What a theory of change actually is
Strip away the sector language and a theory of change is a simple thing. It is your account of why the work you do leads to the change you are trying to make. You write it down so that you, your colleagues and anyone reading your reports can follow the reasoning and see whether it stands up.
It is not a mission statement. A mission statement says what you care about. A theory of change says what you do, what you expect to happen because of it, and what you are assuming in between. That last part is where most of the value sits.
It is also not a document you write once and file. It is a working hypothesis. You put it on paper precisely so that evidence can support parts of it, and challenge others.
Why you get asked for one, and why it helps anyway
Grant applications ask for one because a funder wants to know that money going into activity has a plausible route to something changing. Commissioners ask for something similar in a service specification. Boards ask because trustees are accountable for whether the organisation is doing what it says.
If nobody has asked you, write one anyway. The exercise forces an honest look at the gap between activity and outcome. Most organisations can describe their activity in detail and their outcomes in slogans. Writing a theory of change makes that imbalance visible in about twenty minutes.
For a group with no funding yet, it does another job. It is the clearest way to set out a case before you have results to show. You cannot yet prove what changed. You can set out what you intend to change, why you think it will, and what you will collect to find out. That is a credible starting position, and it beats an application built entirely on enthusiasm.
The core structure
Almost every version of a theory of change uses the same five stages. The vocabulary varies between funders. The logic does not.
Inputs
Six trained volunteer mentors, a room in the community centre, £8,000 for the year.
Activities
Weekly one-to-one mentoring sessions with young people aged 13 to 16.
Outputs
40 sessions delivered to 15 young people over two terms.
Outcomes
Young people feel more confident asking for help and attend school more regularly.
Impact
Fewer young people in the area drift out of education without support.
Reading it left to right
- Inputs are what you put in: people, money, time, space, kit.
- Activities are what you do with them. This is the bit organisations describe best.
- Outputs are the countable results of the activity: sessions run, meals served, people seen.
- Outcomes are what changed for someone. Confidence, skills, health, stability, circumstances.
- Impact is the longer-term or wider change you contribute to. Be careful here: contribution is honest, sole credit usually is not.
One warning about the last box. Impact at that scale is shaped by schools, families, housing, health services and luck. Claiming your programme caused it is the fastest way to lose a careful reader. Say you contribute to it, and be specific about the outcomes you can actually show.
Mistake one: outputs dressed up as outcomes
This is the most common problem in theories of change and in the reports that follow them. An output is something you did. An outcome is something that changed for a person. “We ran 40 sessions” is an output. “Young people felt more confident” is an outcome. The first is easy to count and easy to prove. The second is what the reader wants.
| Output: what you did | Outcome: what changed |
|---|---|
| Youth mentoring40 mentoring sessions delivered to 15 young people. | Young people say they have an adult they trust to ask for help. |
| Food support620 food parcels distributed over six months. | Households report skipping fewer meals at the end of the month. |
| Employment support12 CV workshops run, 48 attendances recorded. | Participants apply for work they previously ruled themselves out of. |
| Befriending1,100 phone calls made to older residents. | Residents describe feeling less isolated week to week. |
A useful test: could a stopwatch or a clipboard produce this number without asking anyone anything? If so, it is an output. Outcomes almost always require someone to tell you something, or a record from elsewhere that shows a change.
Mistake two: never stating the assumption
Most theories of change list activities on the left, outcomes on the right, and leave the middle blank. The middle is the interesting part. It is the belief about why the activity produces the outcome.
Every arrow in a theory of change hides a sentence beginning “we believe this works because…”. Write that sentence down.
In the mentoring example the assumption might be: young people disengage from school partly because they have no adult outside the family who takes an interest, and a consistent weekly relationship with one adult changes how willing they are to ask for help.
Stating that does three things. It makes the reasoning checkable, so a trustee or a funder can agree or push back rather than nodding at a diagram. It tells you what to collect, because the assumption points straight at the question you should be asking people. And it lets you be wrong usefully. If the evidence comes back saying the relationship mattered less than the practical help with coursework, you have learned something specific rather than concluding the programme “did not work”.
Write one assumption per arrow. Keep each to a sentence. If you cannot write it, that arrow is probably a hope rather than a theory, and it is better to know that now.
The template
Here is the whole thing as something you can copy. The worked example uses the youth mentoring programme. Underneath it is the same structure, blank, ready for your organisation. Copy either into a document, a wiki page or a wall of sticky notes. Nothing about this needs software.
Worked example, youth mentoring
Plausible illustrative figures, not real results.
- The problem
- Young people in the area who lose contact with school rarely have an adult outside the family taking an interest in them.
- Who this is for
- Young people aged 13 to 16 referred by two local secondary schools.
- Inputs
- Six trained volunteer mentors, a weekly room booking, £8,000 a year, one part-time coordinator.
- Activities
- Weekly one-to-one mentoring for two terms, plus a termly group session.
- Outputs
- 40 sessions delivered, 15 young people supported, 12 completing both terms.
- Our assumption
- A consistent relationship with one adult outside the family makes a young person more willing to ask for help before a problem becomes a crisis.
- Short-term outcomes
- Young people name an adult they would go to with a problem. They turn up to sessions they arranged themselves.
- Longer-term outcomes
- More regular school attendance, and fewer unresolved issues escalating to exclusion.
- Contribution to wider impact
- Alongside schools and families, fewer young people locally drift out of education without support.
- How we would know
- A short conversation at the start and again at the end of the two terms, mentor session notes, and attendance figures shared by the schools.
- What could make this wrong
- Attendance may improve for reasons nothing to do with us. Young people may tell a mentor what they think the mentor wants to hear.
Blank template, your programme
Fill each row with one or two sentences. If a row is hard to write, that is the row worth discussing with your team.
- The problem
- Who this is for
- Inputs
- Activities
- Outputs
- Our assumption
- Short-term outcomes
- Longer-term outcomes
- Contribution to wider impact
- How we would know
- What could make this wrong
A printable one-pager. Use your browser's print or “save as PDF” option.
Using it once it is written
- Show it to someone who does not work for you and ask which arrow they doubt. That is the one to evidence first.
- Turn the 'how we would know' row into the two or three questions you actually ask people.
- Revisit it once a year, or whenever a piece of evidence surprises you.
- Read the companion guide on outcomes versus outputs before you write your next report.
One last honest note
A theory of change is a starting hypothesis, not a finding. It should look different in two years, because by then you will know things you are currently guessing at. The versions that gather dust are the ones written for an application and never opened again; the useful ones get marked up as evidence comes in.
ImpactProof keeps your theory of change and the evidence for it in one place, so it can evolve instead of gathering dust.
