What needs to be finished first
The stakeholders whose behavior or situation should change through the impact are named and differentiated.
What needs to be ready before start
Large whiteboard or Miro board with columns from right to left (Impact, Outcomes, Outputs, Activities, Inputs); sticky notes in four colors; markers for assumption stickers and risk symbols; timer.
One facilitator with ToC experience; four to twelve participants from program leadership, operations, domain, evaluation; one scribe for assumptions and measurement points.
Mission or program idea in one paragraph; existing impact assumptions; known data or studies about comparable programs; time horizon (typically 3-5 years for impact).
2-6 h
Set up five columns on the board and work from right to left. Keep an example row visible (Activity -> Output -> Outcome -> Impact). Rule: think backward, impact is the starting point, not the endpoint.
The one question this method answers
How do the planned activities lead through outputs and outcomes to long-term impact, and which assumptions and risks do we need to test and measure?
Marker: Phase
| Step | Duration | Action | Hint |
|---|---|---|---|
1Phase 1: Formulate impact | 20-30 min | Formulate long-term impact as an observable end state (typically 3-5 years). Be specific per stakeholder: which state is intended. Support it with measurable indicators. | Impact is not mission. Mission is direction, impact is observable state. "Improve education" is mission; "reduce school dropout share in region X from 12% to 7%" is impact. |
2Phase 2: Derive outcomes backward | 30-45 min | For each impact statement, ask which behavior or state changes must happen before it. Build outcomes as a chain (long-term -> mid-term -> short-term). | Outcomes are changes in stakeholders, not what we do. "Parents attend workshops" is output; "parents apply new parenting techniques" is outcome. |
3Phase 3: Outputs and activities | 30-45 min | For each outcome, note which outputs (countable results of activities) are required and which activities produce these outputs. Defer inputs (resources) until the end. | An activity does not automatically produce an outcome. Check assumptions for every Output -> Outcome connection, because this is where the break usually occurs. |
4Phase 4: Mark assumptions explicitly | 30-45 min | For each arrow in the diagram, mark assumptions with assumption stickers: what must be true for the impact to happen (causal assumption, context assumption, resource assumption)? | Assumptions are where programs usually fail. If no assumptions are marked, the group is documenting a wish, not a theory. |
5Phase 5: Measurement points and risks | 30-45 min | Define at least one measurement point per outcome and impact: indicator, data source, frequency. For each critical assumption, add a risk sticker with likelihood and damage value. Assign owners for measurement and risk monitoring. | Measurement points without data source are wishful thinking. If data must first be collected, that collection is itself an output, not something that just happens. |
What comes out at the end
Theory-of-Change diagram as image or digital board plus supporting document with impact definition, outcome chain, output-activity linkage, explicit assumption list, measurement points with data source and risk plan.
Own ToC per program with date. Annual reviews with update section (which assumptions were confirmed, refuted, modified). On strategy change, create a new version and keep the old one as snapshot.
Compact working template for Theory of Change with context, input, output artifacts, and next step.
What is this method used for?
Which question should be answered at the end?
Which data, observations, or materials are available?
Owner, date, success signal.
Concrete filled scenario, fictional example
Concrete filled scenario, fictional example
Impact (3 years): 60% of solo tax advisors among our customers reduce weekly admin time from 12 h to 8 h and gain at least 5% client growth.
Outcomes:
Outputs: 5-day onboarding flow live, knowledge base with 30 articles, community forum with active moderation.
Activities: Content production, hire community manager, build onboarding tool.
Assumptions (excerpt, critical):
Measurement points: Activation Rate (Mixpanel, monthly), weekly admin time (self-report survey, quarterly), NPS (in-app, every 90 days).
Recognize symptoms and steer against them
"10 workshops delivered" sits in the outcome column, behavior change is missing.
Repeat the outcome definition: change in stakeholders, not in us. If confused, explicitly contrast output and outcome.
Diagram looks stringent, with hardly any assumption stickers.
Require at least one assumption per arrow. If the group cannot think of any, the facilitator asks: what would have to be true for this arrow to work?
Impact is "a better world" or "more happiness", with no indicators.
Enforce Phase 1: impact needs at least one quantitative or clearly observable indicator. Otherwise the program cannot be evaluated later.
Indicators exist, but nobody knows who measures what, where, and how.
Define data source, owner, and frequency per measurement point. If data collection is required, add it as a separate output.
Activities are already decided, ToC is rationalized backward.
Question activities when outcome logic breaks. ToC may change the program; otherwise it is decoration.
Done signals checkable in under a minute
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.