Most impact metrics measure the programme, not the person
Counting trainings delivered tells you what the organisation did. It says nothing about whether anything changed.
Sector reporting is dominated by activity counts: workshops held, beneficiaries reached, kits distributed. These are easy to collect and almost impossible to misreport, which is exactly why they dominate. They are also nearly uninformative about outcomes.
The metrics that matter are usually harder and slower. Not graduates trained but graduates still earning at twelve months. Not referrals made but referrals closed. Not textbooks converted but children reading on the same morning as their class. Each of these requires following a person rather than counting an event.
Not graduates trained. Graduates still earning at twelve months.
Software changes the economics of this. Following a cohort for two years is prohibitive on paper and trivial once records are structured and identity survives a repeat visit. The reason most organisations report activities is not indifference; it is that outcomes were never measurable at their cost base.
We instrument for the slower number from the start, and we publish the shape of the result even when it is flat. A deployment that moved nothing is more useful to the sector than a deployment that was never measured, and it is considerably cheaper than the next organisation repeating it.
The honest version of an impact claim usually has a confidence interval and a caveat. Funders who can live with that get better information than funders who cannot.