Where Logic Models Quietly Lie
Every logic model has an unstated assumption it's hoping no one asks about.
Back in 2003, I was working at Catholic Community Services of Western Washington when I went to a training run by United Way of King County. They were teaching the logic model because they had started requiring one with every grant application. I got it immediately. To me, it was a puzzle: once I understood what each column meant and what it was for, I could fill in the pieces I already knew, then work forward and backward across the columns until the whole thing locked together.
I looked around the room, and most people were baffled.
Table of Contents
Why the logic model is back, and why it matters more now
What I did not realize in that room was that I had walked into the middle of a movement. United Way of America had published Measuring Program Outcomes: A Practical Approach in 1996, and local United Ways across the country were turning the logic model into a condition of funding. The tool I was puzzling over was already on its way to becoming a national expectation.
Then, for a while, it lost its shine. Theory of change became the fashionable term, and the logic model started to feel like paperwork you completed because a funder told you to.
The ground has shifted again, and this is the part my students need to hear. The federal government got serious about evidence, and the Foundations for Evidence-Based Policymaking Act of 2018 pushed federal agencies to build evidence about whether their programs actually work. Logic models came roaring back into grant guidelines. Today, they are required across federal programs, from behavioral health to education to workforce development, and a growing number of foundations ask for one too. The tool I learned as a United Way requirement in 2003 is now something my students will meet in nearly every serious application they touch. That is why I teach it early, and why I care so much that it gets built honestly.
Because here is the catch that trips up even experienced writers: understanding the logic model and building an honest one are not the same thing. A model can be filled in correctly and still quietly lie. The lies are almost always in the same five places, and once you know where to look, you cannot unsee them in your own work.
The output wearing an outcome's name tag
Hello, my name is... Outcome. (It's actually an output. It just borrowed the marker.)
This is the most common one, and it hides in plain sight. A logic model has an outcomes column, and that column is full of outputs. "200 students tutored." "15 workshops delivered." Those are counts of what you did, not evidence that anyone is better off.
Why it happens: outputs are safe. They are countable, certain, and fully within your control, while outcomes are harder to measure and ask you to promise a change you cannot completely guarantee. Under deadline pressure, writers retreat to what they can count. The fix is a single question asked after every output: and then what? If you cannot finish that sentence with a change in someone's knowledge, behavior, or condition, you have an output sitting in the wrong column.
The outcome with no road to it
This is the opposite failure, and it is just as common. The model names a real outcome, often a big and admirable one, with nothing connecting the activity to it. A twelve-week financial literacy class on one side, "increased economic mobility in the region" on the other, and a blank space in between.
Why it happens: the grand outcome is what the funder's mission wants to hear, so writers reach for it to sound aligned, and ambition outruns the evidence. Reviewers call this the leap. The fix is to build the chain in stages: an immediate outcome, then an intermediate one, then the long-term goal, each link close enough to the last that a skeptical reader believes the move. If the only way from your activity to your outcome is a miracle, name a smaller outcome you can actually reach and let the big one be the vision, not the promise.
The number that came from nowhere
"85% of participants will report increased confidence." It looks rigorous. It is often invented.
Why it happens: a precise percentage signals competence, and writers worry that an honest "we will measure the change" sounds weak, so they manufacture precision. A target becomes a quiet lie when there is no baseline behind it and no method in front of it. Eighty-five percent compared to what, measured how, by whom? The fix is humble and specific: state your baseline when you have one, say plainly when you do not, and name the instrument you will use to measure. A defensible 60% you can support beats an impressive 90% you cannot.
Here is a test I run from the reviewer's chair: I do the math backward to see how many real people a percentage requires. Say a program serves 30 people and promises that at least 85% will meet an objective. Eighty-five percent of 30 is 25.5, so the target depends on half a person showing up changed. That fractional person is the tell. The writer reached for 85% because it sounded strong, not because it matched the people in the program. Set the target at 25 of those 30, and you land at about 83%, a number that corresponds to whole human beings. Better yet, write it as a count: 25 of 30 participants. A percentage tied to a number you can actually reach reads as planning. A percentage that dissolves into fractions of a person reads as a guess.
The activity dressed up as impact
Impact isn't a costume you put on an activity. It's what's still true after the activity ends.
This one is about the verb you choose. An activity is the concrete thing you do: provide a training, offer counseling, distribute meals. A result is the change it produces: people gain skills, stabilize, have enough to eat. The lie is describing an activity with a result word, so the doing sounds like the achieving.
"We will empower women" is the classic example. Empowerment is not an activity. It is a result, and a hard one to prove. The activity is to provide a training. When "empower" shows up in the activities column, the proposal has quietly claimed the outcome as something the organization simply does, skipping the part where it has to show how the training leads to empowerment and how that change will be measured.
Why it happens: result words sound more impressive and more mission-driven than plain ones, so "provide a workshop" becomes "transform lives" and "offer mentoring" becomes "build resilience." A reviewer reads those words sitting in the activities column and sees a writer who has named the destination as though it were the trip.
The fix is to match the verb to the column. Use doing words for activities: provide, deliver, offer, host, distribute. Save the result words, empower, improve, increase, reduce, for the outcomes column, where you will have to back them up with a measure. Name the work for what it is, and let the result earn its place.
The assumption nobody wrote down
Every logic model rests on assumptions, whether or not it has a box to write them in. The chain only works if certain things hold true: that participants will show up and stay, that the partner organization will deliver its piece, that the community insight statement is actually what the community would say. Those assumptions are load-bearing, and most proposals never name a single one.
That silence is the quiet lie, and it is not really about the template. Most logic models, including the W.K. Kellogg standard, do not include an assumptions section, so it is easy to skip the step entirely. But the assumptions are there whether the format asks for them or not, and leaving them unstated makes the model look more certain than the work actually is.
Why writers skip it: naming a risk can feel like admitting weakness in a competitive ask. The irony is that surfacing assumptions helps you. Reviewers trust the proposal that states what it is counting on, and ideally says what happens if one fails, more than the one that presents a frictionless chain with no acknowledged risk. Naming an assumption is not a confession. It is evidence you thought the project all the way through.
So make it a practice, not a template feature. Whether your logic model has a place for assumptions or not, write them down somewhere the funder will see them, in the narrative, in a short note beneath the model, or in a theory of change. The point is simply that they get said.
What an honest logic model does
An honest logic model is not the impressive one. It is the one that could survive a reviewer reading it slowly. It keeps the outcomes column for outcomes, shows the road between the work and the change, supports its numbers, lets activities be activities, and names what it is counting on.
This is what I teach first, because the logic model is no longer a nice-to-have. It is required across federal programs and a growing share of foundation applications, which means more reviewers than ever are reading these columns closely. A model that quietly lies is a model that quietly loses points.
It also helps to build the logic model before the narrative rather than after, so the proposal is written from the logic instead of reverse-engineered to fit it. And when your award comes in above or below what you requested, the logic model is the first thing to revisit, because a different funding level changes what you can honestly promise. (More on that in the post on scaling a proposal up or back.)
The quiet lies are easy to tell because they make the work sound surer than it is. The truth is almost always more modest, and on the reviewer's side of the table, modest and well-reasoned wins.
Frequently asked questions
What is the real difference between an output and an outcome?
An output is a count of what you produced or delivered: people served, sessions held, materials distributed. An outcome is the change that resulted: new knowledge, a shift in behavior, an improved condition. A quick test is to ask "and then what?" after each item. If the answer is a change in someone's situation, you have an outcome. If there is no answer, you have an output.
How many outcomes should a logic model have?
Fewer than you think. A handful of outcomes you can actually measure and defend is stronger than a long list you cannot support. Reviewers are not impressed by volume. They are reassured by a clear, believable chain from activity to result.
Should I build the logic model before or after writing the narrative?
Before, whenever you can. A logic model built first becomes the backbone the narrative hangs on, and the proposal reads as one coherent argument. A logic model built last tends to be reverse-engineered to match a story that already exists, and reviewers can usually tell.
Why are logic models showing up in more applications again?
The evidence-based policy movement brought them back. Federal law now pushes agencies to show that their programs work, logic models are required across many federal grant programs, and a growing number of foundations ask for one. A tool that faded as a buzzword is now a hard requirement in much of the field.
My logic model has no assumptions section. Where do I put them?
Most logic models do not have one, so add your assumptions in the narrative or as a short note beneath the model. List the conditions outside your control that the project depends on: participant turnout and retention, partner follow-through, the accuracy of your read on community need, the stability of funding or staffing. Noting how you would respond if one slips signals that you have thought the project through rather than hoped it through.
What if I do not have baseline data?
Say so, plainly, and describe how you will establish a baseline early in the project. An honest "we will measure at intake and compare at exit" is more credible than a confident percentage with no source. Reviewers reward defensible over impressive.
References
United Way of America, Measuring Program Outcomes: A Practical Approach (Alexandria, VA: United Way of America, 1996).
Foundations for Evidence-Based Policymaking Act of 2018, Public Law 115-435, signed January 14, 2019.
If you want to go deeper than weekly tips, the logic model is one of the first things we build together in the Certificate in Grant Writing.
Not ready for that yet? Join the Spark the Fire newsletter and get practitioner level grant writing like this every week.
Course Description
The Spark the Fire Certificate in Grant Writing is a comprehensive, step-by-step course that gives you the skills, strategies, and tools to confidently find grant opportunities, write compelling proposals, and build relationships with grantmakers. Whether you're new to the field or sharpening skills you already have, you'll get practical training rooted in real-world scenarios, not theory. You'll build a solid foundation in finding and researching grant opportunities, funder alignment, narrative development, budgets, evaluation plans, and compliance, plus personalized feedback at every step to help you grow.
Course Format
24 hours of instruction you can move through on your own timeline
Peer review, support, and discussions with others learning alongside you
Weekly group coaching and coworking sessions every Tuesday at 11 am PT, 12 pm MT, 1 pm CT, 2 pm ET
One-on-one office hours with Allison, available with 24-hour notice
Quick help from the instructor and your peers in our 24/7 chat
No deadlines to complete the course
You'll receive individualized feedback from your instructor on every key assignment, so you build real skill and real confidence.
Course Capstone Projects
To earn the certification, you'll complete the following projects, each one designed to set you up for long-term success:
A grant proposal for the project of your choice. You pick the organization and project, and everyone writes to the same boilerplate application form rather than one funder's specific questions, so you master the core questions that show up in nearly every proposal.
A logic model using the Spark the Fire framework
A project budget
A prospect list and grants calendar
Payment Information
Use the payment options below, including payment plan options in the checkout. You may also request a purchase order by email or pay by check.
Bundle and Save
Pair the Certificate with our other courses and save. Adding the entire Mastering Grant Research course costs just $100 more in either Certificate bundle below.
The Complete Grant Writer. For people going all in on grant writing as their own practice. Certificate in Grant Writing + Business of Freelance Grant Writing + Mastering Grant Research. $2,097. (Save $344)
Craft to Career. For anyone who wants to make a living writing grants, from first client to full practice. Certificate in Grant Writing + Business of Freelance Grant Writing. $1,997. (Save $197)
The Full Pipeline. For in-house grant professionals who own the whole cycle, from finding funders to writing the asks. Certificate in Grant Writing + Mastering Grant Research. $1,597. (Save $147)
More Info
Continuing Education Points? Yes.
Your 24 hours of instruction translate directly into 24 continuing education points. Full participation in the Certificate in Grant Writing is applicable for 24 points toward the GPC and qualifies under Category 1.B (Education) of the CFRE International application for initial certification and/or recertification. Every hour you invest counts twice, once toward your skills and once toward your credential.
