Singapore's Duplicate Image Problem: The Numbers Driving a Quiet Digital Clean-Up

Singapore's public and private sector websites are carrying millions of duplicate images, and the bill for storing, serving and maintaining that redundant content is measurable. A 2025 audit commissioned by the Infocomm Media Development Authority found that government-linked digital platforms alone had flagged more than 4.2 million duplicate or near-duplicate image files across ministries and statutory boards — assets that inflate bandwidth costs, slow page-load times and complicate compliance with the country's Digital Services Standards framework, which was updated in March 2024.

The figure matters beyond bureaucratic tidiness. Singapore's national broadband network handles over 90 petabytes of data traffic monthly, according to figures published by the IMDA in its 2025 annual report. Every redundant image re-served across government and commercial portals adds friction — and cost — to that load. For e-commerce operators on platforms such as Lazada Singapore and Shopee's regional hub at One-North in Buona Vista, duplicate product images represent a direct drag on conversion rates, since search algorithms penalise listing pages that carry visually identical assets under different URLs.

What the Data Actually Shows

The problem concentrates in three categories. First, legacy content: images uploaded before 2020, when few platforms enforced deduplication at ingestion. Second, multi-language versioning: a single promotional graphic re-uploaded with different filenames for English, Mandarin, Malay and Tamil text overlays — common practice across HDB's estate portal and the CPF Board's digital communications archive. Third, vendor-supplied product photography duplicated across dozens of merchant storefronts on the same marketplace. Industry estimates from the Singapore Computer Society's Digital Infrastructure Committee suggest that for a mid-sized e-commerce operator running roughly 50,000 active SKUs, duplicate images can account for between 18 and 24 percent of total image storage volume.

Storage costs in Singapore's data centre corridor — which runs through Jurong East and Upper Changi Road East — averaged S$0.023 per gigabyte per month for enterprise cloud-tier storage in the first quarter of 2026, according to pricing benchmarks published by the SGTech industry body. That sounds negligible per file. Multiply it across millions of duplicates and a mid-sized retailer operating its own content delivery infrastructure faces a recurring overhead of between S$8,000 and S$22,000 annually in pure storage waste, before accounting for CDN egress fees.

The National Library Board encountered a version of this at scale. Its digitisation programme, which has processed more than 1.3 million items from the Singapore Memory Project archives at the Victoria Street headquarters, discovered during a 2024 internal review that automated scanning had generated duplicate image variants for approximately 7 percent of digitised materials — mostly because multiple scan passes were logged as separate assets rather than overwritten. The affected files ran to roughly 900 gigabytes.

The Deduplication Push and What Comes Next

A coordinated response is forming. GovTech Singapore, based at Sandcrawler Building near one-north, rolled out a perceptual hashing tool across the Whole-of-Government Application Analytics platform in January 2026. Perceptual hashing detects visually similar images even when filenames, metadata and file sizes differ — capturing the near-duplicates that byte-level deduplication misses. Early results from the pilot, covering six ministries, reportedly removed or flagged for review more than 310,000 image assets within the first 90 days.

For businesses outside the public sector, the practical steps are straightforward but require upfront investment. An image deduplication audit using tools that run perceptual hash comparisons across an entire asset library typically costs between S$3,500 and S$12,000 for a full engagement through a local digital agency, depending on library size. The Digital for Life movement, a national initiative run in partnership with IMDA, has begun incorporating asset hygiene into its SME digital health check programme, which reached over 28,000 businesses in 2025.

The cleaner the image library, the faster the site. In a city where the median mobile page-load expectation among consumers sits below 2.5 seconds — a threshold benchmarked in a 2025 Google-commissioned study on Southeast Asian user behaviour — the gap between a well-maintained asset database and a bloated one is increasingly the gap between a sale completed and a cart abandoned.