Why most M&E frameworks fail in year two
The pattern is so consistent it could be a checklist. Here's what predicts which frameworks survive.
We have audited more than thirty M&E frameworks built by other firms. A pattern emerges: most fail in year two, and the failures are predictable.
The dominant failure mode is what we call “funder-facing capture” — the framework was designed to satisfy a reporting requirement, not to inform the program. Year one is fine because the funder is paying attention. Year two, when the program team needs to actually use the data to make decisions, the framework cannot answer the questions they are asking.
The five symptoms
Indicators that cannot be collected affordably at frequency. A framework that requires quarterly survey rounds with 400 respondents costs more to operate than the program it is measuring. By year two the field team has quietly stopped collecting it.
Indicators that nobody on the team can explain. “Composite well-being score, weighted by domain salience, normalised to baseline.” Nobody knows what a 0.4 improvement means. The framework gets ignored.
No theory of change. The indicators are not connected to a theory of how the program is supposed to work. When the indicators show no change, the team cannot diagnose why.
No baseline. Or worse: a baseline that was collected after the program started, contaminated by exposure effects.
No plan for what to do with the data. Reports go to the funder. Nobody on the program team reads them.
What survives
Frameworks that survive year two are smaller, slower, and more boring. They have 8–12 indicators, not 40. They have a written theory of change that the program team helped author. They have a documented data-collection workflow that a single fieldworker can complete in an afternoon. They have a calendar of decision points where the data actually gets used.
The boring frameworks are the ones that work.
Article body — approximately 1200 words.