Yes the point-in-time count is 8000, but a lot of homelessness services go to preventing people who are on the edge of homelessness.

If you use the estimate here (https://sf.curbed.com/2019/11/20/20973811/sf-homeless-count-...) that 18,000 people are homeless at some time in a given year then the figure goes down to 20,000 per homeless.

Also consider that an additional estimated 9,500 each year would be homeless if not for SF's existing supportive housing services (https://sf.curbed.com/2020/3/4/21152501/san-francisco-homele...)

Plus people being evicted. In 2019 there were 1600 eviction notices (https://sfrb.org/sites/default/files/Document/Statistics/18-...). If we assume that even just 25% of those are 4-person families, that's bringing us to around 10k a year per homeless/at-risk resident (and that's not even counting those who are at-risk of eviction but did not get evicted thanks to SF's existing eviction prevention resources.

This is a really good point, thank you for providing the links!

This is a misleading metric.

Say you had a $100k homeless budget, and 100 homeless folks. You’re spending $1k per resident.

But say your program REALLY works, and now you only have 10 homeless folks. Then, your program is spending $10k per resident!

the program didn’t get worse. As your program gets better, you should be spending MORE per resident, not less.

Another great point, it is entirely likely that happened in SF since I’ve heard the homeless population quoted as 20,000 a couple years ago. As the commenter above also pointed out, my original math was indeed misleading.