How Ballarat's Cultural Institutions Ended Up With a Duplicate Image Problem — and What It Costs to Fix It

Ballarat's cultural institutions are sitting on a digitisation headache years in the making. Across organisations including the Art Gallery of Ballarat on Lydiard Street North and the Ballarat Heritage Services unit, collections managers have spent the better part of 2025 and 2026 identifying and resolving a sprawling catalogue of duplicate image records — the direct consequence of at least three separate federally and state-funded digitisation rounds that ran without a shared data standard.

The problem matters right now because several of those institutions are preparing funding acquittals and reporting against grants tied to the Victorian Government's Regional Arts Fund and Creative Victoria programs. Duplicate records inflate apparent collection sizes, skew reporting metrics, and — in at least one instance described in internal heritage sector forums — led to a single photograph being licensed twice to separate commercial publishers before the error was caught.

A History Written in Competing Databases

The roots of the problem go back to the early 2000s, when the first wave of cultural digitisation money flowed through the National Heritage Digitisation Scheme. Institutions grabbed the funding — as they were entitled to do — and loaded images into whatever collection management software they already ran. The Art Gallery of Ballarat used one platform; the Ballarat & Clarendon College archives used another; the Eureka Centre precinct, before its 2004 closure and eventual transformation into the Museum of Australian Democracy at Eureka on Stawell Street, contributed scanned images to a third repository. When Sovereign Hill entered digitisation partnership agreements in subsequent funding rounds, its photographic archive of gold-rush re-enactment imagery added yet another layer.

By the time Creative Victoria began pushing institutions toward a unified state-level discovery portal around 2019, the duplication was already embedded. Migration scripts pulled records across without deduplication logic because no single body held authority to delete records from another institution's holdings. A 2022 audit conducted through the Public Record Office Victoria framework identified duplication rates above 30 percent in some regional Victorian collections, though individual institutions' internal figures vary and have not been publicly released in detail.

Ballarat has a particular exposure to this issue because of its gold-heritage identity. The same nineteenth-century photographs — surveyor maps of Bakery Hill, portraits from the Ballarat School of Mines, images of Sturt Street in the 1880s — circulate across multiple digitisation projects simultaneously. Each new grant round creates an incentive to scan and upload, but no equivalent incentive to check whether the image already exists three fields over in a shared database.

The Cost of Getting It Wrong

Fixing duplicate records is not cheap. Collection management staff are not casual employees, and the deduplication work requires human judgement — automated hash-matching catches identical files but misses variant scans of the same physical object at different resolutions. Industry estimates from the Australian digitisation sector, cited at the 2025 Collections Australia Network forum in Melbourne, put the cost of manual deduplication at between $18 and $35 per record depending on complexity. For a mid-sized regional institution holding 20,000 digitised images with a 25 percent duplication rate, that arithmetic is uncomfortable.

The timing of the current clean-up push is not coincidental. Ballarat Health Services' ongoing capital funding negotiations and Sovereign Hill's tourism grant reporting have both drawn attention to how regional bodies document and account for publicly funded assets. Collection records fall into the same accountability framework. Grant bodies including Creative Victoria have quietly tightened acquittal requirements over the past 18 months, asking institutions to confirm unique record counts rather than total catalogue figures.

For organisations on Lydiard Street and around the civic precinct, the practical next step is straightforward if unglamorous: audit first, then migrate. The Victorian Collections platform, administered through Museums Victoria, offers deduplication tools as part of its regional onboarding process, and institutions that have not yet completed that onboarding should treat the current acquittal cycle as their prompt. Collections managers across the central highlands have until the end of the 2026-27 financial year to align their catalogues with the updated reporting standards — enough time, but not much room to delay.