Logic Model for Grants: Linking Activities, Outcomes and Impact With Indicators

Linking activities, outcomes and impact in a grant is the job a logic model does on one page. The diagram often arrives as a line on a funder's checklist, close to the deadline, with no word on what it is for. Built early, a logic model for grants tells you whether the proposal holds together before you have written a paragraph of narrative. It shows which promises your budget can actually pay for. Built last, it usually exposes a gap you no longer have time to close.

Quick Answer: How Do You Link Activities, Outcomes and Impact in a Grant?

Build the chain in a grant logic model: each activity produces a counted output, each outcome names who changes, how and by when, each outcome carries an indicator with a data source and a collection date, and the impact claim follows from those outcomes at the program's real scale. A link that needs a paragraph of explanation is broken.

Free download: Logic model builder workbook as Excel .xlsx. No email address required.

What Is a Logic Model?

A logic model has five columns, and each column answers one question about the program. The columns run left to right, and the answers have to agree with each other.

An output is a count of what the program produced. An outcome is a change in the people or conditions the program touched. Confusing the two turns the outcomes column into a second count of the same work, which is the failure mode this format exists to prevent.

The five-column structure comes out of program evaluation practice; no single funder owns it. The W.K. Kellogg Foundation's Logic Model Development Guide defines outputs as "the direct results of program activities" and outcomes as "specific changes in attitudes, behaviors, knowledge, skills, status, or level of functioning expected to result from program activities" (W.K. Kellogg Foundation, Logic Model Development Guide, 2004). Those definitions matter for grant writing, because the model does most of its work during planning. Our guide to how to write a grant proposal puts that planning ahead of drafting, and the logic model is where the planning becomes visible on a page.

Logic Model for Grants: What Reviewers Cross-Check

A reviewer reads a logic model against the rest of the application and looks for places where the documents disagree. The diagram is short enough to hold in view while checking three or four other pages, which is exactly what makes it useful to a reviewer and risky for an applicant.

The budget is the first cross-check. If the model lists a full-time family navigator and the grant proposal budget funds a quarter-time coordinator, the reviewer trusts the budget and discounts the model.

Then the workplan. Every activity in the model should appear on the timeline with a start date and an owner. An activity that appears only in the diagram is aspiration.

The evaluation plan gets read against the outcomes column. Each outcome needs a measure, a data source, and a collection schedule. That is the whole job of the evaluation plan, and the outcomes column supplies its raw material.

Last, the statement of need against the impact column. If your need statement documents a shortage of bilingual caseworkers, the impact column should point at that shortage. General community wellbeing is a different claim, and the two do not substitute for each other.

When these four disagree, the model works against the application. A reviewer who spots one contradiction has a reason to go looking for others.

How to Build a Logic Model for Grants

Build a logic model backwards. Start with the change you are promising and work leftward to the resources you actually have, because that direction surfaces the gaps while there is still time to fix them.

1. Name the impact. Write the long-term condition the project contributes to, in plain language, in one sentence. Keep it honest about scale. A two-year, one-site program contributes to a community condition; it does not eliminate it.
2. Write the outcomes. For each outcome, answer four questions: what changes, for whom, by when, and by how much. Split them into short-term (knowledge, skills, attitude) and medium-term (behavior, practice, status). This column and the goals and objectives section of your application should say the same thing in different formats.
3. Set the outputs. For every outcome, name the countable evidence that the work happened. Outputs are attendance, completion, delivery, and distribution figures. They prove effort, not effect.
4. List the activities. Include only the activities that produce the outputs you just listed. An activity with no output attached is either missing a measure or does not belong in this proposal.
5. Price the inputs. Walk each activity back to the staff time, space, materials, and partner capacity it requires. This step catches an overreaching workplan, because the inputs column has to match the budget line by line.
6. Add assumptions and external factors. Underneath the columns, write what has to be true for the chain to hold, and what could break it. Assumptions might include that partner schools will release student data. External factors might include a funding cut at a referral agency. Naming them signals that you have thought past the best case.
7. Read it back as a sentence. Left to right, the model should read: if we have these resources, we can run these activities, which produce this much service, which leads to these changes, which contribute to this condition. If any link needs a paragraph of explanation to make sense, the chain has a hole.

Steps two through five usually need a working session with program staff. A solo drafting hour will not surface the same gaps, because the people who have to deliver the program are the ones who can tell you which links are thin. Expect the session to change at least one outcome, and expect at least one activity to fall out of the model entirely once someone asks who is running it on a Tuesday afternoon.

Logic Model Grant Writing: Turning the Columns Into Narrative

Each column of a finished logic model feeds a specific section of the application, so the model doubles as your drafting outline.

Write the outcome text once and reuse it. The objectives section and the evaluation plan then carry identical wording, which is harder to achieve than it sounds when three people draft three sections in the same week. The evaluation plan guide shows how that shared wording becomes measures and a collection schedule.

Funders differ on whether the model belongs in the narrative page count or in the appendix, and some ask for a narrative description with no diagram at all. Read the instructions before you design the page. Our logic model template covers the layout options if you want a blank grid to start from.

When the Funder's Template Does Not Use Five Columns

Funder templates rename the columns and sometimes split them, but the chain underneath stays the same. Map the headings on the form onto the five standard columns before you type anything into it, and the model you already built transfers without rework.

Nothing standardizes the labels, so read the column definitions in the instructions instead of trusting the column names.

Two variants cause most of the trouble. The first is a template with three outcome columns, which invites padding, and a medium-term outcome that restates the short-term one in different words is visible to anyone reading down the page. If the program has two real outcomes, leave the third box empty and say why in a line of text. The second is a template that opens with a problem statement box at the far left. That text is your statement of need in compressed form, so lift it from the narrative rather than writing it fresh, because a reviewer will read the two versions side by side.

Fill in the funder's form, then keep your own five-column version as the working copy. The form is an export. The model is what you edit when the program changes.

A Logic Model Example You Can Adapt

A grant logic model for an after-school tutoring request fits on one page, with one row for each chain that runs from an input through to an impact. The version here belongs to Example Literacy Partners (composite), a small literacy nonprofit serving one elementary school with a $25,000 request (illustrative figures).

Illustrative figures, not data from a real organization.

Every output in this literacy-tutoring model is a number the coordinator can produce from a sign-in sheet, and every outcome has a named measure attached. Swap the population and the assessment and the same structure holds for a health clinic or a workforce program.

Notice what the impact column does not say. It does not claim the program will close the district's reading gap. A tutoring project reaching 30 students contributes to a district goal. Claiming ownership of that goal invites the reviewer to check the arithmetic, and the arithmetic will not support it.

What the First Draft of This Example Got Wrong

Building backwards catches problems that a finished table hides. Three things changed between the first pass at this composite example and the finished version.

An outcome had no instrument. The first version promised that students would gain confidence in reading. Nobody could say where that number would come from at report time, and confidence appears on none of the assessments the program already gives. It became a decoding-skills result measured by the fall and spring assessments, which the activities column was paying for anyway.

An activity lost its place. The first list included monthly check-ins with classroom teachers. The idea sounded useful and produced nothing countable. Asked who would schedule the meetings and what record would come out of them, the coordinator answered nobody and none, so the activity came out of the model.

The inputs column forced a cut. The first draft assumed a full-time coordinator. At a $25,000 request (illustrative figures), with volunteer tutors and donated space absorbing most of the delivery cost, that line pays for part-time hours and nothing more. The program dropped from three tutoring afternoons a week to two, and the output count moved with it. Outputs have to survive arithmetic: two sessions a week across a 24-week term is 48, which is the figure in the table. A session count carried over from the larger schedule would have described a program the budget does not fund.

Two of those inputs also belong on the assumptions line, because the organization controls neither one. The donated classroom depends on a principal who can reassign the room, and twelve volunteers depend on a recruitment season that has come up short before. Write both underneath the columns with the fallback in a sentence, so a reviewer can see that the plan survives losing the room.

Linking Activities, Outcomes and Impact in a Grant: The Chain With Indicators

Linking activities, outcomes and impact means every link in the chain points to the next one with evidence attached: an activity to the output it produces, an output to the outcome it should lead to, an outcome to the indicator that will show it, and the outcomes to the impact they contribute to. A model with one link missing reads as a list of good intentions.

Grant writers name this gap themselves. The complaint that "the activities, outcomes, and indicators don't connect well" appears in two Reddit threads about proposal writing (Reddit threads 1vf8jxw and 1sv1f08, read September 2026), and one commenter diagnosed rejections as "no crystal clear outcomes that are linked confidently to the org's capacity and track record" (Reddit thread 1vf8jxw, read September 2026).

Grant Logic Model Outcomes and Indicators, Worked Through

An indicator is a measure that shows whether an outcome happened, written as a count or a percentage of a named group at a named date. The Urban Institute's outcome-management guide gives the wording rule: "Outcome indicators should almost always begin with words such as 'The number of . . .' or 'The percent of . . .'" (Urban Institute, Key Steps in Outcome Management, 2003). This chain follows one program from activity to impact with an indicator under each outcome.

Illustrative figures, not data from a real organization.

The right-hand column is the test. The employment indicator needs follow-up calls after the course ends, so the calls belong in the activities column and their staff time belongs in the budget. Without them, the medium-term outcome has no evidence at report time, and the chain breaks at its most important link.

A Self-Check for Broken Links in the Chain

Run six checks on any draft before the narrative is written around it.

Then read the model right to left. Start at the impact and ask which outcome supports it, which output feeds that outcome, and which activity produces that output. A claim with nothing to its left either comes out or gets the activity it is missing.

Common Mistakes in a Logic Model for Grants

A frequent error is an outputs column dressed up as outcomes. Counting what you delivered does not show what changed, and a reviewer reading the two columns side by side ends up seeing the same information twice.

Illustrative figures, not data from a real organization.

Four more errors show up repeatedly.

The activities column turns into a to-do list. Grant writing, hiring, and reporting are project management tasks, not program activities. The activities column holds what you do for participants.

The impact claim outruns the award. Scale the impact statement to the size and length of the grant. A one-year, single-site project that claims a county-wide condition change is promising more than the money buys.

The model runs to too many rows. It is a one-page format, and a dozen outcome chains crowded onto a page stop being readable at print size. Fewer chains, each one traced all the way through the five columns, does more for the application than a full page of half-finished ones.

You build the model after the narrative is finished. Reverse-engineering a diagram from finished text reproduces whatever gaps the narrative already has. Build it during planning and edit both together.

Using the Model After You Submit

A logic model keeps working after submission because the outputs and outcomes columns are the reporting schedule for the award. Interim reports ask for the numbers already listed in the outputs column and the measures already named beside each outcome.

Teams that handle this well set up the collection dates at award time, months before the first report is due. GrantCue keeps tasks and deadlines on every plan, Free included, with an ICS calendar feed and simple email reminders, so each indicator's collection date can sit on the calendar before the report is due; Essentials, at $49 a month for up to 3 users, adds daily or weekly email digests for deadlines and tasks (grantcue.com pricing, October 2026). The features overview covers task management and deadline tracking.

The harder case is a target you are going to miss. Report the variance rather than quietly restating the target as though it had always been lower. Put the original number, the actual number, and the reason in one paragraph, then say what changes for the next period. Edit the model itself when an activity changes and date the edit, so the final report starts from a document that matches what the program did.

The model also gives you a head start on the renewal. A funder deciding whether to continue support wants the same chain, with the outcomes column showing what actually happened.

FAQ

What Is a Logic Model?

A logic model is a one-page diagram that connects the resources a program uses to the results it promises. It has five columns: inputs, activities, outputs, outcomes, and impact. Some grant applications ask for it as a one-page attachment, and program evaluators use the same format outside of fundraising.

How Do You Write Outcomes and Indicators in a Grant Logic Model?

Write each outcome as a change in a named group by a named date, then attach an indicator that shows it: a count or a percentage, the instrument that produces it, and the collection date. Put the collection step in the activities column and its staff time in the budget, or the indicator will have no data when the report is due.

What Is the Difference Between a Logic Model and a Theory of Change?

A logic model shows what a program does and what results from it. A theory of change explains why those steps should produce that result, including the assumptions and conditions behind each link. A theory of change is usually longer and written in prose. An application that asks for a logic model often covers the reasoning in the project description.

What Is the Difference Between Outputs and Outcomes?

Outputs count what the program delivered, such as sessions held or people served. Outcomes describe what changed for those people, such as skills gained or behavior adopted. A number that could be pulled from a sign-in sheet is an output. A finding that requires a before-and-after measure is an outcome.

Do Grant Applications Require a Logic Model?

Some funders ask for one by name in the application instructions, often as a one-page appendix. Others fold the same information into the project description or evaluation section without using the term. Read the notice of funding opportunity before deciding, and build the model regardless, because it tests the proposal whether or not you attach it.

Where Does the Logic Model Go in a Grant Proposal?

Placement varies by funder. Some put it in the attachments or appendix outside the narrative page limit; others count it against the limit. When the instructions are silent, attach it as a labeled one-page appendix and reference it in the project description so the reviewer knows to look.

How Long Should a Logic Model Be?

One page. The format exists to make a program legible at a glance, and a model that runs to three pages has stopped doing that. If the chains do not fit, the project has too many outcome claims for the size of the request.

Build the Chain Before You Pick a Funder

The same five columns work for a federal notice and for a family foundation, so build the model before you have picked a funder. Then search open grant opportunities for funders whose priorities match the outcomes you can defend. When the evaluation section needs a method to match those outcomes, the guide to grant evaluation methods compares the options by cost and rigor.