How to Count People Who Don't Want to Be Found

The Challenge

Every January, communities across America conduct Point-in-Time counts of people experiencing homelessness. The sheltered count is relatively straightforward—you have rosters. The unsheltered count is something else entirely.

How do you count people in encampments you don't know exist? People who move every night? People who actively avoid being counted because they fear authorities?

What We've Tried

Our first counts were terrible. Volunteers walked predetermined routes between midnight and 6 AM, counting everyone they saw who "appeared homeless." The bias was obvious in retrospect.

Now we combine multiple approaches:

• Street outreach workers survey their regular contacts weeks before the count
• We interview clients at soup kitchens and day centers to understand where unsheltered people actually stay
• We build relationships with encampment residents who help us find hidden sites

The Accuracy Problem

Here's the uncomfortable truth: every count is an undercount. Some people aren't outside on count night. Some are in areas we don't cover. Some refuse to participate.

The solution isn't pretending our numbers are perfect. It's being transparent about methodology so comparisons between years are meaningful. Did homelessness go up, or did we just get better at counting?

Why It Matters Anyway

Flawed counts still matter. They drive HUD funding. They shape local policy. They tell stories that motivate communities.

Our job is to make them less flawed every year—and to honestly describe their limitations.