Nonprofit Logic Model Example With an Alignment Walkthrough

The worked nonprofit logic model

The five-column vocabulary comes from the W.K. Kellogg Foundation’s Logic Model Development Guide, still the sector’s reference document: resources and activities are “your planned work,” while outputs, outcomes, and impact are “your intended results,” with outputs defined as the direct products of program activities. Kellogg also puts rough clocks on the results side: short-term outcomes within 1 to 3 years, longer-term outcomes within 4 to 6, impact at the community or system level in 7 to 10. A 12-week program promising community-level change inside a grant year has answered a question nobody asked.

Logic model inputs examples: nouns with quantities

The inputs column fails by vagueness more than by omission. An input is something the program consumes, written so a stranger could price it.

Entries that pass: 0.5 FTE program coordinator at a stated salary share; 12 trained volunteer tutors averaging 2 hours a week; a $3,600 curriculum license; two donated classrooms with the donor named; a 24 percent documented fringe rate. Entries that fail: community support, strong partnerships, dedicated staff, technology. The failing entries are not false; they are unpriceable, and a reviewer who cannot price the inputs column cannot check whether the activities are affordable. Every passing entry also has a second home waiting for it: the same quantities become the personnel and cost lines of the grant budget, which is the fastest consistency check in the whole application.

Logic model activities examples: verbs with a dose

Activities are what the program does with the inputs, and each one needs a frequency and a duration because dose is what connects an activity to any believable outcome.

Entries that pass: run two 2-hour workshops weekly for 12 weeks; deliver 4 coaching sessions per participant; conduct monthly home visits for 10 months; train tutors in a 12-hour August institute. Entries that fail: provide support, offer mentorship, raise awareness, engage the community. An activity with no dose cannot be costed, cannot be scheduled, and cannot be tested against the outputs column, so it quietly exempts itself from every review the model exists to enable.

The alignment walkthrough

A logic model earns its keep when its numbers survive contact with the other proposal documents. Walk the Fairview model through four checks.

Inputs against the budget. Every input traces to a budget line or a named in-kind commitment: the $27,000 salary share, the $18,000 of coach hours, the $3,600 license. If the model lists a laptop lab and the budget has no equipment, replacement, or insurance line, one of the two documents is wrong.

Activities against the outputs. The multiplication above either works or it does not. First drafts routinely fail it, because the columns get filled on different days by different optimism levels.

Dose against the evidence. The 80 percent attendance output is not decoration: the curriculum publisher’s evidence assumes participants actually receive the workshops, so attendance is a stated delivery requirement. Whatever supports your outcome claim, published evidence, prior program data, or a declared professional rationale, it assumes a dose, and the model must deliver that dose on paper.

Outcomes against the evaluation plan. Each outcome names its instrument: the portfolio checklist, the mock-interview rubric, the 6-month follow-up. The evaluation plan should measure exactly these, with the same wording. The CDC’s program evaluation framework makes describing the program the second step of any credible evaluation precisely because an unaligned model leaves the evaluator measuring things the program never promised. One translation note from the CDC’s own materials: its format treats outputs as an indicator that can live in the narrative rather than as a mandatory column, so when a funder supplies a form whose labels differ from Kellogg’s, map your entries onto their structure instead of arguing with it.

Three misalignments and their fixes

The first draft of the Fairview model contained all three of the failures reviewers see most, planted here on purpose.

Output inflation. The draft said 90 workshops. The schedule supports 72. The fix is not averaging; it is recomputing the output from the activity schedule and letting the smaller, checkable number stand.

An outcome hiding in the outputs column. The draft listed “48 participants gain interview skills” as an output. The sorting rule: if the number can go up without any participant changing, it is an output; otherwise it is an outcome and needs an instrument and a timeframe. Skill gain moved to short-term outcomes and acquired the rubric.

An unresourced activity. Employer panels appeared in the activities column, but no coordinator hours or budget line paid for recruiting employers. The honest fixes are to resource the activity or cut it; the draft that leaves it unresourced is promising free work from a column that is supposed to be priced.

A clean, aligned logic model cannot guarantee an award, and no diagram can. What it does is remove the contradictions that make reviewers stop trusting a proposal, and it hands every downstream document a single set of numbers to inherit. The column mechanics and the four arrow tests behind this walkthrough are covered in the logic model template; the budget side of the same discipline, explaining each cost’s basis and purpose, is in the budget narrative template.