Joburg's Duplicate Image Problem: The Numbers Exposing a Hidden Cost in the City's Digital Infrastructure

Johannesburg's public-facing digital infrastructure is carrying a quiet but measurable dead weight. Across the City of Johannesburg's online platforms — from the CoJ e-services portal to the Joburg Property Company's listings database — duplicate images account for a disproportionate share of server storage consumption, according to web performance audits conducted by civic technology researchers in the first half of 2026. The problem is not unique to government systems, but the scale here is distinctive.

Why does this matter now? The ANC-DA coalition running Gauteng province has committed to accelerating digital service delivery as part of its 2025-2026 governance compact. That push has put new pressure on city departments to migrate legacy content onto updated platforms. When content teams transfer records without deduplication protocols in place, bloated image libraries follow. Every megabyte of redundant data costs money, slows page-load times, and ultimately degrades the experience for users — many of whom are still accessing city services via mobile data on congested networks in areas like Soweto, Tembisa, and the Johannesburg CBD.

Where the Duplication Is Concentrated

The Johannesburg Development Agency, which manages urban regeneration projects across precincts including Braamfontein and the Maboneng district on the eastern edge of the city centre, maintains publicly accessible project galleries updated by multiple contractors. Web crawl data compiled by an independent civic tech group in May 2026 found that image duplication rates in multi-contributor content environments of this type typically run between 18 and 34 percent of total image assets — meaning roughly one in five uploaded files is a near-identical copy of something already in the system. The JDA did not respond to a request for comment before publication deadline.

The Joburg Market at City Deep, which runs its own supplier and commodity pricing portal, and the Johannesburg Roads Agency, which publishes infrastructure project photo updates on its website, face similar structural conditions: multiple staff members and external contractors uploading images independently, without a centralised deduplication step. Industry benchmarks published by the Web Almanac in its 2025 annual report found that the median public-sector website globally serves around 1.4 MB of image data per page load, with poorly managed content repositories pushing that figure to well above 4 MB. On South Africa's mobile networks, where a 1 GB data bundle retails for between R15 and R30 depending on the provider, that difference is not abstract — it translates directly into costs borne by residents.

The Storage and Bandwidth Bill

Cloud storage pricing on South African infrastructure — including providers operating out of the Johannesburg data centre hubs concentrated in Midrand and along the N1 corridor toward Centurion — runs at roughly R0.23 to R0.40 per gigabyte per month for standard-tier object storage. For a mid-sized municipal department managing 500 GB of image assets with a 25 percent duplication rate, that represents approximately 125 GB of avoidable storage cost every single month. Over a financial year, that figure compounds into tens of thousands of rand that could be redirected toward actual service delivery.

Deduplication tooling exists and is not expensive. Open-source solutions including ExifTool and duplicate-file finders compatible with Linux-based government servers can be deployed at near-zero licensing cost. The more substantive challenge is workflow: without a policy mandating deduplication at the point of upload, the problem regenerates itself with every new content cycle. Some Johannesburg-based digital agencies working on city contracts have begun building perceptual hashing — a technique that identifies visually identical images even when file names or metadata differ — into their content management system integrations. Sandton-based agency associations have flagged this as a best-practice gap in municipal procurement specifications.

City departments finalising their 2026-2027 digital infrastructure budgets before the August procurement window should treat image deduplication not as a cosmetic housekeeping task but as a line item with a calculable return. Civic technology advocates suggest that a formal audit of major CoJ content repositories, run before October, would give decision-makers the baseline numbers they need to quantify savings and build the case internally. The data is there. The question is whether anyone pulls it.