Zurich's Digital Archive Crisis: The Key Decisions Ahead on Duplicate Image Replacement

Zurich's public cultural and administrative institutions are confronting an increasingly costly reckoning with their digital image archives. Thousands of duplicate photographs, scans and digital assets — accumulated across decades of digitisation drives — are clogging databases at the Zentralbibliothek Zürich and the Stadtarchiv on Alfred-Escher-Strasse, forcing administrators to decide by autumn 2026 how aggressively to purge, replace or consolidate those redundant files.

The problem crystallised this spring after ETH Zürich's Digital Humanities Lab published a technical assessment of interoperability failures between Zurich's major public-facing image repositories. The report identified systematic duplication rates that, in some collections, exceeded 30 percent of total stored assets — meaning roughly one in three image files was either a near-identical copy or a lower-resolution legacy version of a file stored elsewhere in the same system. Storage costs are not abstract: commercial-grade archival server capacity in Switzerland currently runs to roughly CHF 4,000 to CHF 7,000 per terabyte annually when factoring in redundancy, security and access infrastructure.

What the Duplication Problem Actually Costs

For institutions operating under tight cantonal budgets, that arithmetic matters. The Stadtarchiv holds records stretching back to the medieval period, and its digitisation programme — accelerated under a 2019 cantonal initiative — generated enormous file volumes quickly. Speed, at the time, was the priority. Deduplication was not. The Zentralbibliothek, which houses the Graphische Sammlung and maintains a separate image portal for public access, faces a parallel challenge: its online catalogue links to image files stored in at least three separate backend systems, some of which date to pre-merger database environments from the early 2000s.

The decisions ahead are neither simple nor cheap. Automated deduplication tools — the kind now standard in commercial media asset management — can process large volumes quickly, but cultural heritage institutions require human review before any file is permanently deleted. A photograph flagged as a duplicate by an algorithm may carry distinct metadata, a different provenance note, or a marginally better scan of a historically significant document. Getting that wrong is irreversible. The Stadtarchiv's technical team has flagged that a full manual review of flagged assets could require the equivalent of two full-time archivists working for 18 months, at a combined personnel cost the institution has not yet publicly budgeted for.

Across the Limmat, at the Museum Rietberg in Rieterpark, curators have already piloted a hybrid approach: algorithmic pre-screening followed by subject-specialist sign-off for any file older than 1950. The pilot, which ran through the first quarter of 2026, cleared roughly 12,000 flagged files in eight weeks. That model is now being studied by the Stadtarchiv and is expected to feature prominently in a joint working-group report due before the Zurich city council in September.

The Fork in the Road This Autumn

Three options are on the table. The first is a centralised replacement programme, consolidating all public-institution image assets onto a single Zurich-administered platform — a solution that offers long-term savings but requires upfront capital investment and political consensus across institutions with fiercely independent governance structures. The second is a federated model, where each institution manages its own deduplication but adopts common metadata standards brokered through the Kanton Zürich's digital government programme, eZürich. The third is minimal intervention: patch the worst redundancy problems and defer structural reform until a future budget cycle.

Few observers of Zurich's cultural sector expect the third option to survive scrutiny. Storage costs compound annually, and the cantonal government has signalled that digital infrastructure reform is a prerequisite for any new digitisation funding after 2027. The September council session will likely set the terms. Institutions that present costed, interoperable plans stand to secure earmarked cantonal support; those that do not may find themselves frozen out of the next funding round entirely.

For anyone who uses Zurich's public digital archives — researchers at ETH, journalists, genealogists working through the Stadtarchiv's online portal on Alfred-Escher-Strasse — the practical message is straightforward: expect some catalogue disruption in late 2026 and early 2027 as files are assessed and links are restructured. The Zentralbibliothek has committed to publishing a migration notice at least 60 days before any major catalogue change goes live. Whether the other institutions follow suit will itself be a measure of how seriously Zurich takes the coordination challenge it has set for itself.