Here's How San Francisco Counts Unhoused Residents

Elester Hubbard sees San Francisco’s housing crisis up close on a daily basis. As a supervisor for San Francisco’s homelessness outreach team, he works every day to connect unhoused people with social services.

But late in the night on Tuesday, while winding through residential and industrial areas of the Bayview neighborhood, Hubbard had a realization.

“We see homelessness all the time. But actually counting it, and just seeing those numbers go up and keep growing, you really see how there’s always work to do,” he said.

Hubbard, who grew up in San Francisco, was among dozens of workers who fanned out across the city to pursue an almost impossible task: estimate the number of unhoused people in the city.

Known as the point-in-time count, the survey measures the number of people staying in shelters, as well as those who are living in cars, tents, sleeping bags or other places not meant for human habitation. The U.S. government requires all local regions and counties to participate in the count every two years in order to receive federal dollars for homelessness services.

Preliminary data from San Francisco’s survey will be released in Spring 2024, and the full report will be available this summer.

Last time around, in 2021, San Francisco and the many Bay Area counties canceled the count due to the COVID-19 pandemic. In 2022, another winter surge of COVID-19 delayed the count from its usual January date to February.

Ultimately, city workers tallied 7,754 people experiencing homelessness two years ago — a 3.5% decrease from 2019. Also, during that period, the number of people staying in San Francisco’s shelters, rather than sleeping outside, increased as the city simultaneously grew its inventory of shelter beds.

In any year, the survey is widely considered an undercount because counters have to make judgment calls and can also easily overlook people who are couch surfing or simply hidden from plain sight.

“There are several well-known limitations of the point-in-time count,” said Margot Kushel, director of the Benioff Homelessness and Housing Initiative at UCSF. “It doesn’t do a good job counting young adults and homeless youth, or people who are cycling in and out of homelessness.”

But the information from the count is still valuable, Kushel said, especially when compared to other data sets.

Last year, Kushel and a team of researchers at UCSF released the largest of its kind study on homelessness in California, which asked thousands of unhoused people in-depth questions about how they became homeless, their health, drug use, family history, work history and more.

“It’s easy to understand that there are way too many people experiencing homelessness, but it would help us to understand what is happening in people’s lives right before they’re homeless,” Kushel said. “What are the signals just before someone becomes homeless, and how can we get them legal services or funding to pay back rent? How do we end their homelessness quickly?”

How San Francisco conducts its point-in-time count

In San Francisco, teams of workers and volunteers — most of whom work within the city’s homeless services agencies — gathered downtown around 7:30 p.m. on Tuesday to hear instructions. Each crew was handed a detailed map of streets and alleyways to scan for signs of life, along with instructions to record findings in a mobile app.

Counters did not conduct interviews with people they identified but were given tips before the assignment on visual cues to look out for and others to be wary of.

“You might see people laying down, count them. People walking around, count them,” Jose Torres, program manager for the SF Homeless Outreach Team, said to surveyors before they took off for the night. “If they are carrying a lot of stuff and bags, that might tell you the person is unhoused. Shoes can tell you a lot about a person too, seeing if they have really old shoes or no shoes at all.”

Groups took off by foot in densely populated areas like the Tenderloin and Mission. But out in the sprawling Bayview neighborhood, where RVs are much more common, Hubbard and his team drove in a city vehicle to cover more ground.

Immediately upon getting to the group’s map area, the team spotted a line of RVs with tents and other makeshift shelters scattered between a FedEx distribution center and Amazon warehouses.

“I’m going to say that RV is confirmed because there’s music playing inside,” said Sean Bullock, an outreach specialist with SFHOT who was riding in the passenger seat while recording the data. At another row of RVs, the team stopped beside a smaller trailer packed with items but dark inside. “That looks like storage; let’s move on,” Hubbard said.

Block by block, the team scanned the streets, making these types of judgments over whether a person or multiple people may be living in a vehicle or tent. But some calls were easier than others. As the team drove slowly under a freeway overpass, four people stood beside their tents and huddled around a small fire to keep warm.

After nearly four hours, Hubbard’s crew called it a night. They recorded 257 encounters, including individuals confirmed and vehicles and RVs that appeared to be lived in.

Unlike counties that conduct their point-in-time count in the early morning hours, San Francisco does its count at night. It also relies on trained workers, as opposed to a volunteer crew, to perform the tallies. Other cities across the country, and even across California, do it differently — another challenge researchers have with the accuracy of the count.

In New York City, 95% of people experiencing homelessness are in shelters, according to Kushel, whereas the majority of unhoused people in San Francisco are unsheltered. And in more rural parts of the country, “it’s way easier to not be counted or found,” Kushel said.

Looking at highlighted maps and the final tally on his phone, Hubbard said the count itself made him appreciate the work that’s done beyond just number-gathering.

“We’ll come back 10 times to build rapport with a client to make progress,” Hubbard said. “Some people may think that’s wasted time, but we don’t think of it as that. That’s how we get results with clients we work with.”