HUD system performance measures, explained

The System Performance Measures are seven numbers that describe a whole homeless response system rather than any single project inside it. Your Continuum of Care compiles them from HMIS data across every participating provider and submits them once a year through HDX 2.0, covering the federal fiscal year from October 1 to September 30.

Which is why most providers file them mentally under someone else's problem. That is the mistake this page is about.

Why a system measure lands on your desk

HUD requires CoCs to run a local competition that ranks projects, and it sets a floor on how that ranking is built. In the FY2026 CoC Application, a CoC has to show that at least 25 percent of the total available points in its local process were based on system performance measures when reviewing, selecting and rating project applications.

What that means in practice

At least a quarter of the score that determines where your renewal sits on the priority list comes from performance data. Your project's contribution to these measures is not an abstraction. It is a line in a scoring rubric, and the projects at the bottom of a ranked list are the ones exposed if the CoC does not get everything funded.

The FY2026 application also names specific metrics in its own questions. Metric 1.2 on length of time homeless, Metric 5.2 on first-time homelessness, Metric 7a.1 on street outreach outcomes, and Metric 7b.1 on exits to permanent housing. Those four are worth knowing by number, because your CoC will be asked about them directly and will look to its providers for the answer.

The seven measures

Each one draws from particular project types. If your organization does not run that project type, that measure is not counting you, and it is worth knowing which of the seven you actually appear in.

MEASURE 1 Length of time persons experience homelessness

Average and median days homeless. Metric 1a counts only the days recorded in HMIS. Metric 1b adds self-reported time homeless before entry, and time spent in permanent housing before a move-in date. The two produce different numbers on purpose, and 1b is almost always the larger one. Emergency shelter · safe haven · transitional housing · permanent housing for 1b

MEASURE 2 Returns to homelessness after a permanent housing exit

Takes everyone who exited to a permanent destination two years ago and asks how many came back within 6, 12 and 24 months. Because it looks two years back, this year’s number is a verdict on work your team did in a previous funding cycle. Street outreach · emergency shelter · transitional housing · safe haven · permanent housing

MEASURE 3 Number of people experiencing homelessness

Metric 3.1 is the change in Point-in-Time counts, sheltered and unsheltered, entered by hand rather than pulled from HMIS. Metric 3.2 is the change in the annual count of sheltered people in HMIS, unduplicated. Emergency shelter · safe haven · transitional housing

MEASURE 4 Employment and income growth

Six metrics, comparing income at entry against income at annual assessment or exit. Earned income, non-employment cash income, and total income, each split between people still enrolled and people who left. Adults only, and only in CoC-funded projects. Stayers count only once they have been enrolled 365 days. Transitional housing · safe haven · all permanent housing types, CoC funded only

MEASURE 5 People becoming homeless for the first time

Counts entries by people with no prior enrollment anywhere in HMIS in the preceding 24 months. Metric 5.1 covers shelter and transitional housing. Metric 5.2 adds permanent housing, and it is the one the CoC Application asks about. Emergency shelter · safe haven · transitional housing · permanent housing for 5.2

MEASURE 7 Successful placement and retention

Metric 7a.1 is the rate at which street outreach clients exit to acceptable destinations. Metric 7b.1 is the rate at which people leaving shelter, transitional housing or rapid rehousing exit to permanent housing. Metric 7b.2 covers permanent supportive and housing-only projects, and counts people who stayed housed as well as those who exited to another permanent destination. Street outreach for 7a · shelter, TH, RRH for 7b.1 · PSH and housing only for 7b.2

MEASURE 6 Category 3 prevention and placement

Applies only to communities designated as high-performing under Category 3 of HUD's homeless definition. Most CoCs never touch this one. If yours has not told you it applies, it does not.

The tension nobody puts on the slide

HUD says the measures are interrelated and should be read against each other. That is a mild way of saying some of them pull in opposite directions, and a program can improve one by damaging another without anyone deciding to.

The clearest pair is Measure 1 and Measure 2. Measure 1 rewards moving people out of homelessness quickly. Measure 2 punishes exits that do not hold. A program under pressure on length of stay can exit people faster into thinner placements, look better this year on Measure 1, and see the cost land on Measure 2 two years later, in a different funding cycle, when nobody connects the two.

Measure 5 has its own version of this. First-time homelessness counts people with no prior HMIS enrollment in 24 months. It rises when prevention is failing, and it also rises when your system starts reaching people it was previously missing. A number going up is not automatically bad news, and a CoC that improves its outreach should expect to see exactly that.

If you present these to a board

Never show one measure alone. A single number out of context invites the wrong conclusion, and the wrong conclusion here usually means pressure to do the thing that damages the next measure. Show Measure 1 next to Measure 2, and say plainly that fast exits only count if they hold.

HUD's own guidance adds one more caution worth repeating: comparisons should be made between projects serving similar populations. Youth programs, domestic violence programs and programs serving chronically homeless people face different circumstances, and ranking them against each other on a shared measure produces a fair-looking number and an unfair result.

Which of your fields feed which measure

This is the part you can act on. Four data elements carry most of the weight, and each one is a field somebody either fills in or does not on an ordinary afternoon.

The field	What it moves, and what happens when it is blank
Exit destination	Drives Measure 7 entirely and starts the clock on Measure 2. A blank or unknown destination does not sit out the calculation. It counts as a placement that was not permanent.
Housing move-in date	Decides whether a permanent housing enrollment counts under 7b.1 or 7b.2, and separates time spent housed from time spent still homeless in Measure 1b. Missing it can put a housed person in the homeless column for months.
Income at entry, annual and exit	All six metrics in Measure 4 are a comparison between two points. Miss either point and the client contributes nothing, so a program doing real income work can post a flat number.
Project type on the HMIS project	Set once, years ago, and rarely revisited. It decides which measures include you at all. A project typed wrong is not scored badly, it is scored in the wrong place, which is harder to notice and harder to explain.

One number worth knowing

In the FY2026 CoC Application, a CoC can earn up to 2 points by showing that the rate of successful exits to permanent housing from emergency shelter, safe haven, transitional housing and rapid rehousing is at least 50 percent.

That is a system-level threshold, not a project-level one, and no single provider clears it alone. It is still the most useful benchmark to hold your own exit data against, because it is the number your CoC is trying to reach and it is computed from the destinations your staff record.

What to actually do about it

Four things, in order of how much they change the number relative to the effort.

Close exits properly, every timeAn exit with no destination is the single most expensive blank field in this whole framework. It damages Measure 7 immediately and removes the person from Measure 2 entirely, so you lose both the score and the ability to learn anything.

Check your project types once a yearTen minutes, once, against what the grant actually funds. This is the cheapest correction available and the one nobody schedules.

Record move-in dates the week they happenNot at exit, not at the annual assessment. A move-in date entered late is not wrong in the file, but it was wrong for every day it was missing, and Measure 1b is counting those days.

Ask your CoC for your own sliceMost HMIS Leads can produce the measures filtered to your projects. Almost nobody asks. Seeing your own contribution once is usually enough to make the previous three items feel worth doing.

None of this is a reporting exercise. Every item above is a question about what happens in the hour after a client leaves, which is the hour when the shelter is full and someone is already waiting at the desk.

About COMPASS

We make case management software for organizations doing this work. It captures exit destinations, move-in dates and income at each assessment point as part of the session a worker is already writing, and it names the required fields missing from a record before an export is generated. It does not calculate these measures. Your CoC does that.

That is the only pitch in this guide. The rest is true whatever system you run.

Something here not matching what your CoC is showing you?

Write to us. A person answers, and if this page is wrong we will fix it.

hello@illumipath.io

Sources

Requirements change and local competition scoring varies by CoC. This page reflects what HUD published as of August 2026 and we update it in place. Your own CoC's rating factors are the authority on how your project is scored.