What to Put in the Impact Section of Your Grant Report
What funders ask for when a grant report reaches its impact questions, and how to answer with evidence beyond citations. You get the outputs, outcomes and impacts split, a test for evidence a program officer can check, and a worked before and after example.
The report form has a box labelled impact, and the cursor is blinking in it. You have a list of papers, two conference talks and a vague sense that a policy team read something. The papers go in another section. The vague sense is not evidence.
This post is about filling that box well: what funders actually ask for, what counts as evidence besides citations, and how to turn a list of activities into a few sentences a program officer can believe.
What the Report Actually Asks For
Impact questions look different from funder to funder, but they all want the same thing: who outside your immediate field was reached, and how you know. The details below are accurate at the time of writing (October 2026), so check the live pages and your award terms before you submit.
| Funder | Report | What it says about impact and reach |
|---|---|---|
| NSF (US) | Annual, final annual, and Project Outcomes Report | Project reports have an Impact component, annual reports cover broader impacts work, and the public outcomes report (no more than 800 words) is posted exactly as submitted |
| UKRI (UK) | Annual outcomes submission in Researchfish | Report outcomes from early in the project until at least five years after it ends; outcomes are made public on Gateway to Research |
| SSHRC (Canada) | End of grant report | Report on the outcomes and impact of the research, with outputs, outcomes and impacts defined separately |
| NIH (US) | Research Performance Progress Report (RPPR) | At least annually; asks about accomplishments, plans, and manuscripts and publications produced, among other items |
NSF
NSF's guide to award administration (the 24-1 edition its reporting page links to) lists the components of NSF project reports: accomplishments, products, participants and organizations, impact, changes and problems, and special requirements. Annual reports should address all activities, including broader impacts work that is not intrinsic to the research.
At the end comes the Project Outcomes Report for the General Public: a brief summary, generally two to three paragraphs, written for the lay reader, with PIs strongly encouraged to avoid jargon, terms of art and acronyms. NSF's reporting page caps it at no more than 800 words. NSF posts it exactly as submitted, with a disclaimer, and it is due no later than 120 days after the award ends. The same reporting page also asks you to tell your program officer about significant discoveries and broader impacts accomplishments without waiting for the annual report, and to give advance notice of high-profile publications or media coverage.
UKRI
UKRI's outcomes reporting page asks the lead researcher on a UKRI grant or fellowship for an annual update in Researchfish, usually as part of the award terms unless you have an exemption, from early in the project until at least five years after it ends. You can only submit once a year, during the submission window (1 February to 11 March 2027 for the next one), but you can update outcomes all year round. Outcomes are published on Gateway to Research, and UKRI says funding may be withdrawn if you do not report.
Between NSF's public outcomes report and Gateway to Research, assume that anything in a public summary will be read by people who never saw your proposal, and write it for them.
SSHRC
SSHRC's knowledge mobilization guidelines say all funding recipients must submit an end of grant report documenting how funds were used and the outcomes and impact of the research. The same page gives the clearest definitions of the three words most reports blur together.
NIH
NIH's reporting requirements page says the RPPR is submitted at least annually and asks about accomplishments, plans for the next year, manuscripts and publications produced, personnel, and challenges, among other things. Read your institute's instructions for where outreach and dissemination belong.
Outputs, Outcomes, Impact: Put Each Item in the Right Box
The most common weakness in an impact section is a list of outputs presented as impact. SSHRC's guidelines draw the line well: outputs are short-term results such as publications and event attendees, outcomes are what people do as a result (using the findings, training, new policies), and impacts are long-term changes in thinking and behaviour. Sorting the items from the worked example below shows how it works in practice:
| Item | Box | Why |
|---|---|---|
| Three papers and three plain-language summaries | Output | You produced them |
| 1,900 reads of the summaries | Output | Reach, not yet use |
| A briefing requested by wellbeing staff | Outcome | Someone acted on the work |
| A partner using your checklist in its workshops | Outcome | The findings are in use |
| Fewer students reporting exam-season sleep problems | Impact | Not measured yet, so it is not claimed |
A two or three year grant will rarely prove impact in the long-term sense, and reviewers know it. A report that is honest about outputs, shows real outcomes, and names where longer-term change might show up later reads as more credible than one that claims transformation.
Evidence Beyond Citations
Citations take years and mostly measure reach inside your field, so the impact box needs other evidence too: counts of who read, listened or watched, requests and invitations that came because of the work, and signs of use by people outside academia. Our guide to a knowledge mobilization plan lists the evidence types and how to plan for them.
The test for each item is simple: could a program officer check it? A number with a source, a named partner, a document or a letter passes. An adjective does not. SSHRC suggests developing indicators at the outset, such as newsletter or blog subscribers and recommendations to policy-makers that were adopted, which is the easiest way to make sure the checkable items exist when the report is due.
A Worked Example: Before and After
Say you led a three-year psychology project on sleep and exam stress in first-year university students, and you are writing the impact section of the final report. The numbers below are invented for illustration.
Before
Our findings have had significant impact on student wellbeing practice and have been widely disseminated to academic and non-academic audiences.
Nothing in that sentence can be checked, and every report says it.
After
We published three papers and a plain-language summary of each. The summaries were read about 1,900 times over the final year, mostly through links in two student services newsletters.
Wellbeing staff at two partner universities asked for a briefing after reading the summary, and one now uses our checklist in its exam-season workshops (letter of support attached).
The French audio version was opened about 240 times, most of them from the partner campus in Quebec, which matters because we committed to dissemination in both official languages.
We have not measured changes in student outcomes; a follow-up with the partner workshops is planned.
Each sentence is a claim, the evidence, and one line of context. The last sentence says what you do not know, which makes the rest easier to believe.
Collect It Before You Need It
None of the after version can be reconstructed from memory at report time. The 1,900 reads, the briefing request and the letter of support each had to be written down when they happened, by someone who knew the report was coming. Give that job to a named person on the team, and file the proof (an email, a newsletter issue, a screenshot) next to the entry.
UKRI lets you update outcomes in Researchfish all year round, and NSF wants significant accomplishments reported as they happen, so a running log also matches how funders prefer to hear from you.
Where Amplify Fits
Two lines in the after version depend on having something people outside the field could open, and on counting who did: the summary reads and the French audio. Amplify makes those materials from one paper (a one-page print-ready infographic, an audio brief in English with 21 more languages on request including French, a 45 to 60 second vertical short and a video abstract of about 3 minutes), generated from your PDF alone and released only once you approve them.
Put them beside the paper on your lab site or a partner page with the embeddable widget, and the analytics record reads, listens, questions asked, languages, and which format moved people. Those counts are the kind that go into lines like the reads and the French audio above. Where readers came from and what they did next still belong in your own log.
Write the Box You Would Want to Review
A program officer reading your impact section wants three things: who was reached, what they did, and how you know. Start the log now, keep outputs and outcomes in separate boxes, and leave the word "significant" out unless a number follows it.
Quick answers
What is the difference between outputs, outcomes and impact in a grant report?
Outputs are what you produced, outcomes are what people did with it, and impacts are long-term changes in thinking and behaviour. SSHRC's guidelines draw that line, and it sorts most report items: three papers are outputs, a partner using your checklist is an outcome. A two or three year grant will rarely prove impact in the long-term sense, and reviewers know it.
Are grant outcome reports made public?
Often, yes, so write them for a reader outside your field. At the time of writing (October 2026), NSF posts the Project Outcomes Report exactly as submitted, and UKRI makes research outcomes reported in Researchfish publicly available on Gateway to Research. Check your own funder's pages and award terms, and assume that anything in a public summary will be read by people who never saw your proposal.
How do I write an impact statement for a grant report?
Write each item as a claim, the evidence for it, and one line of context, and say what you have not measured. "Our findings have had significant impact" cannot be checked. "Wellbeing staff at two partner universities asked for a briefing after reading the summary, and one now uses our checklist" can, especially with the letter of support attached.
How long should the NSF Project Outcomes Report be?
No more than 800 words, according to NSF's reporting page at the time of writing (October 2026), and NSF's award guide says it should generally be two to three paragraphs written for the lay reader, without jargon, terms of art or acronyms. It is posted publicly exactly as submitted and is due no later than 120 days after the award ends.
