The number has sat in the steering pack for two years: eighty per cent mapped, risk rated low. Nobody built it. Everybody uses it. Then the cutover weekend arrives, six on Saturday morning. The delivery lead is on. The vendor is on. Someone from the original business case is on, camera off, mic on mute.
Nobody asks her anything.
Not because she does not know. Not out of malice. The question simply does not form, not in front of a sponsor, on a call this exposed, when everyone already suspects what the honest number costs.
Here is what a convenient number is for.
A baseline is not a measurement. It is an agreement. It sat in the RAID log, green. It sat in the contract, assumed. It sat in the steering deck, unquestioned. None of that made it true. It made it load-bearing.
Convenient numbers survive on small refusals. The gap raised in a design review, ruled out of scope. The flag buried in an appendix nobody opened. The milestone that needed one more quarter, so the question waited for the quarter after. Each refusal reasonable. None of them a correction.
She built the eighty per cent so the unknown would look small enough to fund. She has known for a year the truth is closer to fifty. She never lied. Nobody ever asked precisely enough for that to count as a lie.
You cannot govern your way to an accurate baseline. No steering committee can endorse it into being true. No RAID log retires a risk by marking it green. Correction happens when somebody with nothing left to protect says the real number out loud.
Every programme has two dependency maps. The one in the business case, and the one the cutover team builds at six in the morning, from whatever is still running. Only one of them was ever going to be tested.
The funding was approved against the first one.
— Signal / unfiltered observations from the field. Fortnightly.
#VelocityArchitecture #EnterpriseArchitecture #ArchitectureLeadership #ITGovernance #ITStrategy #Execution #CIO #DigitalTransformation