How Ballarat's Public Image Archive Ended Up Full of Duplicates — and What It Cost to Fix

Ballarat City Council's digital asset library contains more than 40,000 image files. Somewhere between a third and half of those, according to a review flagged in a council information technology working group discussion paper circulated in March 2026, are duplicates — the same photograph filed under different names, in different folders, by different departments over roughly a decade of uncoordinated digital storage.

The problem did not emerge overnight. It is the accumulated result of decisions — and non-decisions — made from about 2014 onward, when individual teams across the council, Sovereign Hill, Ballarat Health Services, and regional arts bodies like the Art Gallery of Ballarat each began building their own digital filing systems with no shared taxonomy and no central oversight. Every time a communications officer left, a new contractor arrived, or a major event generated a burst of photography, more images entered the system with whatever filename seemed logical at the time.

How the Duplication Built Up

The mechanics are straightforward enough. A photographer shoots Sturt Street during the 2019 Ballarat Begonia Festival. The raw files go to the events team. A council communications staffer pulls several for a media release and saves them to a shared drive. A tourism officer downloads the same images from a Dropbox link and renames them for a grant acquittal to Regional Tourism Victoria. Six months later, a web developer finds the originals on a USB, uploads them to the content management system, and tags them differently again. Four copies of the same photograph now live in four locations, none of them linked.

Multiply that process across eight years and dozens of events — Ballarat's 170th anniversary celebrations in 2021, the Eureka Centre precinct redevelopment consultation, the opening of the Ballarat Base Hospital's new inpatient unit in late 2023 — and the scale becomes clearer. Storage itself has a cost. Council's IT working group paper put the annual expenditure on cloud storage for the broader corporate asset library at roughly $280,000 for the 2025–26 financial year, a figure the paper noted could be materially reduced through deduplication.

Sovereign Hill faced a parallel version of the same issue. The living museum on Bradshaw Street draws on a photographic archive stretching back to its opening in 1970. As the organisation digitised older print photographs through a Heritage Victoria-supported program running from 2018 to 2022, scanned files were saved in multiple resolutions and formats without consistent naming conventions. Staff would later find three or four versions of the same 1975 gold-pour demonstration photograph catalogued under entirely different accession numbers.

The Push for a Unified System

The impetus for fixing this came partly from external pressure. The Victorian Auditor-General's Office published guidance in 2024 on digital record management standards for local government, noting that duplicate and untagged assets created both efficiency losses and potential compliance risks around copyright and consent documentation. Several regional councils, Ballarat among them, took the guidance as a prompt to audit what they actually held.

The Art Gallery of Ballarat on Lydiard Street North had already been through a smaller version of this process. In 2023, the gallery completed a Collections Management System migration that required staff to manually reconcile approximately 6,200 digital image records, collapsing duplicates and standardising metadata. That project, funded partly through a Museums and Collections grant from Creative Victoria, took 14 months and cost the gallery around $47,000 in contractor and staff time.

That experience is now being cited inside council's own review as a practical model. The gallery's collections registrar worked with a Melbourne-based digital asset management firm to write deduplication rules that could be adapted for larger municipal libraries.

For residents and ratepayers, the practical upshot is this: the council's image deduplication project, expected to be tendered in the third quarter of 2026, should produce a searchable, rights-cleared archive that any department can draw on without recreating work already done. Whether that delivers the storage savings projected in the March discussion paper will depend heavily on which vendor is selected and how consistently staff are trained to use whatever system replaces the current patchwork. The tender process will tell that story.