A forecast built as a best-estimate becomes a commitment the moment it leaves the spreadsheet. Three warning signs the conversion has already happened.
🎧 Hear this post
It’s 7:42pm on the last Thursday of Q3 and the CFO is rebuilding next quarter’s revenue forecast in a new tab because sales leadership doesn’t like the old one.
The old one is not wrong. It is uncomfortable. So it gets a new assumption — upsell conversion bumped two points — and becomes the number that goes into the board deck.
Four weeks later, that number is a target. Four weeks after that, the sales team is measuring themselves against it. Four weeks after that, someone is getting fired for missing it.
Nobody in that 7:42pm meeting thought they were making a commitment. They thought they were doing math.
A forecast is supposed to be a prediction — a best estimate of what will happen given the inputs on hand. The moment it is presented to a board, a lender, a new hire, or a sales team, it stops being a prediction. It becomes a promise. The person who built it didn’t mean to make a promise. That doesn’t matter.
The gap between forecast as prediction and forecast as promise is where most ops dysfunction lives. The forecast was supposed to inform the hiring plan. Now the hiring plan is defending the forecast. The forecast was supposed to set investor expectations. Now investor expectations are setting the forecast. The forecast was a range with a midpoint. Now it is a single number, and the range has quietly disappeared.
Three warning signs that a forecast has converted from prediction to promise:
One — the sales leader has memorized it. Nobody memorizes a range. People memorize commitments.
Two — hitting the forecast has replaced hitting the business plan in casual conversation. The forecast was supposed to be the map. It has become the terrain.
Three — the person who built it has stopped including a downside case. They know, correctly, that the downside case will be ignored. So they stop writing it down. Now the forecast has no bottom.
The fix isn’t more rigor in the model. It’s a rule: any forecast presented externally carries three numbers, not one. The 10th percentile, the midpoint, and the 90th. If the room wants to pretend only the midpoint matters, that is the room’s choice — but the other two numbers are on the page, and anyone reading the deck in eighteen months can see them.
The forecast built in a tab at 7:42pm on a Thursday will be read back to whoever built it in eighteen months by someone they haven’t met yet. Make sure it has a range.
0 Comments