Outputs vs. Outcomes: The Difference Your Funder Is Really Asking About

I've seen the same grant report more times than I can count. It's a beautiful document. Glossy photos, a quote from the executive director, and a big bold number on page two: 1,200 people served.

And then the funder writes back with one question. "Great. What changed for them?"

Usually nobody in the room can answer it. Not because the program didn't work. Most of the time it did. But everything the team collected was about the program, and nothing was about the people. That gap has a name, and it's the most useful idea in nonprofit reporting.

The one-sentence difference

Outputs are about the program. Outcomes are about the people.

That's it. That line comes almost straight out of the United Way of America's 1996 manual, Measuring Program Outcomes, which defined outcomes as "benefits or changes for participants during or after their involvement with a program" and said plainly that merely attending a program is not an outcome (Education Northwest, summarizing the manual). That manual sold more than 100,000 copies and is a big reason funders started asking the question my funder asked (NASW Press).

The W.K. Kellogg Foundation's Logic Model Development Guide draws the same line. Outputs are "the direct products of program activities," things like classes taught, meetings held, participation rates and hours of service. Outcomes are "specific changes in attitudes, behaviors, knowledge, skills, status, or level of functioning" (Kellogg Foundation guide).

So when a funder asks "what changed?", they're asking for that second list.

Side by side, with real programs

Here's what the difference looks like in the kinds of programs I work with every week.

Those outcome numbers are illustrations of the shape, not real results from a real program. The point is the shape: every outcome names a group of people, a change, and a way you'd know.

Look at the left column again. Every one of those numbers is honest and hard-won. Somebody stayed late for those 41 distribution days. Outputs matter. They prove the work happened at the scale you promised. They're just not the answer to "so what?"

Outputs still count

I want to be clear about this, because I've seen teams overcorrect. Funders still ask for outputs, and they should. Government grants almost always want them. The Appalachian Regional Commission, for example, tells grantees they must report actual outputs and any progress toward outcomes in every interim report (ARC Grant Administration Manual).

Outputs answer "did you do what you said you'd do?" Outcomes answer "did it work?" A good report has both, in that order, and connects them.

The tricky middle cases

"People served" is an output, almost always. But the United Way manual itself gives the exception: if your program's whole purpose is getting people to seek a service, like getting someone into alcohol treatment, then the number who showed up can be the outcome (summary of the manual's example). Context decides.

Satisfaction is another one. "92% said the workshop was helpful" feels like an outcome, but it's really a measure of how the activity landed. It's a quality output. Useful, keep it, but don't let it stand in for change.

And there's impact, which people mix up with outcomes constantly. Kellogg describes impact as change at the organization, community or system level (Kellogg Foundation guide). Lower neighborhood unemployment. Fewer evictions countywide. Your program contributes to impact alongside a hundred other forces. Your outcomes are what you can actually show.

Writing an outcome that holds up

Every outcome I write has three parts, and if one is missing I keep going:

1. Who. Which participants? All of them, or the ones who finished?
2. What change. Knowledge, skill, behavior, condition or status. Use a verb that describes the participant, not the program.
3. How you'll know. The indicator, the question, the source and when you'll ask.

Here's the difference in practice:

• Weak: "Improve digital skills in the community."
• Better: "Participants will improve their digital skills."
• Strong: "At least 70% of participants who complete four or more sessions will be able to send an email with an attachment on their own at the 60-day check-in, measured by a short skills task."

The strong version is longer. It's also the only one you can report on honestly a year from now.

Turning the counts you have into outcomes

Most teams don't need a research study. They need one follow-up question asked at the right time. That's the move: find the output you already track, and ask what it was supposed to lead to.

• You count devices distributed. Ask at 60 days: "Are you still using it? What for?"
• You count workshop attendance. Ask at the end: "What can you do now that you couldn't when you walked in?" Then ask again in three months.
• You count referrals made. Track whether the referral was completed and what happened.
• You count resumes reviewed. Ask at 90 days: "Did you get an interview? An offer?"

Ask the same few questions at intake and at follow-up. Same wording, same scale, every cycle. That consistency is what turns a pile of surveys into a before-and-after you can stand behind.

Where this usually breaks

Honestly, it rarely breaks on the definition. It breaks on the data. The intake survey lives in one spreadsheet, the follow-up lives in a Google Form, the sign-in sheets are paper, and "completed" means three different things depending on who entered it. By report time, nobody can say which 212 rows are the same person.

That's the problem I built Storyhouse to solve. It holds the plan, the questions and the records in one locked place, asks the same questions the same way every cycle, and links every outcome back to the de-identified rows behind it, so a funder can click "68% connected" and see where it came from. You don't need it to do any of this. A clean spreadsheet and discipline will get you far. But if you're juggling five funders and three spreadsheets, that's what it's for.

Start here this week

1. List your outputs. You already know them.
2. Next to each, write the change it's supposed to cause, in a sentence about participants.
3. Pick the two or three that matter most to your mission and your biggest funder.
4. Write the indicator for each: who, what change, how you'll know.
5. Add one follow-up question to your process for each, and put the date on the calendar.

If you want the next step after this, read my plain-English guide to building a logic model, then the one on writing a grant progress report funders actually read. And if you're wondering whether a chatbot can just do this for you, I answered that honestly in Can't We Just Write the Grant Report With ChatGPT?

The 1,200 people were real. Next time, tell your funder what changed for them.