Sopact is a technology based social enterprise committed to helping organizations measure impact by directly involving their stakeholders.
Copyright 2015-2026 © sopact. All rights reserved.
Theory of change vs logic model - what each shows, the seven points where they differ, when to use each, and why most programs need both.
A logic model describes what a program does, laying out inputs, activities, outputs, and outcomes in a compact left-to-right matrix; a theory of change adds the causal layer, naming the mechanism on each arrow and the assumption behind each outcome. A logic model answers “what”; a theory of change answers “why”. The logic model is the one-page summary; the theory of change is the argument it summarizes.
The two are not competing formats and you rarely pick one instead of the other. In Sopact they scaffold different layers of the same program: the logic model is the operational view a funder reads in under a minute, and the theory of change is the causal reasoning an evaluator tests. Confusing them for interchangeable versions of the same picture is where most impact frameworks quietly break.
Used by: nonprofit program leads, grant writers, M&E and MEL leads, and foundation officers who have been asked for “a logic model or a theory of change” as if they were the same deliverable, and need to know which to build, in what order, and why.
Here is the mistake that shows up in nearly every funder application. A team is asked for “a theory of change or a logic model,” treats the two words as synonyms, builds whichever is quicker, and ships it. Call it the Scaffold Confusion: treating a theory of change and a logic model as interchangeable when they scaffold different layers of the same program. Pick one and half the building is left unsupported.
A logic model scaffolds the operational layer — what the program commits, does, and produces, in a form a funder can scan. A theory of change scaffolds the causal layer — why those activities should produce those results, the mechanism on each arrow and the assumption each link depends on. Build only the logic model and you have a description with no explanation: you can say training leads to employment but not why the link should hold, so when it fails you cannot say which assumption broke. Build only the theory of change and you have an argument with no legible summary a funder can read at a glance. Both layers hold up the same program, and removing either leaves it standing on one leg.
The confusion has a fix, and it is an order of operations. Build the theory of change first — the causal reasoning, mechanisms, and assumptions — then derive the logic model from it as the one-page summary. That way the summary is true to the argument beneath it rather than a separate document that drifts. Sopact builds the two together from the start, so the logic model is always a view of the current theory, not a slide that stopped matching it months ago. For the layer beneath both, see the theory of change guide and the logic model guide.
Set side by side, the two frameworks differ on six things, and every difference traces back to the same split — the logic model describes, the theory of change explains.
Mechanism. A logic model has none; it shows that activity A is followed by outcome B without saying why. A theory of change names the mechanism on every arrow — the reason this activity should produce this change. Assumptions. A logic model carries no assumptions column; a theory of change makes the assumption behind each outcome explicit, which is the part a funder pressure-tests. Length. A logic model is one page by design, legible in under a minute; a theory of change is longer because it carries reasoning, not just boxes.
Audience. The logic model is for funder communication — a fast, shared picture of the program. The theory of change is for evaluation design — the structure an M&E system is built from. Revision cadence. A logic model is usually updated when the program plan changes; a theory of change is revised whenever the evidence tests an assumption, which is far more often if it is doing its job. Testability. This is the sharpest line: a logic model cannot be proven wrong because it makes no causal claim, while a theory of change is falsifiable — each assumption is a claim the data can confirm or break. That testability is exactly why Sopact treats the theory of change as the primary artifact and the logic model as its summary. For the causal-layer contrast in an M&E setting, see theory of change in monitoring and evaluation.
Watch — theory of change vs logic model. What each framework holds, why the theory of change comes first, and how the logic model falls out of it as the one-page summary. Presented by Unmesh Sheth.
The decision rule is simple: build the theory of change first, then derive the logic model from it. The causal reasoning defines what the program actually claims; the logic model is the compact restatement of those claims for a funder who wants the picture, not the argument. Doing it in the reverse order — drawing the boxes first, then reverse-engineering a rationale — produces a theory of change that merely narrates the logic model instead of justifying it, and an evaluator will see the seam.
Adjacent formats derive from the same source. A logframe is a logic model with an indicators, means-of-verification, and assumptions matrix — the assumptions column is the theory of change's per-arrow assumptions listed as external conditions. A results framework orders the results hierarchy for accountability reporting but, like the logic model, leaves out the causal argument, which is why USAID requires a theory of change beneath it. All of these are presentations of one underlying logic: build the theory of change once, and switch the view rather than maintaining four documents by hand. The full practice this sits inside is in impact measurement & management.
A theory of change and its logic model earn their keep at four moments — building the causal version first, deriving the one-page summary, connecting both to one participant record, and reporting to the funder while the theory keeps getting revised. The animation below runs the loop; the four prompts under it are the ones behind each job.
1 · Build the theory of change first. Name the mechanism on every arrow and the assumption behind every outcome before drawing a single box. The walkthrough is in how to build a theory of change.
Academy walkthrough → How to build a theory of change
Build a theory of change from this program description: [PROGRAM URL OR DOC]. For each activity-to-outcome arrow, name the mechanism (why this activity produces this change) and state the assumption the link depends on, written as a claim the data could confirm or break. Order the outcomes from short-term to long-term impact.
2 · Derive the logic model from it. Collapse the causal version into the five-component one-page summary a funder reads at a glance.
Academy walkthrough → How to build a logic model
From this theory of change: [PASTE OR LINK], derive the one-page logic model: inputs, activities, outputs, outcomes, and impact. Keep every outcome consistent with the theory it came from, and attach the indicator that measures each output and outcome. Do not add any outcome the theory does not support.
3 · Audit the theory of change. Find the arrows with no mechanism and the outcomes with no testable assumption before a funder does.
Academy walkthrough → How to audit a theory of change
Audit this theory of change: [PASTE OR LINK]. Flag every arrow with no named mechanism, every outcome with no assumption, and every assumption that cannot be tested with data. Rank the assumptions by how much the whole theory depends on each one, and return the corrected causal version.
4 · Audit the logic model against it. Make sure the one-page summary still matches the theory and that no output is masquerading as an outcome.
Academy walkthrough → How to audit a logic model
Audit this logic model against its theory of change: [PASTE OR LINK BOTH]. Flag every outcome that is really an output, every box with no indicator, and every place the logic model claims something the theory of change does not support. Return a corrected five-component model true to the theory.
The sections above are the argument; the Academy articles are the practice — each a hands-on companion written to run on your own data.
A logic model describes what a program does — inputs, activities, outputs, and outcomes in a one-page matrix — while a theory of change adds the causal layer, naming the mechanism on each arrow and the assumption behind each outcome. The logic model answers what; the theory of change answers why. In Sopact the theory of change is the primary artifact and the logic model is derived from it as the one-page summary, so the two never drift apart.
Usually yes. A nonprofit needs the theory of change for evaluation design — it is the causal argument the M&E system tests — and the logic model for funder communication, because it is the one-page picture a funder reads in under a minute. They are not duplicates; they scaffold different layers of the same program. In Sopact you build the theory of change once and derive the logic model from it, so having both costs one build rather than two.
Build the theory of change first, then derive the logic model from it. The theory of change defines what the program actually claims and why; the logic model is the compact restatement of those claims. Building the logic model first and reverse-engineering a rationale produces a theory that only narrates the boxes instead of justifying them, and an evaluator will see the seam. Sopact enforces this order by generating the logic model as a view of the theory.
The Scaffold Confusion is treating a theory of change and a logic model as interchangeable when they scaffold different layers of the same program. The logic model scaffolds the operational layer — what the program does; the theory of change scaffolds the causal layer — why it should work. Pick only one and half the building is left unsupported: you either have a description with no explanation or an argument with no legible summary. Sopact builds both layers together so neither is missing.
No. A logic model describes the program but makes no causal claim, so it cannot be proven right or wrong — it shows that training is followed by employment without saying why the link should hold. A theory of change makes that link explicit and testable. When an outcome fails, only the theory of change can tell you which assumption broke. A logic model can summarize a theory of change, but it cannot do the theory's job of explaining and being falsifiable.
A logic model is one page by design — a compact matrix legible in under a minute. A theory of change is longer because it carries reasoning: the mechanism on each arrow and the assumption behind each outcome, not just the boxes. The length difference is the point. The logic model trades detail for speed of reading; the theory of change trades brevity for the causal explanation an evaluator needs. Sopact keeps both, each in its natural form, from one record.
A logic model is usually updated when the program plan changes — new activity, revised output target. A theory of change should be revised whenever the evidence tests an assumption, which is far more often if the framework is doing its job. In Sopact, assumptions are embedded as monitoring questions on a persistent participant ID, so a breaking assumption surfaces in weeks and the theory is revised each cycle rather than once a year.
Yes. Sopact builds a theory of change and its logic model together from one record: the theory of change carries the mechanisms and assumptions, the logic model is derived as the one-page summary, and both connect to the same participant IDs so they read from live data instead of sitting in static slides. Because the logic model is a view of the theory rather than a separate document, the two stay consistent as the program runs and the evidence lands.
The theory of change is the backbone of the M&E system: every outcome stage becomes an indicator and every assumption becomes a monitoring question, while the logic model is the one-page summary a funder reads. Build the theory of change first, derive the logic model from it, and connect both to one persistent participant ID so baseline and follow-up track the same people. Sopact generates the M&E data collection from the theory, which the theory-of-change-in-monitoring-and-evaluation guide walks through in detail.