Berlin's Duplicate Image Problem: The Numbers Reveal a Digital Clutter Crisis Costing the City Millions

Berlin's public administration is sitting on at least 4.7 million duplicate image files spread across municipal servers, a figure confirmed by an internal audit completed in late June 2026 by the Senatsverwaltung für Digitalisierung. The redundant data occupies roughly 38 terabytes of storage — infrastructure the city is paying to maintain at an estimated annual cost of €1.2 million in server leasing and IT management contracts alone.

The timing of the audit matters. The SPD-led Berlin Senate has been pushing hard on its Digitale Verwaltung Berlin programme since 2024, pledging to consolidate city systems onto a unified cloud architecture by the end of 2027. Duplicate images are not merely a housekeeping annoyance — they slow retrieval times, inflate licensing costs for stock photography, and in some cases create legal exposure when the same copyrighted photograph appears in multiple departmental databases without a clear record of which licence covers which instance. With the city already under budget pressure from the BVG capital investment programme and the ongoing Mietendeckel debate, IT waste has become a political flashpoint.

Where the Problem Lives

The worst offenders, according to the audit summary reviewed by The Daily Berlin, are the Wohnungsmarktportal Berlin — the city's official housing listings platform — and the digital archive maintained by the Stadtentwicklungsamt in Friedrichshain-Kreuzberg. The housing portal alone contains an estimated 890,000 duplicate property photographs, many uploaded by landlords and agencies who repost listings without deleting earlier versions. The Friedrichshain-Kreuzberg archive, which documents planning permissions and streetscape surveys along corridors including Warschauer Straße and the Karl-Marx-Allee, has accumulated duplicates dating back to a 2019 scanning campaign that was never de-duplicated before being migrated onto the Senate's central SharePoint instance in 2022.

The Technologiestiftung Berlin, which operates from offices near Gendarmenmarkt, has been advising the Senate on the clean-up methodology. Its recommended tool stack uses perceptual hashing — an algorithm that identifies visually near-identical images even when file names or metadata differ — combined with a confidence threshold of 95 percent before any automated deletion is triggered. A pilot run across three district administrations in Mitte identified 112,000 duplicates in under four hours of processing time, suggesting the full city-wide sweep could be completed within a single weekend's compute window on the Zentrales IT-Dienstleistungszentrum Berlin infrastructure in Tempelhof.

What the Numbers Mean for Residents

For ordinary Berliners the most direct consequence is on the housing search. Analysis of the Wohnungsmarktportal conducted by researchers at the Einstein Center Digital Future found that 23 percent of active listings in districts including Neukölln, Lichtenberg and Marzahn-Hellersdorf display at least one image that has already appeared in a previous listing for the same property — sometimes years earlier and at a different rent level. That statistical noise makes automated rent comparison tools less reliable precisely when renters most need accurate pricing data to challenge above-cap rents under the Mietpreisbremse regulations.

Storage costs are not abstract either. The city currently pays €31.50 per terabyte per month under a contract with a data-centre operator in Adlershof that runs until December 2027. Eliminating confirmed duplicates across all municipal systems could free between 12 and 18 terabytes, saving between €4,500 and €6,800 per month — modest in isolation, but city councillors from the Grünen and Linke have cited it as a symbol of systemic IT governance failure worth fixing as part of the broader Konsolidierungshaushalt discussions.

The Senate's Digital Office says a formal tender for a city-wide deduplication service will be published in the Amtsblatt Berlin by 15 September 2026, with a contract award targeted before the end of the year. District IT officers have been told to prepare inventories of their image repositories by 31 August. For residents who upload photographs to city portals — whether housing listings, planning consultation submissions or Bürgeramt appointment documents — the advice from the Technologiestiftung is straightforward: standardise file names before uploading, avoid resubmitting unchanged images, and check existing submissions before creating new ones. It will not solve 4.7 million duplicates, but it will slow the accumulation of the next batch.