What is grant reporting?
Grant reporting is the work a funded organization does to show each funder how an award was used, what was delivered and what changed, counted the way the grant agreement defines it and sent on the funder’s schedule. A grant report is one submission in that cycle; the reporting itself runs from the day the agreement is signed.
For a program lead with several funders, the hard part is rarely the writing. Each funder sends its own template and deadline, and the numbers get rebuilt for each until nobody knows which count is right.
THE SHORT VERSION
- Record what each award asks, how it defines each number and when it is due, in one requirements register, the week the agreement is signed.
- Collect once, on one ID per person, and answer every funder’s template from the same records and the same definitions.
- Where a funder defines a number differently, send their number with a one-line label, and mark anything not yet measured as not available instead of estimating it.
What does a funder expect in a grant report?
A funder expects the results named in your grant agreement, counted the way the agreement defines them, for the right period and in its own format, with spending explained against the approved budget.
When an award begins, open a requirements register with one row per thing the funder asks for; the editable grant reporting evidence map gives you the columns as a spreadsheet.
| Field | The question it answers |
|---|---|
| Award and period | Which award and which dates does this submission cover? |
| Required result | Which objective, metric or question, in the funder’s own words? |
| Definition and source | How is the figure counted, and which records hold it? |
| Responsibility | Who collects, checks and approves it? |
| Submission rule | Which form, portal, deadline and sign-off apply? |
“Describe progress toward employment” becomes answerable once it has a definition, a time window, a population and a source. For U.S. federal awards, Grants.gov explains the recipient’s administrative, financial and programmatic reporting responsibilities; other funders set their own rules, so confirm unclear ones with the program officer.
How do you report to several funders without collecting everything twice?
Keep one evidence base, meaning the same records, the same data dictionary and one ID per person, and treat each funder’s template as a view of it rather than a new data collection.
Take Partner C, a fictional workforce training program from the free course and one of four partners of a fictional regional workforce fund. Its reports go to four readers: Funder A’s quarterly template, Funder B’s annual narrative, Funder C’s IRIS+ metrics and a two-page board update.
What joins a template to the records is a map: each funder’s ask in its own words, the dictionary metric that answers it, and a note wherever definitions differ. The chapter Map every funder’s ask to one evidence base builds it row by row against a shared data dictionary.
One rule keeps the map honest: never change your data to fit a template. If Funder A counts visits and your dictionary counts unique people, send visits in their field and add the unique count below, with one line saying which is which.
What does a grant reporting calendar look like?
A workable grant reporting calendar has four beats: map the requirements when an agreement is signed, check the numbers each quarter, assemble the annual reports from the same records, and revisit the map at renewal. Most deadline scrambles trace back to skipping the first.
| Beat | When | What you do | What you have after |
|---|---|---|---|
| Agreement signed | Week one | Match each ask to a dictionary row | Gaps named before any report is due |
| Every quarter | Two weeks before due | Run the counts, compare with the agreement | Numbers that match the agreed definitions |
| Annual | Each funder’s year-end | Write narratives from the same records, in each funder’s taste | Reports that agree with each other |
| Renewal | Before the next agreement | Reread guidance, update the map, raise gaps | Cleaner definitions next time |
When a funder’s year and your program year differ, the ID does the sorting. Girls Inc. of Metropolitan Dallas runs a coding program with pre, mid and post surveys and follow-ups at 6 months and 1 year, on a cycle tied to a July 1 fiscal year; each follow-up lands in the right period because every wave belongs to the same participant.
How do you write a grant report?
Use the funder’s required structure first; where you can choose, six sections cover what a program officer checks. Put the main findings near the top and supporting schedules where readers can find them.
- Executive summary. Award, period, progress, difficulties and any decision you need. Write it last.
- Population and reach. Who the work was meant to serve and who took part, as unique people unless the template asks for visits.
- Activities against the work plan. Delivery compared with the approved plan, with delays and changes explained.
- Outcomes against the objectives. Each agreed outcome with its definition, time window and the number of people it covers.
- Budget and expenditure. Finance-reviewed figures for the right period, with material variances explained. An underspend is not a saving until you know whether delivery also slipped.
- Learning and next steps. What will change, who owns it and when you will review it.
Keep each caveat beside its finding. The impact report writing guide and these report examples show finished layouts.
What does a grant progress report example look like?
A useful progress report sets each agreed metric beside the number sent and a one-line note wherever the definitions differ. Partner C’s quarterly report to the fictional regional workforce fund shows what happens without that note.
The fund’s agreement asked for unique people enrolled, placement in a job within 90 days of exit, retention in the same job at 12 months, and hourly starting wage by track. Partner C sent 80 people enrolled, placements counted “within 6 months”, no retention figure and one average wage.
| Metric | Agreement says | Partner C sent | What the report should say |
|---|---|---|---|
| Enrolled | Unique people | 80 people | Matches; send as is |
| Placed in a job | Within 90 days of exit | 55, within 6 months | The 90-day count; the 55 goes to Funder B, labelled |
| Retained | Same job at 12 months | Not reported | “Not yet available”, with a date |
| Starting wage | Hourly, by track | One average | One figure per track |
None of these fixes needs new data. Exit and placement dates already sit on one ID per person, so the 90-day count can be computed. Meanwhile the fund held C’s 55 out of its portfolio total and sent three questions a labelled report would have answered.
Can AI write a grant report?
AI can draft a grant report from records you choose, but a person checks every number against its source before anything is sent. The risk is a figure that looks like a count and was never counted.
In the course’s demonstration on a demo workforce dataset, the first AI draft flagged job placements and starting wage by track as missing, because the placements survey had not been selected as a source. Once it was added, the answer was complete.
In Sopact Sense, the AI Assistant stays locked until you declare which surveys it may use, and every line of its answer links to a record you can open. The chapter Write each funder’s report with AI, then check it covers the full loop.
PROMPT · PASTE INTO CLAUDE, CHATGPT OR YOUR AI TOOL
You are checking a grant report draft before it goes to the funder. Inputs: the funder's template and award terms [ATTACH]; our data dictionary [ATTACH]; our draft report [ATTACH]. 1. List every item the funder requires, quoting its wording and section. 2. For each, show where the draft answers it, or write "not answered". 3. For each number in the draft, say whether its definition matches the funder's. If it differs (people vs visits, 90 days vs 6 months, one average vs by track), draft a one-line label. 4. List any number in the draft that is not in the dictionary or records. Rules: do not invent or estimate any number. If a required figure is not in the inputs, write "not in our data". Return a table: Requirement | Where answered | Definition match | Fix.
The video below explains why reports built on clean data still get ignored. Watch for how the reader’s context changes what a funder takes from the same numbers.
What should you test in grant reporting software?
Test any grant reporting tool on one real cycle, including the awkward parts: two funders with different time windows, a late update and a revised budget. A convincing demonstration reproduces one reported figure from its source records; the grant management software guide covers wider selection.
| Test | What to try | A pass looks like |
|---|---|---|
| One ID | Link intake, exit and follow-up for one person | One record, nothing retyped |
| Definitions | Visits for one funder, unique people for another | Both from the same records, labelled |
| Missing data | Ask for a number you never collected | “Not in our data”, not an estimate |
| Tracing | Open a figure in a draft | It leads to the records behind it |
| Access | Give a colleague one program only | They see that folder and nothing else |
What should a grant report not claim?
A grant report should not present outputs as outcomes, treat a before-and-after change as proof the grant caused it, or describe a self-reported result as if everyone answered. These are the first things a careful program officer looks for.
Enrolments are outputs; placement and retention are outcomes. A rising placement rate may reflect the local job market as much as the program, and without a comparison group the honest verb is “changed”, not “caused”.
Follow-up surveys lose people, and those who answer may differ from those who do not, so give the number who responded beside every result. AI drafts are suggestions: a named person checks them against the record and decides what goes to the funder.
The video below shows what happens when a report claims more than its records support: months of work look convincing until a funder asks one hard question. Watch for what the answer depends on, whether each number traces to its records.
Start with one funder’s next report
Pick the report due soonest and run it through a requirements register before you write a sentence.
- List every funder you report to, plus your board, with format, rhythm and due dates.
- For the soonest report, copy each required item word for word into the register.
- Match each item to a dictionary row and mark it same, different or no match.
- Write the one-line label for every “different”, and a dated “not yet available” for every gap.
- Draft from the records, then check each number against its source before sign-off.
- Add the funder’s follow-up questions to the register for next cycle.
After the first cycle you have a register for one funder, labels you will reuse and gaps to raise at renewal.
Frequently asked questions
What is the difference between grant reporting and a grant report?
Grant reporting is the whole cycle behind an award: recording what the funder requires, collecting the evidence, checking numbers against the agreed definitions and sending reports on schedule. A grant report is one output of that cycle, such as a quarterly template or an annual narrative. Treating reporting as year-round work keeps each report from becoming a last-minute rebuild.
What should a grant report include?
Follow the funder’s template first. Where you choose, cover a short summary, who took part, activities against the work plan, outcomes against the agreed objectives, budget against the approved plan, and what you will change. Give each outcome its definition, time window and number of people covered, with any caveat beside the finding.
How do you standardize grant reporting across multiple funders?
Standardize the evidence, not the reports. Keep one set of records, one data dictionary and one ID per person, then map each funder’s questions to dictionary metrics. When a funder’s definition differs, such as visits instead of unique people, report their number with a one-line label. Do not change your own definitions to match whichever funder reported last.
Can grant reporting be automated?
Parts can: reminders, pulling the same counts each quarter, mapping a new template to your dictionary and drafting narrative from checked numbers. The judgment stays with people: whether a definition matches, how to explain a variance, and approval before submission. A good AI draft flags a missing number instead of estimating it.
How should we explain a missed target in a grant report?
Give the target and the actual result for the same period and definition, then say what you know about the gap and what you do not. Separate verified causes from possible ones, name the next step and its owner, and follow the award’s process for changes. A miss stated plainly reads as care; a miss hidden in an average loses trust.
