San Francisco's Duplicate Photo Problem: Why Bad Data Wastes City Resources and Hurts Residents

San Francisco's municipal databases are quietly drowning in duplicate images. Scanned building permits on file with the Department of Building Inspection, ID photos cycling through the Department of Homelessness and Supportive Housing's case management software, and property records maintained by the Assessor-Recorder's Office all carry redundant files that inflate storage costs, slow processing times, and—critically—delay services that Mission District families and SoMa shelter residents depend on every day.

The issue has taken on new urgency in 2026. The city is pushing a housing production emergency agenda meant to fast-track approvals along the Market Street Octavia corridor and in the Sunset District, where rezoning under state housing mandates has put permit volumes on track to hit levels not seen since the post-2008 reconstruction boom. Duplicate image files buried inside permitting workflows can trigger system errors that kick applications back to the start of the queue—sometimes by weeks.

How Duplicate Images Clog City Hall's Pipes

The mechanics are mundane but the consequences are not. When a contractor at 49 South Van Ness—the new city office complex that consolidated several departments—submits a building application, staff scan floor plans, site photos, and owner identification documents. Older intake systems at DBI lacked automatic deduplication logic, meaning the same JPEG could be saved under multiple file names across shared drives. By some estimates from city IT assessments conducted in early 2025, redundant files accounted for roughly 18 percent of total storage consumption across several legacy municipal platforms, though the city has not published a comprehensive audit covering all departments.

At Larkin Street Youth Services, a nonprofit serving homeless young people in the Tenderloin, staff who interface with city case management databases have long flagged that repeated uploads of client photos—required for eligibility verification—cause lag times in the Coordinated Entry System. Each duplicate image has to be manually resolved before a bed assignment can be confirmed. During a crisis intake surge, that lag can mean a young person waits hours longer than necessary on Turk Street in 40-degree fog.

The Department of Technology, headquartered at 1 Dr. Carlton B. Goodlett Place, has been piloting a deduplication protocol as part of the broader Digital Services Modernization initiative that the Mayor's Office of Innovation flagged as a budget priority in the fiscal year 2025-26 spending plan. The protocol uses hash-matching algorithms to identify identical image files before they are written to disk a second time. Early internal benchmarks from a six-month test run covering the Recreation and Park Department's permitting system showed a 23 percent reduction in storage overhead—a modest but measurable win that the city hopes to replicate across agencies.

What Residents Can Do—and What Comes Next

For ordinary San Franciscans, the practical upshot is this: if you have submitted documents to any city agency—DBI, the Planning Department at 49 South Van Ness, or the Assessor-Recorder's Office at City Hall Room 190—and received confusing error messages or inexplicable delays, a duplicate image conflict may be part of the explanation. Staff at those offices can manually pull and reconcile file records, but residents have to ask explicitly. Simply resubmitting photos tends to compound the problem rather than fix it.

Housing advocates at the Council of Community Housing Organizations have been pressing the Department of Technology to expand the deduplication rollout to the permitting modules most directly connected to affordable housing approvals. Their concern is that processing slowdowns disproportionately hit community land trust applications and nonprofit developers who lack the legal staff to shepherd stalled files through bureaucratic review.

The Department of Technology has indicated the expanded rollout is scheduled for completion by December 2026. Whether that timeline holds will depend partly on how aggressively the city funds the initiative in the next budget cycle—a debate that will play out at the Board of Supervisors this fall. Until then, the duplicate images accumulate, and the wait for a permit, a shelter bed, or a city service stretches just a little longer than it should.