans of the same document, near-identical aerial photographs, and redundant export files generated by automated pipelines have accumulated over years of distributed research work.

Why This Week Matters

The timing is not accidental. July 1 marked the start of a new fiscal half-year for cantonal bodies, making it a natural moment for Zurich's city administration — headquartered at Stadthaus on Stadthausquai — to assess digital infrastructure costs. Cloud storage is not free, and the canton of Zurich has been under sustained pressure to demonstrate efficient use of public funds, particularly after the prolonged legislative debates over housing subsidies tied to the Wohnungsnot crisis.

At the Stadtarchiv Zürich on Neumarkt, staff have been working since late June to implement a new deduplication protocol across its photographic holdings. The archive holds more than 1.2 million digitised images covering the city's history from the mid-nineteenth century onward. Duplicates entered the system in waves: first during a bulk scanning drive in the early 2010s, then again during the Covid-era rush to digitise paper collections when reading rooms closed. Identifying and replacing low-quality duplicates with canonical versions is painstaking work, but the archive expects to complete a first pass on its post-1945 photographic series by the end of August 2026.

The pharmaceutical sector, which anchors much of the Swiss economy and has a significant administrative and research presence in and around Zurich's Technopark on Technoparkstrasse in Zürich-West, faces its own version of the challenge. Regulatory submissions to Swissmedic require precise image documentation, and duplicated or mislabelled files in submission packages have drawn scrutiny during audits. Industry bodies have been circulating updated technical guidance on image-file naming conventions since May 2026.

Tools, Costs, and What Comes Next

The practical dimension is significant. Enterprise deduplication software licences — tools capable of comparing image hashes and flagging near-duplicate variants using perceptual algorithms — typically run between CHF 8,000 and CHF 40,000 per year for mid-sized institutional deployments, depending on volume and vendor. Smaller offices within the city administration have sometimes leaned on open-source alternatives, but these require technical staff time that is itself a scarce resource.

ETH Zurich's IT services division has piloted one such workflow this spring, processing roughly 600,000 image files from three separate research groups over a six-week period. The pilot identified a duplication rate of around 18 percent across that sample — a figure consistent with benchmarks published by digital preservation bodies in Europe. The university has not yet announced a broader rollout, but the pilot results are expected to inform decisions in the autumn semester planning cycle starting September 2026.

For smaller organisations — neighbourhood Quartiervereine uploading photographs to shared civic platforms, or the planning teams at Amt für Städtebau who maintain visual records of building permits — the practical advice this week from cantonal IT advisers is straightforward: establish a single canonical folder structure before adding new images, use consistent file-naming tied to date and project code, and run a free hash-check tool at the point of upload rather than waiting for an archive backlog to develop.

The broader lesson from this week's convergence of audits and reviews is structural. Digital collections grow faster than the governance frameworks designed to manage them, and Zurich's institutions — however well-resourced by European standards — are not immune to that pattern. The work of replacing and rationalising duplicate images is unglamorous, but archivists and IT managers here will argue, with some justification, that it is the kind of maintenance that determines whether a city's institutional memory remains genuinely accessible a generation from now.