Nobody Ever Comes Back: Why the Savings You Approved Last Year Don't Show Up
Weather forecasters get graded.
Every forecast is checked against what actually happened. There is an entire discipline built around it, and the scores are published every year to a body whose job is to look at them.
It is the reason a seven day weather forecast today is about as good as a three day forecast was forty years ago. Somebody kept score, so the forecasts got better.
Now count the business cases you have approved. Then count the ones anybody checked.
The Number Nobody Checked
Ellen approved a scheduling system eighteen months ago. The case promised four hours a week back for every charge nurse on the floor, plus a reduction in agency coverage.
She has no idea whether any of it happened.
Not because she stopped caring. Because there was never a mechanism.
The project shipped, the team moved on, the savings became something everyone assumed, and the only signal she would ever get is if something went badly wrong.
The forecast was made. Nobody scored it.
This is not a failure of diligence. Approving is a decision with a date on it, and verifying is a chore with no date at all, so it loses every week to something urgent.
The loop stays open because nothing in the calendar ever closes it.
What That Teaches the Room
An unscored forecast is not neutral. It trains everyone who watches.
The person whose savings never materialized paid no price. The person whose savings landed exactly as promised got no credit.
Both submitted, both got approved, and eighteen months later the two are indistinguishable.
So the accurate submitter learns that accuracy is invisible, and the optimistic one learns that optimism is free. Give that a few cycles and every number crossing your desk is inflated, and you have no way to tell which.
Then you do the only thing left. You discount everything, which taxes the honest along with the rest, and the spiral tightens another turn.
You did not create that spiral by being too trusting. You created it by never closing the loop.
Require the Mechanism, Not the Number
The instinct is to scrutinize harder. Tougher questions, more evidence, a colder room.
It does not work, and the reason is structural. Everything in a business case is a projection, so no amount of scrutiny at approval time separates a good projection from a confident one. You are grading the writing.
The move is to stop asking whether the number is right and start asking who brings it back.
Two questions, both before you approve.
What specifically will be different, stated plainly enough that somebody outside the room could check it.
Who reports back, by name, and on what date.
If either answer comes back vague, you have not found a weak number. You have found a request nobody intends to be measured on, which is far more useful to know.
It Changes the Ask Before It Arrives
Once people know the question is coming, the numbers change on the way in. Not because anyone was lying, but because a person writing a figure they will have to defend in nine months writes a different figure than one writing into a void.
You will get smaller numbers. You will get better ones.
Smaller is the part people flinch at, and it should not be.
A projection that survives being checked is worth more than a larger one that evaporates, because you can build a plan on the first and you cannot build anything on the second.
And the first person who comes back and reports the savings landed at sixty percent of projection is worth more to you than ten cases nobody ever checked, because that is the first real data point you have ever had.
Monday
Pick one thing you approved last year. Find out what actually happened.
Not to punish anybody. Just so you know whether your own approvals are worth anything.
Then on the next one, ask for a name and a date before you sign.
-BZ
Comments