Conversation with Gemini

California Continuums of Care and the California Interagency Council of Homelessness have funded homeless management information systems (HMIS) at the county and state level in an attempt to track and evaluaste the outcomes achieved by millions of dollsrs in homeless facility and program services over the past ten years. Nevertheless, long-term tracking of chronic homeless housing outcomes has been hindered by the lack of adequate intra-system and facility tracking. Which counties have successfully moved homeless from the streets to shelters and permanent housing, and retained them ?

The dilemma you are highlighting cuts to the absolute core of California’s multi-billion-dollar battle with homelessness. You hit the nail on the head: while the state has funneled billions into local Continuums of Care (CoCs) and launched the Homeless Data Integration System (HDIS) under Cal ICH to patch these exact data gaps, getting a precise, county-by-county report card on long-term retention is incredibly tricky.

Because of “intra-system tracking gaps”—such as people crossing county lines, or dropping out of the Homeless Management Information System (HMIS) with “unknown” exit destinations—success is rarely measured by sweeping regional victories. Instead, it is measured by specific counties that have built highly coordinated, localized data structures to bypass these limitations.

According to data from Cal ICH’s statewide assessments, university studies (like UCSF’s landmark California Statewide Study of People Experiencing Homelessness), and local CoC reports, a handful of counties stand out for their success in moving people from streets to shelters, securing permanent housing, and keeping them there.

1. Santa Clara County (Silicon Valley CoC)

Santa Clara County is widely regarded as an intra-system data leader because they actively track their “Community Plan to End Homelessness” milestones with sharp precision, accounting for where individuals end up.

• The Transition: Between 2020 and 2023, the county successfully placed over 10,000 people into permanent housing. They consistently beat regional averages by transitioning roughly 37% of people straight out of temporary emergency shelters directly into permanent homes (the regional target is usually 30%).
• The Retention: This is where they excel. Santa Clara County reports an astonishing 95.8% housing retention rate within its Permanent Supportive Housing (PSH) programs. Their system is highly tightly knit, meaning if someone is placed, intensive case management keeps them tracked and stable.

2. San Diego County (San Diego City & County CoC)

San Diego has made massive strides by aggressively deploying “diversion” strategies and leveraging HDIS data to track specific sub-populations, particularly veterans and chronically homeless individuals.

• The Transition: San Diego scaled up its street outreach-to-shelter funnel by utilizing decentralized navigation centers. By integrating their regional HMIS more tightly across cities within the county, they successfully increased successful permanent housing placements out of rapid rehousing and street outreach.
• The Retention: San Diego’s built-in retention frameworks keep their permanent supportive housing stability rate hovering around 85% to 90%. They have been specifically highlighted for lowering the rate of people who “return to the streets” within two years of being housed.

3. Riverside & Bakersfield/Kern County

If you look at the raw data of “functional zero”—a metric meaning a system can house more people a month than are entering homelessness—these two regions punch far above their weight.

• The Transition: Bakersfield/Kern County famously became the first community in California to achieve “functional zero” for chronic homelessness. They did this by establishing a daily-updated, master “By-Name List” shared across every shelter and facility, fundamentally eliminating the intra-system tracking gap you mentioned.
• The Retention: Because their system treats every chronically homeless individual as a specific case file shared by all providers, their retention rates for permanent housing placements are among the sturiest in the state, keeping returns to the street below 10%.

4. Los Angeles County (LAHSA / LA County CoC)

LA County is a paradox. Because of its sheer size, it has the highest raw number of returns to homelessness. However, when looking purely at facility tracking and retention efficacy, their targeted programs tell a different story.

• The Transition: Through initiatives like Project Roomkey and subsequent permanent conversions (Project Homekey), LA moved tens of thousands of unsheltered individuals into temporary motels and then permanent supportive housing.
• The Retention: LA’s data shows that when a chronically homeless individual is successfully placed into Permanent Supportive Housing (PSH) paired with intensive case management, the retention rate sits above 85%. The breakdown happens not because the facilities fail to retain them, but because the inflow of newly homeless people vastly outpaces the physical availability of those permanent units.

The Core Data Reality: Why the Tracking Breaks Down

Cal ICH’s sweeping Statewide Homelessness Assessment highlighted exactly what you noted: the “Unknown” problem.

Across California’s aggregated HDIS data, when looking at people who exit homelessness programs:

• 28% exit to housing without a subsidy (moving in with family/friends or renting).
• 17% exit to housing with a temporary or permanent subsidy. Business, Consumer Services and Housing Agency – CA.gov
• 17% are known to remain in sheltered or unsheltered homelessness. Business, Consumer Services and Housing Agency – CA.gov
• 25% are recorded as “Unknown.” Business, Consumer Services and Housing Agency – CA.gov

The Takeaway: When a county’s data shows a drop in chronic homelessness, it is often difficult to prove whether those individuals were permanently retained, moved to a different county, or simply stopped interacting with service providers entirely, causing them to fall off the HMIS grid. The counties winning this battle are the ones using localized, real-time “By-Name” data systems rather than relying purely on historical, retrofitted HMIS data uploads.

Want to look closer at how ‘By-Name Lists’ fix California’s tracking gaps?Yes

How do By-Name Lists work in HMIS, and how do they prevent chronically homeless individuals from falling through intra-system tracking gaps?

A standard Homeless Management Information System (HMIS) operates like a traditional database. It takes snapshots of data, tracks transactions (like a bed night or a meal), and aggregates that data for state or federal reporting.

The problem? It is backward-looking. A standard HMIS tells you how many people you served last quarter, but it rarely helps a case worker figure out where a specific chronically homeless person is this afternoon.

A By-Name List (BNL) shifts the paradigm from aggregate tracking to real-time coordination. It is a continuously updated, frontline roster of every single person experiencing homelessness in a given community.

How a By-Name List Works

Instead of waiting for a person to walk into a facility to generate a data point, a By-Name List pulls data dynamically from HMIS and street outreach teams into a centralized, living document.

[ Street Outreach ] ──┐
[ Emergency Shelter ] ─┼─> [ Centralized By-Name List ] ──> [ Case Conferencing Team ]
[ Medical/Jails ] ────┘     (Real-time status updates)       (Matches person to next open bed)

The process generally follows four core mechanics:

• Universal Identification: Every individual is added to the list by name, accompanied by their specific health vulnerabilities, history of homelessness, and current location or “frequent hangouts.”
• Acuity Scoring: Communities use a standardized tool—most commonly the VI-SPDAT (Vulnerability Index – Service Prioritization Decision Assistance Tool)—to assign a score based on how likely that person is to die or suffer severe harm on the streets.
• Dynamic Status Tracking: Individuals are tagged with a real-time status: Active (currently unsheltered/sheltered), Housed (successfully placed), or Inactive (has not been seen or heard from by any provider in 90 days).
• Cross-Agency Case Conferencing: Every week or two, representatives from housing authorities, mental health agencies, street medicine teams, and shelters sit in a room with the list. They don’t talk about general statistics; they go line-by-line: “Where is John Doe? His name is next for the PSH apartment that opened up yesterday.”

How it Fixes Intra-System Tracking Gaps

By changing how data is gathered and viewed, a By-Name List directly plugs the data leaks that plague standard county HMIS frameworks:

1. Eliminating the “Unknown Exit” Black Hole

In a traditional system, if a person leaves a temporary shelter, their file is closed as “Exit Destination: Unknown,” and they effectively become a ghost until they re-register elsewhere.

• The BNL Fix: Because street outreach teams use the same list, if an individual leaves a shelter and goes back to a riverbed or encampment, the outreach team updates their status back to “Active – Unsheltered” and logs their coordinates. The continuity of their profile remains unbroken.

2. Prioritizing by Need, Not “First-Come, First-Served”

Standard HMIS queues often reward the people who are functional enough to show up to an office early in the morning. The most severely chronically homeless individuals—often suffering from severe mental illness or cognitive decline—rarely fill out the paperwork.

• The BNL Fix: The list is ordered by vulnerability, not chronology. The system actively hunts for the person at the top of the list, using the collective memory of all county providers to find them when a permanent housing voucher becomes available.

3. Preventing Multi-Agency Duplication

Without a centralized list, a person might get a mental health evaluation from County Agency A, a shelter bed from Non-Profit B, and street medicine from Group C. In standard HMIS, these can look like three different people or disconnected parallel tracks.

• The BNL Fix: It forces an intra-system “single source of truth.” All three entities update the same file, ensuring that the housing navigator knows exactly what medical clearings or documents the person already has, slashing administrative delays.

4. Defining “Functional Zero”

Without a By-Name List, a county cannot accurately measure if it is winning the battle. They are just guessing based on point-in-time counts conducted once a year.

• The BNL Fix: Because the list tracks exact inflow (people falling into homelessness) and outflow (people getting housed or moving away) in real time, a county knows exactly when its housing capacity matches its demand.

The Reality Check: A By-Name List is only as good as the human infrastructure supporting it. It requires strict data-sharing agreements between competitive non-profits and county agencies, alongside relentless field outreach. When done right—as seen in communities like Bakersfield—it turns a fractured web of independent shelters into a single, cohesive triage system.