San Francisco's City Websites Are Drowning in Duplicate Images — Here's What the Numbers Show

San Francisco's municipal web infrastructure is carrying an estimated 14,000 duplicate image files across city-managed domains, according to a digital asset audit completed in June 2026 by the Department of Technology's Office of Digital Services. The redundancy is costing the city roughly $340,000 annually in excess cloud storage and content delivery network fees — money that could otherwise fund Muni route improvements or Navigation Center beds in SoMa.

The audit matters right now because the city is in the middle of a $12 million overhaul of SF.gov, the consolidated public-facing portal launched in 2020 to replace dozens of fragmented department sites. As migrators pull content from legacy platforms into the unified Salesforce-backed system, duplicate images are being ingested twice, sometimes three times, inflating storage bills and degrading page-load speeds for residents trying to access permit applications, shelter waitlists, and transit maps.

Where the Bloat Lives

The problem is concentrated in a handful of high-traffic corners of the city's digital estate. The San Francisco Department of Building Inspection's online portal, which serves thousands of contractors and homeowners filing permits for projects ranging from Noe Valley basement conversions to Mission District seismic retrofits, accounts for roughly 2,100 duplicate image instances on its own. The San Francisco Public Library's digital collections system, administered from the main branch on Larkin Street in the Civic Center neighborhood, holds an additional 1,800 redundant file copies — many of them scanned historical photographs uploaded by staff across 28 branch locations without a centralized deduplication check.

The Office of Digital Services identified three root causes. First, content management system migrations between 2018 and 2022 lacked automated hash-matching — the process by which a system identifies two files as identical regardless of their names. Second, individual department webmasters, operating without shared asset libraries, routinely reuploaded the same press-release photos and infrastructure images. Third, the city's previous content delivery arrangement with Akamai did not flag duplicate origin files before caching them separately at edge nodes, compounding both storage and retrieval costs.

The Numbers Behind the Cleanup

Digital asset deduplication at municipal scale is not cheap upfront. The Department of Technology has budgeted $185,000 for a six-month remediation contract, which will use open-source perceptual hashing tools to cross-reference image libraries and flag near-duplicates — images that are technically different file sizes or formats but visually identical. Industry benchmarks suggest a project of this scope typically recovers 60 to 75 percent of wasted storage within the first 90 days of implementation.

For San Francisco, that math is meaningful. City servers currently hold approximately 2.4 terabytes of image data across sf.gov and its federated agency subdomains. If the remediation hits the midpoint of those benchmarks — about 67 percent recovery — the city would claw back roughly 1.6 terabytes of active storage, dropping its monthly cloud bill with Amazon Web Services by an estimated $22,000. Over a full fiscal year, that return eclipses the cost of the remediation contract itself.

The broader context is a city government that has been leaning harder than ever on its digital infrastructure. Pandemic-era service shifts pushed residents toward online portals for everything from applying to the Homekey affordable housing program to registering for Healthy SF. Web traffic to SF.gov climbed 38 percent between fiscal year 2021 and fiscal year 2025. More traffic hitting a bloated back-end amplifies the performance penalty of every redundant file sitting on a server.

Residents and contractors who interact regularly with city systems will likely notice results first in page-load times on the Department of Building Inspection portal, where the remediation is scheduled to begin in August 2026. The Office of Digital Services says a public-facing dashboard tracking storage reduction progress will go live on SF.gov by September 1. For anyone building a home addition in the Sunset or pulling a business license in the Tenderloin, a faster, leaner city website is the practical payoff from what amounts to a very unglamorous piece of digital housekeeping.