Pathfinders for Hope · Operation Iron Gate North Bay Pilot

How Can a City Measure Whether Homelessness Referrals Worked?

To measure homeless referral outcomes, define the outcome before calculating a percentage. Referrals sent, referrals acknowledged, appointments attended, and services delivered are different measures. A city needs to know which one is reported, who verified it, and how pending or unknown results are handled.

Separate activity, coordination, and service outcomes

Activity: how many referrals entered the defined workflow.
Coordination: how many received an authorized acknowledgment, had an assigned next action, or remained unanswered.
Service outcomes: which specific event was confirmed, such as an assessment attended or a service started.
Longer-term outcomes: changes such as housing stability, which require their own definitions, evidence, and appropriate follow-up.

Do not move between these categories without saying so. An improved acknowledgment rate may show better communication. It does not by itself prove that more people obtained housing or that a program caused a change in homelessness.

Specify the denominator and reporting window

For a proposed receipt-confirmation rate, use the number of referrals with an authorized acknowledgment divided by the number of in-scope referrals sent in the same defined cohort. State which requests are included, how duplicates or corrections are treated, and the date through which responses were checked.

For a response-time measure, define the start event and the qualifying response. Say whether elapsed time or working time is used. Report pending and unanswered volume alongside the statistic so fast responses do not conceal unresolved cases.

A worked example with fictional numbers

Illustration only; these are not Pathfinders for Hope results: a city examines 40 in-scope referrals sent during a particular month. By its stated cutoff, 28 have an authorized receipt acknowledgment. The receipt-confirmation rate is 28 divided by 40, or 70%. The other 12 remain unconfirmed at that cutoff.

Of those 28 acknowledged referrals, suppose 18 have an assessment scheduled and 10 are still under review. Neither figure establishes that anyone attended an assessment or obtained housing. If the report later learns about attendance, that is another measure with a clearly defined population and verification rule.

Keep unknowns and person-directed choices visible

Use distinct categories for pending responses, an unavailable service, an option the person declined, and an outcome the team cannot verify. Unknown is an information state, not proof of failure. Declining an option is not automatically a negative outcome for the person.

Avoid a single league table that rewards organizations for selecting easier requests or closing records quickly. Compare like-for-like service scopes and time periods. Include enough context to explain a change in demand, partner participation, or data completeness without disclosing personal information.

Publish cautiously and improve the workflow

Have the responsible data and privacy teams review public aggregates, especially small counts and detailed subgroups. Removing names alone may not remove identification risk. Keep case-level information in approved systems with appropriate access.

A useful monthly review ends with a practical question: which communication or routing problem can the participating teams investigate next? Track whether that change is implemented, then assess the next comparable period. Do not describe a before-and-after difference as proven causation without an appropriate evaluation design.

Five ways to measure homeless referral outcomes

A practical scorecard needs more than one success percentage. To measure homeless referral outcomes responsibly, report several linked stages and keep each denominator visible. The five measures below can help a city or participating organization distinguish communication performance from service delivery and longer-term change.

1. Confirm referral receipt

Start with the share of in-scope referrals that received an authorized acknowledgment by the reporting cutoff. State who may send that acknowledgment, what counts as confirmation, and whether automated notices qualify. The denominator should include every referral in the defined cohort, including those still unanswered. This measure helps teams measure homeless referral outcomes at the receipt stage. It shows whether a handoff entered a visible workflow; it does not show that an appointment occurred or a service was delivered.

2. Identify the next action and owner

Measure how many acknowledged referrals have a documented next action, a responsible team, and a target follow-up date. Useful categories may include review pending, information requested, appointment offered, rerouted with authorization, or unable to proceed for a stated reason. Reporting the unresolved group separately helps measure homeless referral outcomes without allowing a high receipt rate to hide stalled work. The owner is a role or participating organization, not a public disclosure of a staff member or the person seeking help.

3. Record barriers without blaming the person

Count barriers that affected the next step, such as unavailable capacity, incomplete authorized information, transportation, communication access, eligibility questions, or a person choosing not to continue with the offered option. Use neutral categories and allow more than one barrier when appropriate. A person-directed choice should not be mislabeled as noncompliance. When teams measure homeless referral outcomes, barrier reporting is most useful when it leads to a concrete workflow review rather than a ranking of people or providers.

4. Verify service milestones separately

To measure homeless referral outcomes at the service stage, define each milestone before reporting it. Examples include an appointment scheduled, an assessment attended, an eligibility decision completed, or a service started. Specify the authorized source used to verify the milestone and the date through which the record was checked. Do not collapse these events into a single “successful referral” label. Each answers a different question, and none automatically establishes housing stability or a reduction in homelessness.

5. Review follow-up and durability

For services where follow-up is appropriate and authorized, define a later review point and the exact question being asked. A 7-day confirmation may assess whether the next contact occurred; a 30-day measure may examine whether an agreed service remained active. Longer windows can create more unknown outcomes, so publish the follow-up denominator and loss-to-follow-up count. Never imply that a short-term communication measure proves a durable housing outcome.

Build a transparent monthly reporting table

A clear table helps a city measure homeless referral outcomes while showing both progress and uncertainty. For each measure, include the cohort dates, numerator, denominator, percentage, pending count, unknown count, verification source, and reporting cutoff. If a definition changes, mark the change and avoid presenting the new series as directly comparable with earlier periods.

Cohort: the referrals included and the dates they entered the workflow.
Measure: the exact event being counted.
Numerator and denominator: both numbers, not only the percentage.
Status detail: completed, pending, unknown, unavailable, or declined where appropriate.
Verification: the authorized record or participating source used.
Cutoff: the date and time window through which outcomes were checked.
Limitations: missing participation, incomplete data, small counts, or other constraints.

For example, “28 of 40 referrals received an authorized acknowledgment by September 30; 12 remained unconfirmed” is more informative than “70% success.” The fuller statement shows what was measured and what remains unresolved. Small numbers and sensitive subgroups should be reviewed by the responsible privacy and data teams before publication.

Questions to ask before interpreting the results

Are the measures answering different questions?

Receipt, next action, attendance, service start, and housing stability belong in separate fields. Combining them can make a dashboard look simple while obscuring where a workflow is functioning or breaking down.

Are pending and unknown outcomes visible?

Exclude neither group silently. Pending means the reporting window has not resolved the status; unknown means the authorized information is not available. Both affect how readers should interpret the percentage.

Are comparisons genuinely comparable?

Compare the same service scope, cohort rules, and reporting window. Differences in demand, capacity, partner participation, or data completeness can affect results. A before-and-after change is a signal to investigate, not automatic proof that one intervention caused the difference.

Can teams act on what the measure reveals?

Each monthly review of how teams measure homeless referral outcomes should end with a named workflow question, a responsible participating team, and a date to revisit it. The purpose of the scorecard is to improve communication and follow-through while protecting privacy – not to create a public league table.

Use authoritative standards and local definitions

Any plan to measure homeless referral outcomes should align with the systems and rules already used by participating organizations. The Agency for Healthcare Research and Quality care-coordination guidance provides a useful federal reference for referral and follow-up practices. A local government example, the Nashville Office of Homeless Services HMIS overview, explains how an HMIS supports community-level information and planning.

These references do not create a universal metric set. Local legal, privacy, program, funding, and data-governance requirements control. Pathfinders for Hope presents this framework as a planning resource for organizational partners; it is not reporting actual pilot results and does not guarantee a service or placement.

Local standards and related resources

These are proposed operational measures, not substitutes for statutory reports, funder measures, or HUD requirements. Where HMIS applies, consult the current HUD data standards and the local HMIS lead before adding or redefining fields. Use HUD’s Continuum of Care program information for broader official program background.

Use the county tracking guide for a proposed operational record and the receipt-confirmation guide for milestone definitions. Partnership planning can help teams agree on definitions before collecting data.

Frequently asked question

Does a high referral acknowledgment rate prove better housing outcomes?

No. It measures confirmed communication within a defined workflow. Housing or service outcomes need separate definitions and authorized evidence, and causal claims need an appropriate evaluation.

The metric set follows from the standard being applied. See referral accountability and Operation Iron Gate.

Operation Iron Gate North Bay Pilot

Pathfinders for Hope is developing the pilot with organizational partners. Public intake is not open, and no service or placement is guaranteed. GATEHOUSE, our community platform and a smaller version of Operation Iron Gate, is open now: register on GATEHOUSE. Discuss the pilot with Pathfinders for Hope or browse the resource hub.

Prepared by Pathfinders for Hope. Updated September 5, 2026. Resource scope and sources.