Birmingham's Digital Archive Problem: The Numbers Behind the City's Duplicate Image Crisis

Birmingham City Council's digital content estate contains an estimated 340,000 image files spread across at least seven internal content management systems, and a growing proportion of those files are duplicates, low-resolution placeholders, or images that have been replaced without the originals being removed. The result is a sprawling, expensive-to-maintain archive that is actively undermining the city's push to modernise its public-facing digital services.

The problem matters now because Birmingham is mid-way through a £14.2 million digital transformation programme, launched in January 2025, that is supposed to consolidate all resident-facing web platforms onto a single GOV.UK-style design framework by the end of 2026. Clean, properly tagged visual assets are a prerequisite for that migration. Every duplicate image that survives into the new system carries a metadata footprint that slows search indexing, inflates cloud storage costs, and — in the worst cases — means residents see outdated or factually wrong photographs on official pages.

Across Birmingham's Digbeth Creative Quarter, where the council co-funds the Digital Birmingham partnership alongside organisations including Brindleyplace management company Argent Related and the West Midlands Combined Authority, content audits carried out earlier this year flagged that roughly one in five publicly accessible council web pages contained at least one image marked internally as a duplicate or pending replacement. The audit covered more than 4,200 pages.

What the Numbers Actually Show

Cloud storage is not cheap at scale. The council's current hosting contract, held with a major public-sector framework provider under Crown Commercial Service arrangements, charges storage at a rate that makes retaining redundant files a measurable budget drain. Duplicate image sets — defined as files sharing more than 85 percent pixel similarity — were consuming an estimated 2.3 terabytes of billable storage as of the Q1 2026 internal review, according to documentation tabled at a Digital Services subcommittee meeting in March. At current contracted rates, that excess storage adds roughly £18,000 per year to the council's running costs for no functional benefit.

The scale of the duplication became harder to ignore after the council migrated its parks and leisure content last autumn. The Birmingham Leisure Trust, which manages sites including Cannon Hill Park in Moseley and the Midlands Arts Centre on Cannon Hill Road, contributed image libraries that had been built up independently over more than a decade. When those libraries were merged into the central CMS, automated scanning tools identified 6,700 duplicate or near-duplicate image pairs in a single batch. Staff had to manually review and clear approximately 1,400 of those pairs because the automated deduplication tool could not determine which version was the higher-quality or most current image.

The problem is not unique to leisure. Highways content relating to streets including Broad Street, the A38 Bristol Road corridor, and the inner-ring road around the Queensway tunnel showed similar patterns, with road condition photographs taken at different dates and resolutions sitting untagged alongside each other in the same asset folder.

The Path to Cleaner Data

Birmingham's digital team has piloted an AI-assisted deduplication tool across two departments since April 2026, covering roughly 40,000 files. The tool, procured through the council's existing agreement with a UK-based public sector software supplier, reduced reviewable duplicate pairs by 67 percent in the planning department's image library in its first eight weeks of operation. The libraries involved covered planning application photographs for areas including Jewellery Quarter conservation zone applications and Eastside development corridor submissions.

Spreading that approach across all seven CMS platforms is the stated goal for the second half of 2026. The Digital Birmingham steering group meets again in September, when it is expected to consider a paper on whether the current deduplication pilot should be expanded into a full programme before the end-of-year migration deadline.

For community groups and local organisations that submit images to council platforms — including neighbourhood forums in areas like Balsall Heath, Ladywood, and Kings Heath — the practical advice is straightforward: submit files at their original resolution, with consistent naming conventions, and keep a local record of what has been uploaded and when. That discipline at the point of submission is the cheapest fix available, and it costs nothing except time.