Why is our forecast wrong?

A forecast is wrong when it's built on statistics instead of commitments. The typical pipeline calculation — volume times historical win rate — looks mathematically sound but says nothing about whether real buying decisions stand behind the deals. Precisely wrong is more dangerous than roughly right, because it feels like control.

I experienced this on a large scale: a 300-million pipeline, comfortable on paper. The forecast rested on a statistical funnel — not on commitments. When we checked deal by deal who had actually decided at the customer, only a fraction remained. The number had been correct. The system behind it hadn't.

The mechanics behind this are human, not mathematical. Three patterns create the gap between forecast and reality.

First: stages reflect activities, not customer decisions. 'Proposal sent' is a salesperson's step. It says nothing about whether the customer has budgeted, prioritised and sold internally. A forecast based on salesperson steps measures your own effort, not the probability of purchase.

Second: the pipeline isn't cleaned up out of politeness. Deals with no activity for months stay in because removing them looks like giving up. Every quota calculated on this pipeline is calculated with dead entries.

Third: corrections get individually explained away. Every slippage has a plausible story — procurement, holidays, quarter-end. Individually, each story is true. That they're piling up is the real information, and that's exactly what gets lost in the individual explanations.

The way out is uncomfortable but simple to state: the forecast must shift from the question 'What have we done?' to the question 'What has the customer decided?' That changes stages, reviews and leadership conversations. And it changes the number in the short term — usually downward. But then it's correct.