Innovation Advisory · Framework 08
A 2×2 matrix · 4 quadrantsThe Rumsfeld matrix, adapted for programme design and MEL work. Four quadrants to map what the programme knows, what it knows it does not know, what it has decided not to look at, and where the surprises will come from.
The evidence base
What do we already know, and how strong is the evidence?
The established literature, previous evaluations, baseline findings, settled facts about the population and the system. List them honestly. Then ask which of them are still true today, and which are quietly out of date but still on the slides.
The questions
What gaps in the evidence have we agreed to investigate?
The hypotheses the MEL plan is designed to test. The questions the funder wants answered. The bits of the pathway nobody has measured yet. If this quadrant is small, you have either solved the problem or stopped looking.
The tacit and the suppressed
What do we already know but have not named — including what we have decided not to look at?
The frontline worker who has run this programme before and was not invited to the design meeting. The biases and assumptions baked into the choice of indicator. The findings from the last evaluation that were politically inconvenient. This is the quadrant most programmes pretend does not exist.
The blind spots
Where will the surprises come from?
The shocks the design has not modelled, the cross-sectoral consequences nobody mapped, the populations the sample frame cannot reach. Run a pre-mortem: imagine the project failed in eighteen months and write the obituary. The cause of death usually lives here.
The matrix has two axes. The horizontal axis is the knowledge dimension — what the programme has versus what it lacks. The vertical axis is the awareness dimension — what the programme has noticed versus what it has not.
Most programme design lives in the top half — knowns and acknowledged gaps. The bottom half is where the trouble usually starts. Unknown knowns is the quadrant for tacit knowledge, suppressed findings, and silent assumptions: the things in the room that nobody is naming. Unknown unknowns is the quadrant for the genuine surprises — the system-level shocks, the equity dimensions you did not think to disaggregate by, the second-order consequences in adjacent sectors.
Used as a pre-mortem, the matrix forces the team to populate the bottom row before the project starts. Used as a review, it makes the team write down what it has agreed not to look at. Pair it with the Causal Pathway Review when stress-testing a Theory of Change, and the Indicator Test when stress-testing the metrics.