Ask a programme team to show you their theory of change, and you will often be handed a colourful diagram that was drawn once for a proposal and never opened again. That is a missed opportunity. A good theory of change is not a graphic. It is a working argument about how and why your activities will lead to the change you want, and it should guide decisions long after the proposal is approved.
What a theory of change is, and what it is not
A theory of change explains the pathway from what you do to the change you seek. Activities lead to outputs, outputs contribute to outcomes, and outcomes contribute to longer-term impact. Crucially, it also states the assumptions that must hold for each step to happen.
A logical framework, or logframe, is a related tool that organises the same logic into a matrix of objectives, indicators, means of verification and assumptions. A useful way to think of it is that the theory of change is the narrative and the logframe is its tabular summary.
A theory of change is not a list of activities, and it is not a diagram with arrows pointing in every direction. If you cannot explain each arrow in a sentence, the thinking is not finished.
Why theories of change fail
- Outputs described as outcomes. “One hundred people trained” is an output. “Trained people are earning more” is an outcome.
- No assumptions. Without them, there is nothing to test when results fall short.
- Too many pathways. A model that tries to explain everything helps no one focus.
- Built without the delivery team. A framework drawn up at head office or by a consultant alone rarely survives contact with the field.
- Never revisited. The model drifts from reality because nobody owns it.
Seven steps to build one that works
- Start with the problem. Define it using evidence, including who is affected and why it persists.
- Define the long-term change. Be specific about who will be better off, in what way and by when.
- Work backwards. Identify the outcomes needed to reach that change, then the outputs and activities that produce those outcomes.
- Surface assumptions and risks. For each link, ask what must be true for this to work and what evidence you have that it is.
- Choose indicators. Select a small number at output and outcome level that will show whether the pathway is working.
- Test with stakeholders. Walk the draft through with programme staff, partners and, wherever possible, the people the programme serves.
- Plan to review. Schedule a review at least once a year, and after any major change in context.
A short illustrative example
Imagine a youth employability programme. The problem is that young people in a county complete training but struggle to find work. The activities are skills training, mentoring and employer engagement. The outputs are the number of young people trained and placed in internships. The outcomes are the share who are in paid work or self-employment six months after completion, and their earnings. The impact is improved livelihoods for young people and their households.
The assumptions are what make the model useful. Employers in the area have vacancies and are willing to host interns. Trainees can afford transport and have access to mentors. Training content matches what employers actually need. When placement results come in below target, the team knows where to look first: the assumptions about employer demand and training fit, rather than simply blaming effort.
|
Level |
Example |
Possible indicator |
|
Activities |
Skills training, mentoring, employer engagement |
Number of training cohorts delivered |
|
Outputs |
Young people trained and placed in internships |
Number completing training; number placed |
|
Outcomes |
Graduates in paid work or self-employment |
Share in work six months after completion; average earnings |
|
Impact |
Improved household livelihoods |
Change in household income over time |
From theory of change to indicator matrix and data plan
Once the pathway is agreed, the rest of the monitoring system follows naturally. The logframe or results framework sets out each objective with its indicators and targets. An indicator matrix then specifies, for every indicator, the definition, baseline, target, data source, collection method, frequency and responsible person. A data collection plan schedules the surveys, routine monitoring and review meetings.
This is the point at which a theory of change either becomes useful or remains decorative. If an indicator cannot be traced back to a step in the pathway, question whether it is needed. If a step in the pathway has no indicator, question whether you will ever know if it is working.
Using the theory of change to learn
The most valuable use of a theory of change is in regular reflection sessions. Each quarter, bring the team together to ask: What do the data show? Which assumptions are holding and which are not? What should we change? Record the answers and update the model accordingly.
This is the core of adaptive management, and it is also what donors increasingly mean when they ask for “learning”. A programme that can show how its understanding changed, and what it did differently as a result, is far more convincing than one that reports every target met without explaining how.
How Melanel can help
Melanel designs complete monitoring, evaluation and learning systems, including the theory of change, logical framework, indicator matrix, results framework and data collection plan. We also train programme teams in results-based management so that they can maintain and adapt the system themselves.
