The Hidden Discipline Behind Every Useful Theory of Change

Jul 18, 2026

10 min read

2 views

74%

The real question is not whether a plan is right

Most plans fail for a boring reason that gets disguised as a strategic one: they confuse a story about causality with causality itself. A logic model, a results chain, a theory of change, a pathway to impact, all of them are elegant until the world gets involved. Then the uncomfortable question appears: what exactly has to be true for this plan to work?

That question matters because no intervention succeeds by force of intent alone. Every serious attempt to create change depends on a long chain of conditions, behaviors, institutions, incentives, and timing. Somewhere in that chain are assumptions so ordinary that they remain invisible, until one of them breaks. The deeper discipline is not building a prettier diagram. It is learning how to make those hidden assumptions visible, testable, and discussable.

A theory of change is not a promise of what will happen. It is a map of the conditions under which change might happen.

That shift sounds subtle. It is not. It changes the purpose of planning from prediction to examination, from certainty to disciplined doubt.

The illusion of the clean causal chain

People often want causal plans to look like a staircase: if we do A, then B, then C, and eventually impact. The appeal is obvious. A clean chain makes action feel controlled, legible, and fundable. But real change rarely behaves like a staircase. It behaves more like a network of doors, each one opening only if several unseen hinges are aligned.

Consider a youth employment program. A polished diagram might say: train participants, improve skills, connect them to employers, increase job placement, reduce unemployment. Yet this chain only works if many other things are also true. Employers must trust the credential. Participants must be able to attend training regularly. Transport must be affordable. The labor market must have openings. Hiring managers must not discriminate. None of those are side notes. They are the actual load bearing beams.

This is why assumptions matter so much. They are not merely caveats tucked in the margins. They are the silent architecture of causality. If they are false, the chain does not merely weaken, it may never have existed in the first place.

The common mistake is to treat assumptions as annoyances, when they are actually the most strategic part of the plan. They tell you where the intervention is vulnerable, where learning is needed, and where confidence is misplaced. In that sense, assumptions are not a weakness of the theory of change. They are the part that makes it intelligent.

Assumptions are hypotheses wearing a disguise

The best way to think about an assumption is not as a belief, but as a testable hypothesis. That single reframing changes everything. A hypothesis can be examined. It can be stressed, refined, falsified, or strengthened. A vague assumption, by contrast, often survives because no one has named it clearly enough to challenge it.

Suppose a nonprofit believes that sending reminders will increase clinic attendance. That looks straightforward. But underneath sits a bundle of assumptions: people receive the reminder, they trust the sender, they have the means to travel, and attendance is constrained by forgetfulness rather than cost, fear, or stigma. If the real barrier is transport, then more reminders simply create more guilt.

This is where a good theory of change becomes less like a blueprint and more like a scientific instrument. It does not just say what should happen. It surfaces the propositions that make the path plausible. In practice, this means asking not only, “What are we doing?” but also:

What must be true for this step to work?
Which of those truths are most uncertain?
Which would most damage the strategy if false?
How would we know early enough to adapt?

That last question is crucial. A lot of organizations discover broken assumptions after the project is over, when the data can no longer save them. The value of naming assumptions lies in creating an early warning system. It turns strategy into a learning process rather than a verdict.

The point of identifying assumptions is not to eliminate uncertainty. It is to know where uncertainty matters most.

Contribution is not the same as proof, and that is liberating

There is a second tension hiding inside causal planning: the difference between claiming credit and understanding contribution. In complex social change, almost no intervention acts alone. Many forces are operating at once, often in contradictory ways. That makes the demand for absolute proof feel both tempting and unrealistic.

Contribution analysis offers a more honest posture. Instead of pretending we can isolate a single cause with laboratory purity, it asks whether the observed change is consistent with the intervention’s expected role in a broader causal story. In other words, did this work contribute plausibly to the outcome, and is there enough evidence to rule out more compelling alternative explanations?

That is a more modest standard, but not a weaker one. It respects reality. If a city reduces homelessness after expanding rental assistance, mental health outreach, and landlord incentives all at once, asking which one alone “caused” the decline is often the wrong question. The better question is: what part of the change can reasonably be attributed to each element, under what conditions, and through what mechanisms?

This matters because organizations often overpromise certainty in order to look accountable. Yet the harder they try to claim direct causation, the more they risk fooling themselves. Contribution thinking replaces heroic narrative with causal humility. It accepts that impact is usually the result of overlapping causal pathways, not a single decisive lever.

The real insight is that contribution does not weaken accountability. It improves it. When a program can explain its pathway, identify its assumptions, and examine whether the expected sequence actually unfolded, it becomes more accountable than one that merely points to a final outcome and says, “We were part of that.”

A practical framework: from story to stress test

A useful theory of change should behave like a bridge design. A bridge is not judged by how beautiful the sketch looks. It is judged by whether it can bear weight under plausible stress. The same should be true for causal plans.

Here is a simple framework for making that happen.

1. State the mechanism, not just the activity

Do not stop at “train staff” or “launch campaign.” Name the mechanism. For example: “training staff will improve counseling quality, which will increase participant trust, which will raise follow through.”

Mechanisms matter because they reveal what the intervention is actually trying to change in the world, not just what it is doing internally.

2. List the assumptions at each link

For every step, ask what must already be true. If the mechanism is “reminders increase attendance,” the assumptions might include:

people can reliably receive the reminder
the message is understandable and credible
attendance is mostly a problem of forgetting
recipients can act on the reminder without major barriers

The goal is not exhaustive completeness. The goal is to expose the most important points of fragility.

3. Rank assumptions by risk

Not all assumptions deserve equal attention. Some are low stakes. Others are existential. A strong planning process identifies the assumptions that are both uncertain and consequential. Those are the ones worth testing first.

A simple lens helps: imagine each assumption on two axes, likelihood of being false and damage if false. High uncertainty and high damage equals immediate attention.

4. Design small tests before large commitments

Instead of waiting for the final outcome, build pilot checks that can validate the pathway early. If you think clinic reminders work because people forget appointments, test whether attendance rises when reminders are sent. If attendance does not rise, ask why. Maybe the real issue is transit, not memory.

This approach saves organizations from scaling elegant theories that do not survive contact with reality.

5. Track contribution, not just outputs

Outputs tell you activity happened. Contribution asks whether the activity moved the causal chain in the expected direction. Did trust increase? Did referrals improve? Did employer uptake shift? These intermediate signals are the pulses of the theory, and they often reveal failure long before the final metric does.

The most important assumption is often political, not technical

Many teams think assumption work is mostly about logistics or behavior. In practice, the most important assumptions are often institutional and political. A program may have a technically sound pathway but still fail because it assumes cooperation where there is competition, or neutrality where there is resistance.

Imagine a data sharing initiative between agencies. The technical plan is easy to describe. But the real assumptions are harder: agencies are willing to collaborate, data definitions can be aligned, privacy concerns can be resolved, and no one fears being exposed by the numbers. If those conditions do not hold, the project is not just underperforming. It is being asked to succeed in a system that has not consented to the change.

This is why assumption work is also a power exercise. It forces people to name what they are implicitly relying on from others. It can reveal hidden dependencies, mismatched incentives, and conflicts that polite planning often leaves untouched. In that sense, a theory of change is not merely a management tool. It is a negotiation tool.

The most honest causal plans are the ones that admit where they are relying on trust, cooperation, legitimacy, or institutional bandwidth. If you cannot name those requirements, you do not really understand the pathway. You only understand your own side of it.

Why this changes how we evaluate success

Once you see assumptions as hypotheses and contribution as plausible inference, evaluation stops being a ritual of judgment and becomes a process of learning. That is a profound shift.

Instead of asking only, “Did it work?” you begin asking:

Did the expected mechanism activate?
Which assumptions held, which did not?
Where did the pathway bend, stall, or accelerate?
What alternative explanations remain credible?

This makes evaluation far more useful to decision makers. A program can miss its headline target and still generate valuable learning if it clarified which part of the causal chain failed. Maybe the intervention design was sound, but the context changed. Maybe the mechanism worked, but at the wrong scale. Maybe the target population was not the right one.

Equally, a program can hit its target and still be poorly understood. Without contribution analysis, success can be falsely attributed to the intervention when in reality the environment did most of the work. That is dangerous because it invites overexpansion of a fragile model.

In other words, the point of evaluation is not to crown winners. It is to separate useful causal confidence from convenient storytelling.

Key Takeaways

Treat assumptions as hypotheses. If a theory of change depends on something important, name it clearly enough that it can be tested.
Focus on the weak links. The most valuable assumptions are the ones that are both uncertain and critical to the pathway.
Use contribution, not purity, as your standard. In complex systems, ask whether the intervention plausibly contributed to change, not whether it single-handedly caused it.
Test the mechanism early. Small pilots and intermediate indicators reveal whether the causal chain is actually moving.
Look for political and institutional assumptions. Many interventions fail not because the idea is weak, but because the system around it does not cooperate.

The deepest lesson: clarity comes from doubt

The paradox of causal planning is that the more seriously you take uncertainty, the more useful your plan becomes. Teams often believe confidence is what makes strategy strong. In reality, the strongest strategies are the ones that know exactly where they might fail.

That is the hidden gift of working with assumptions inside a theory of change, especially when paired with a contribution mindset. It forces a move away from fantasy causality and toward accountable realism. You stop asking for impossible certainty and start building a better relationship with evidence.

The result is not just a smarter evaluation framework. It is a better way of thinking about change itself. Change is rarely a straight line from intention to impact. It is a chain of conditions, many of them fragile, some of them political, all of them worth examining.

So the next time a plan looks neat, ask the question that matters most: what has to be true for this to work? The answer may not make the plan simpler. But it will make it real.