The Ownership Metric: the number we steer on.
Ask a firm what number it steers on and you will learn what it is building. Revenue under contract builds a longer stay. Utilisation builds a bigger bench. We wanted a number that builds an exit.
What counts
The Ownership Metric is the share of AI-enabled work your team ships without us. Not logins, not usage, not a satisfaction score. Work that ships, unaided.
"Without us" means exactly that. Your people run the execution, the quality control, the prioritisation and the maintenance. Work that still needs H3 in any of those roles does not count.
How it moves
It starts low, on purpose. A founding partner runs the first cycle with your people, inside the real work. Each cycle after that, your team takes more of it: first the execution, then the quality checks, then the calls about what to build next. We pull back as ownership grows. The number rises.
It is designed to rise every cycle. The plan is for us to become unnecessary.
A first cycle usually runs ninety days, from diagnosis to handover. Long enough to watch the number move, and to see who moved it.
Usage counts the wrong thing
Usage is the number most programmes report, because usage is easy to count. It tells you a tool was opened. It says nothing about whether the work changed hands. A team can be fully adopted and still ship nothing without help. That pattern has its own field note: why pilots stall.
So usage is a signal we read, never the proof. We steer on returns, on how the work feels for the people doing it, and on the share of the work you own.
What it does to us
A metric only means something if it can hurt the one reporting it. This one prices our exit. Every point it rises, our role shrinks and the handover gets closer. That is the point. The goal is not dependence on H3. The goal is a capability that stays when we leave.
Our number. Nobody else can honestly quote it.
← All field notes