Four figures appear across these pages. This page gives each one its dataset, its observation window, the population it was computed over, the formula behind it, and — more usefully — what it does not establish.
A figure is only useful if its limits travel with it. These are the ones that matter, stated plainly rather than buried.
The coaching agreements preclude disclosing engagement detail to a third party. An external auditor would need exactly what those agreements protect — who was coached, at which employer, into which role, at what compensation.
So the absence of an external audit here is a consequence of the confidentiality terms, not an unwillingness to be checked. The definitions above are published in full precisely because the underlying records cannot be. The First Discipline applies: no coaching client is named anywhere, ever.
Two figures previously published here have been removed rather than defended. A “35% faster time-to-impact than open-ended coaching arrangements” comparison was withdrawn because no comparison group, method or window supported it. A 92% repeat-client ratio was withdrawn because it had not been given a definition precise enough to publish.
Both are recorded here rather than quietly deleted. A methodology page that only lists surviving claims is not a methodology page.