Reference

Which report needs which HMIS field

One set of enrollments goes out to several funders every year, in several different shapes, on several different clocks. None of the shapes are wrong and none of them agree. This page lays out which report each funder wants, what each one is built from, and what is actually inside the export your HMIS hands you. Every fact below links to the source it came from.

The recurring reports

Two of these you file yourself. Two your Continuum of Care files, using data you sent upstream. The last one is however many reports your non-federal funders ask for.

CoC APR

Continuum of Care Annual Performance Report

Required by: HUD, from every recipient of CoC Program grant funds
How often: Every operating year, keyed to each grant's own operating year rather than the calendar
Filed through: Sage, the HMIS reporting repository, as a CSV upload your HMIS generates

Built from the standard HMIS data elements for the projects on that grant: who was enrolled, when they entered and left, where they went, income and benefits, disabling conditions, services.

Read the source

ESG CAPER

Emergency Solutions Grant Consolidated Annual Performance and Evaluation Report

Required by: HUD, from ESG recipients — states, metro cities, urban counties and territories
How often: Annually, on the grantee's program year
Filed through: Sage, also as a CSV upload; form CR-60 is filed separately in the Consolidated Plan

Not a section of the APR and not a longer version of it. It is a separate report for a separate funding stream, on a different clock, built from much of the same underlying data.

Read the source

LSA

Longitudinal System Analysis

Required by: HUD, from the Continuum of Care rather than from individual agencies
How often: Annually
Filed through: HDX 2.0

It replaced the older AHAR data collection and feeds HUD's Annual Homeless Assessment Report to Congress. Your agency's data reaches it through the CoC, so what you submit upstream shapes it.

Read the source

PIT and HIC

Point-in-Time count and Housing Inventory Count

Required by: HUD, from the Continuum of Care
How often: Annually on one night in January for the sheltered count and the bed inventory; the unsheltered count is required in odd-numbered years
Filed through: HDX

A snapshot rather than a period. The same beds and the same enrollments you report over a year appear here as a single night, which is why the totals will never match your APR and are not supposed to.

Read the source

State reporting

Whatever your state housing agency asks for, in its own shape

Required by: State agencies, counties, cities and private funders
How often: On each funder's own schedule
Filed through: Usually a document or a spreadsheet, not a data upload

Washington's Department of Commerce, for one, reports to its legislature under RCW 43.185C by combining HMIS data with expenditure data into narrative and tables that look nothing like a Sage upload. Most states have an equivalent.

Read the source

What is inside an HMIS CSV export

The current specification took effect on October 1, 2025 and defines twenty-four files in one zip archive. Most reporting questions are answered by two of them: Enrollment.csv, one row per enrollment, and Exit.csv, one row per exit, joined on EnrollmentID.

HUD HMIS CSV Format Specifications

Export.csv
Organization.csv
User.csv
Project.csv
Funder.csv
ProjectCoC.csv
Inventory.csv
Affiliation.csv
HMISParticipation.csv
CEParticipation.csv
Client.csv
Enrollment.csv
Exit.csv
IncomeBenefits.csv
HealthAndDV.csv
EmploymentEducation.csv
Disabilities.csv
Services.csv
CurrentLivingSituation.csv
Assessment.csv
AssessmentQuestions.csv
AssessmentResults.csv
Event.csv
YouthEducationStatus.csv

The fields that carry the most reporting weight are EntryDate, HouseholdID, RelationshipToHoH, LivingSituation, DisablingCondition and MoveInDate in Enrollment.csv, and ExitDate with Destination in Exit.csv. Almost every housing outcome a funder asks about is some question about those last two.

What Grantwell reads

For the sake of being specific: these are the fields Grantwell looks for in a file, and what each one is used for. It recognizes the HUD export by name and also plain enrollment spreadsheets, where the column headings are rarely the official ones.

Entry date: Enrollment.csv · EntryDate — The one field a file cannot be read without
Exit date: Exit.csv · ExitDate — Left blank for anyone still enrolled
Destination: Exit.csv · Destination — Where someone went, for housing outcome figures
Program: Project.csv · ProjectName — How figures are split by project
Household: Enrollment.csv · HouseholdID — Household counts, distinct from people
Person: Client.csv · PersonalID — Counted, then discarded — never stored
Date of birth or age: Client.csv · DOB — Age at entry, computed and then discarded
Veteran status: Client.csv · VeteranStatus — For funders who ask for it
Disabling condition: Enrollment.csv · DisablingCondition — For funders who ask for it

One thing Grantwell does not do, so that it is clear: it does not file anything. The Sage and HDX submissions are uploads of your HMIS’s own generated CSV and they should stay that way. Grantwell is for the reports underneath those — the state agency, the county, the United Way chapter, the family foundation — where somebody has to read the numbers and write them out again in someone else’s format.

If a report like that is on your desk now, there is a finished one you can read without an account and without uploading anything, and your first one is free.