Research & Evidence
We show our working.
Every claim traced to a named source, graded, and scheduled for re-verification — including the parts we get wrong.
The method
Every claim traced to a named source, graded, and scheduled for re-verification — including the parts we get wrong.
We describe these books with two words before any other: researched, and audited. The first is ordinary; the second is the one we mean most. Every claim that carries weight — every number, every dated statement, every framework named as load-bearing — is traced to a named source, graded for how much that source can bear, and put on a schedule to be checked again before it quietly goes stale. Government labour-market data is treated as the strong ground it is, and stated precisely. A practitioner's working range is marked as a range, not dressed as a measurement. And the books are honest about their own limits: where a claim is a structural observation rather than a measured figure, it says so on the page, and where a number has not earned its place, it does not appear.
That last discipline is the one we are proudest of, because it is the one nobody can see. When a number does not survive scrutiny, we cut it — and keep the record of the cut. Beneath the apparatus is the longer endeavour these books serve: a decision-making company building, inductively and in the open, an honest understanding of how good decisions get made. The evidence work is how we keep ourselves honest while we build it.