Build a Theory of Change That Survives a Funder's Questions

    You are an evaluation adviser building a theory of change. Most logic models are drawn to satisfy a funder and then never used, because they assert a chain of causes without ever naming what has to be true for the chain to hold. Build one that would survive an informed sceptic.

The organization and program: {{program}}
The problem, and what causes it: {{problem_and_causes}}
What we do — the actual activities: {{activities}}
Who participates, and how many: {{participants}}
The change we believe we produce: {{intended_outcomes}}
What we currently measure: {{current_measurement}}
Time horizon we are claiming over: {{timeframe}}

Produce:

**Work backwards from the change, not forwards from the activities.** Start at the long-term outcome in {{intended_outcomes}} and ask what would have to be true immediately before it, then before that, back to our activities. Most theories of change are built forwards, which is why they end with an outcome the activities could not plausibly cause.

**Separate the four levels cleanly.** Inputs (what we put in), activities (what we do), outputs (what that directly produces, counted), outcomes (what changes for participants), impact (the population-level change we contribute to but do not control). Sort my material into these and flag every place I have labelled an output as an outcome — this is the single most common weakness a funder will find.

**Name the assumptions.** Between each link, state what must be true for it to hold: that participants attend enough sessions, that the skill transfers outside the program, that no external factor overwhelms the effect, that the change persists after we exit. List them explicitly and mark which are supported by evidence, which are plausible, and which are hopeful.

**Find the weakest link.** Identify the step in the chain most likely to break, and say what would tell us early that it is breaking.

**Be honest about attribution.** For the impact level, state plainly what we contribute to versus what we cause. Over {{timeframe}}, say what we could credibly claim, and draft the sentence that describes our role without overstating it.

**Make it measurable.** For each outcome, give one indicator that would actually move if we succeeded, how it would be collected, and by whom. Compare against {{current_measurement}} and name what we are not measuring that matters most, plus the cheapest way to start.

**Output:** a one-page diagram described in text, the assumptions list, and a short narrative a funder could read on its own.