Nonprofit Logic Models: What Funders Expect and How to Tie Yours to the Evaluation Plan
GrantCopilot Team
November 1, 2026
11 min read
TL;DR
A nonprofit logic model lays out inputs, activities, outputs, outcomes, and impact, usually in the format of the W.K. Kellogg Foundation's Logic Model Development Guide. Foundations want one readable page that matches your narrative and budget. Government funders vary: some require one and score it, and some, CDC among them, publish their own logic model and ask you to show how your work fits it. Build yours from the problem and your SMART objectives, keep outputs (what you count) separate from outcomes (what changes), then write the evaluation plan straight from it: an indicator, data source, timing, and target for each outcome you measure.
A logic model is the one page a grant reviewer can read in two minutes and come away knowing what your program does, for whom, and what changes because of it. Some foundations ask for one by name. Many more ask for its pieces under other headings: goals, objectives, outcomes, and how you plan to measure them. Some government funders require one and score it; others publish their own and ask you to show how your program fits it. A logic model also does a second job that applicants underuse: it is the skeleton of your evaluation plan. Every output becomes something you count, and every outcome becomes something you measure. When the model and the evaluation plan come from the same source, reviewers see one consistent program instead of a narrative and an evaluation that describe two different things. This guide covers what nonprofit funders expect, how to build the model in a reviewer-friendly order, and how to turn it into an evaluation plan, with a worked example. Writing a federal research proposal instead? See our companion guides to NSF logic models and NIH logic models.
What a Nonprofit Logic Model Shows
A widely used nonprofit format comes from the W.K. Kellogg Foundation's Logic Model Development Guide, last updated in 2004. Its components are resources (often called inputs), activities, outputs, outcomes, and impact, and it reads left to right as a chain of if-then statements: if you have these resources, you can run these activities; if you run them, you produce these outputs; if you produce them, participants change in these ways.
Many templates, including the one in this guide, split outcomes into three steps. A common way to tell them apart is knowledge first, then behavior, then condition. Above the table, state the problem or need the program addresses and your SMART objectives. The problem tells the reader why the model exists. The objectives tell them which outcomes you are committing to measure.
- Inputs: what the program draws on. Staff, volunteers, funding, facilities, partners, equipment.
- Activities: what the program does with those resources. Workshops, services, outreach, training. Keep them to a handful of high-level statements, not a task list.
- Outputs: the direct, countable products of the activities. People served, sessions held, materials distributed.
- Short-term outcomes: changes in knowledge, skills, or attitudes.
- Mid-term outcomes: changes in behavior or practice.
- Long-term outcomes: changes in condition or status.
- Impact: the lasting change in the community or system that the program contributes to.
What Foundations Expect
Foundation requirements vary widely. Some provide their own logic model template and want it as an attachment. Some ask for "outcomes and evaluation" in a narrative section and never use the words logic model. A few ask for a theory of change instead. Before you start, read the guidelines for three things: whether a logic model is required, recommended, or not mentioned; whether there is a required template or page limit; and where it goes (an attachment, the program design section, or the evaluation section).
Whatever the format, foundation reviewers tend to look for the same qualities:
- One readable page. Reviewers should be able to follow it without a legend or a second read.
- The same program as the narrative and budget. A staff position in the budget should appear as an input, and a workshop series in the narrative should appear as an activity.
- Outcomes the program can plausibly influence in the grant period, not the whole community problem solved.
- A clear line between outputs and outcomes, covered in its own section below.
- Scale that fits the request. A $25,000 grant does not end food insecurity in a county, and the model should not suggest it does.
What Government Funders Expect
Public funders are more prescriptive than foundations, and they use logic models in different ways, so read the notice of funding opportunity closely.
Some ask you to build one and score it. A 2026 Administration for Children and Families announcement (HHS-2026-ACF-OCS-EAH-0027) required applicants to submit a logic model for designing, managing, and evaluating the project, placed it in the project narrative, and gave it its own line in the scoring criteria, worth up to 5 points.
Some give you theirs. Some CDC announcements include a program logic model that lists the strategies CDC expects, and at least one 2026 CDC competition scored applicants' strategies on whether they were consistent with it. In that case your job is not to draw a new model but to show, strategy by strategy, how your activities feed the agency's outcomes, in the agency's own words.
Some never use the term. Many federal announcements ask instead for a work plan, goals and objectives, and performance measures. A logic model is still the quickest way to build those pieces so they agree with each other, even if you never submit it.
Behind all of this is the federal grants rule at 2 CFR 200.301, which requires agencies to measure recipients' performance against program goals and objectives, with expected outcomes, indicators, targets, and baseline data. That vocabulary maps directly onto the right half of a logic model and the evaluation plan built from it.
Before you start, check the notice of funding opportunity for these four things:
- Whether it requires a logic model, provides one, or does not mention it.
- Where it goes (the project narrative, an attachment, or a form) and whether it counts toward a page limit.
- Whether the scoring criteria give it points, and what they reward.
- Which performance measures the agency already requires, so your outcomes and indicators use them.
Build the Model in This Order
Most weak logic models are built left to right: list the staff, list what they do, then guess at outcomes. Build it the other way, starting from the change you want, so every activity has a reason to be there.
- Start with the problem or need. One or two sentences: what is wrong, for whom, and where. Use local data if you have it.
- Write your SMART objectives. Up to five. Each names who changes, what changes, by how much, and by when, for example: "By June 30, 2027, 70 percent of enrolled families will cook at least three dinners a week at home, measured by a 90-day follow-up survey."
- Name the long-term outcomes and impact. The change in condition you are working toward. This is your motivation, not a promise to measure.
- Work backward to activities. For each outcome, ask what the program does that causes it. If an activity leads to no outcome, cut it or explain it.
- List the outputs each activity produces. Every activity should produce at least one countable output.
- Fill in the inputs last. List the resources the activities actually need. They should match your budget and your partners' commitments.
- Read it left to right as if-then statements. If a link sounds like a leap to you, it sounds like one to a reviewer too.
Outputs Are Not Outcomes
Mixing these up is one of the most common logic model mistakes. An output is evidence that the program happened. An outcome is evidence that it worked. A quick test: if you can report it from your attendance sheet, it is an output.
Here is one program traced from output to long-term outcome:
- Output: 120 parents attend a six-session financial coaching course.
- Short-term outcome (skill): parents who complete the course can build a household budget.
- Mid-term outcome (behavior): participating households pay bills on time in the six months after the course.
- Long-term outcome (condition): participating households keep stable housing.
How the Logic Model Becomes Your Evaluation Plan
Write the evaluation plan from the model, not alongside it. The model already lists what you count and what you expect to change. The evaluation plan adds how you know, when you check, and what result counts as success.
The two halves of the model map onto the two kinds of evaluation most funders ask about:
Process evaluation reads the left side. Did the inputs arrive, did the activities run as planned, and did they produce the outputs you projected? Data sources are attendance logs, service records, and staff reports.
Outcome evaluation reads the right side. Did participants change in the ways your short-term and mid-term outcomes predict? Data sources are pre and post surveys, skills assessments, follow-up interviews, and administrative records.
For each outcome you plan to measure, the evaluation plan names five things:
- Indicator: the observable sign that the outcome happened.
- Data source and tool: the survey, assessment, or record that captures it.
- Timing: when you collect it, such as at intake, at program exit, and 90 days after.
- Target: the number or percent you expect, taken straight from your SMART objective.
- Who collects it: a staff role, a partner, or an external evaluator.
Measure What the Grant Period Allows
Measure short-term outcomes in full. Measure mid-term outcomes where the grant period gives them time to show up. Long-term outcomes and impact usually arrive after the award ends: say how you plan to track them if you can (for example, through a partner's administrative data), and do not promise to measure them if you cannot.
Reviewers trust a narrow plan the program can deliver more than a broad one it cannot. An evaluation plan that commits to measuring "reduced poverty in the county" after a one-year grant tells the reviewer the applicant has not thought about data collection.
Worked Example: A Family Cooking Program
A neighborhood nonprofit requests $40,000 for a one-year cooking and produce program for families at two elementary schools. Here is the logic model.
- Problem: families in the two school attendance zones buy most groceries at corner stores and report cooking dinner at home fewer than three nights a week.
- SMART objective: by June 30, 2027, 70 percent of enrolled families will cook at least three dinners a week at home, measured by a 90-day follow-up survey.
- Inputs: a half-time program coordinator, 12 volunteer cooking instructors, a partner grocer's produce donations, a school kitchen two evenings a week, and the requested grant.
- Activities: run a six-session family cooking course each semester; distribute weekly produce boxes to enrolled families; train volunteer instructors.
- Outputs: 80 families enrolled, 24 sessions held, 1,600 produce boxes distributed, 12 instructors trained.
- Short-term outcome: families can prepare five low-cost dinners from the course recipes.
- Mid-term outcome: families cook dinner at home at least three nights a week.
- Long-term outcome: household diets include more fruits and vegetables.
- Impact: lower rates of diet-related illness in the neighborhood.
Worked Example: The Evaluation Plan It Produces
Each row of the evaluation plan traces back to one box in the model.
- Outputs (process): families enrolled, sessions held, and boxes distributed, from the enrollment list and session sign-in sheets, reported monthly against the projected 80, 24, and 1,600.
- Short-term outcome: share of families who prepare five course recipes, from an instructor-scored cooking checklist at the last session. Target: 80 percent.
- Mid-term outcome: share of families cooking dinner at home three or more nights a week, from a survey at intake and 90 days after the course. Target: 70 percent, the SMART objective word for word.
- Long-term outcome: daily fruit and vegetable servings, from the same follow-up survey. Reported as a trend without a target, because one year is too short to see a stable change.
- Impact: not measured by the program. The narrative describes it as the reason for the work.
Where the Logic Model Goes in a Nonprofit Proposal
Follow the funder's instructions first. When they say nothing, the logic model fits best as a one-page attachment or figure that the program design and evaluation sections both refer to ("see the logic model in Attachment C"), so reviewers know it exists and read it alongside the narrative. Keep the wording of activities and outcomes identical in all three places.
Check it against the budget before you submit. Every input that costs money should appear in the budget, and every budget line should support an input or activity; our guide on how to build a grant budget walks through that crosswalk. Inputs that a partner provides should be backed by a letter of commitment, not a letter of support.
Common Mistakes Reviewers Notice
Most of these are consistency problems, not format problems:
- Outputs listed in the outcome columns.
- Activities written as a task list with twenty items.
- An evaluation plan that promises to measure long-term outcomes or impact within a one-year grant.
- A model that describes a different program from the narrative, budget, or timeline.
- Inputs the budget does not pay for and no partner has committed.
- One generic model reused across funders, with objectives that do not reflect each funder's priorities.
- Outcomes with no indicator, data source, or target anywhere in the evaluation plan.
How GrantCopilot Helps with Nonprofit Logic Models
GrantCopilot's logic model builder has a nonprofit format with the Kellogg columns used in this guide: inputs, activities, outputs, short, mid, and long-term outcomes, and impact, with the problem or need and up to five SMART objectives above the table. Here is what it does at each step, from a blank page to the exported proposal.
- Build it from your problem and objectives. You enter the problem or need and up to five SMART objectives. If objectives are the hard part, GrantCopilot suggests SMART objectives from your problem statement, with placeholders like [X%] where your text gives no number, and you keep the ones that fit.
- Draft the columns from your own text. GrantCopilot drafts inputs, activities, outputs, outcomes, and impact from your problem, your objectives, and the text of a linked proposal. Your problem and objectives stay as you wrote them. Where your text says nothing about a column, entries are described by type, such as "Program staff" or "Community partners"; GrantCopilot does not name organizations, partners, or people, or add numbers you did not give. You can edit every entry.
- Link it to your proposal. In a nonprofit proposal, the Logic Model row sits first in the section list. Choose one of your saved logic models there or create a new one. One model can serve several proposals; if you redraft a model that other proposals use, you choose whether to update it for all of them or make a copy.
- Your drafts follow it. When you write or revise a section with AI, GrantCopilot describes activities, outputs, and outcomes the way your model words them and does not promise ones the model leaves out. Section and full-proposal analysis flag activities or outcomes that do not match the model, and model items a section should cover but misses. That keeps your evaluation plan measuring what the model says the program does.
- It exports next to the evaluation plan. In PDF and Word exports, the model prints as a figure titled "Program Logic Model" on its own landscape page, right after the first section with Evaluation in its title, or at the end if the proposal has none.
Which GrantCopilot Plans Include Nonprofit Logic Models
Three GrantCopilot plans include nonprofit logic models. Building a model is a planning tool; linking it to a proposal and exporting it with one needs a plan that writes nonprofit proposals. Every plan starts with a 7-day free trial, and the pricing page has annual prices and full plan details.
- Nonprofit, $45/month: the nonprofit logic model format with drafting and SMART objective suggestions, plus nonprofit and foundation proposal writing, so you can link the model to your proposals and export it with them.
- Professional, $125/month: everything in the Nonprofit plan, plus every other logic model format and proposal type, for grant writers who work for many clients.
- Planning, $25/month: nonprofit logic models, including drafting and SMART objective suggestions, but no proposal writing. It suits planning a program before you write; to link the model to a proposal and export it, you need the Nonprofit or Professional plan.
Frequently Asked Questions
No. Some require one and supply a template, some recommend one, and many never mention it but still ask for goals, outcomes, and an evaluation plan. Read each funder's guidelines. When a logic model is optional, a one-page version still gives reviewers a quick view of how your program works.
A theory of change explains why you expect your activities to cause the change you want, including the assumptions behind each step. A logic model lays out the chain from inputs through activities and outputs to outcomes, usually on one page. Many funders treat the logic model as the visual summary of the theory of change.
One page, unless the funder says otherwise. If it runs longer, the activities are usually too detailed or the outcomes repeat each other. Keep activities to a handful of high-level statements and move task-level detail into the work plan or timeline.
Measure every short-term outcome and the mid-term outcomes the grant period allows. Long-term outcomes and impact usually happen after the award ends, so describe them as the reason for the work and say how you would track them if you can, without promising to measure them.
Above the table, next to the problem statement. Each SMART objective names a specific outcome from the model with a target and a deadline, and that target becomes the success measure for the matching row of the evaluation plan.
A nonprofit logic model earns its page when it does two jobs at once: it shows the reviewer how the program works, and it hands the evaluator a list of what to count and what to measure. Start with the problem and your SMART objectives, work backward to activities, keep outputs and outcomes apart, and write the evaluation plan row by row from the model. For the rest of the proposal, see our nonprofit grant writing tips and our guide to local grants from community foundations and corporate funders. For the vocabulary, see the logic model glossary entry. And if you also write federal research proposals, the NSF logic model guide and NIH logic model guide cover how those funders place and use the model.
Topics
logic model
logic model template
program evaluation
nonprofit logic model
evaluation plan
SMART objectives
nonprofit grants
outcomes measurement