The forecast is the most consequential number sales gives the board. It is also the only number in the company reconstructed from the memory of people who are tired after eight calls.
Anatomy of a Friday forecast
It always looks the same. Friday afternoon, the manager asks for a pipeline update. The rep opens twelve opportunities and tries to recall what happened in each over the last fortnight.
The first three go well. By the ninth, all that remains is a vague sense that it was going fine. The stage stays where it is, because changing it would require a judgement and there is no evidence for one. The amount stays too, because where would a different number come from?
Forecasts are not falsified. They are reconstructed — and memory has a strong bias in favour of one's own work.
Three systematic errors
Forecast errors are not random noise. They lean consistently in the same direction, which makes them far more dangerous than noise would be.
- Stage optimism. A deal where the buyer said 'let's revisit this' is marked active. So is a deal that has gone silent for three weeks. Both look identical in the report.
- Anchoring on the first number. The value entered when the opportunity was created rarely changes, even after the scope visibly shrank. Nobody goes back to revise it downward.
- Invisible blockers. The real reason a deal is stuck was usually stated out loud — a CFO change, a budget freeze, an internal alternative. There is no CRM field for it, so it evaporates.
What changes when call content reaches the CRM
When a summary of every conversation lands in the CRM automatically, forecasting stops being a memory exercise and becomes a reading of current state.
Stages move when something actually happened, not when someone remembered. Every opportunity carries the specific reason it is stuck. And the manager can stop asking 'how's it looking?' and start asking 'I see the buyer mentioned a change on the board — what's our move?'
A signal score — a 0-10 rating computed after each call from budget, decision-maker, timeline and blockers — adds a second layer. Deals that look healthy on paper but score low surface themselves. Those are usually the ones that will not close this quarter.
A test you can run this afternoon
You do not have to take anyone's word for this. There is a simple check that takes half an hour.
Pick ten deals from your last closed quarter. For each, note when the stage last changed and when the last customer conversation happened. If the average gap exceeds a week, your forecast describes last week — and you are making decisions with it today.
Second check: count how many closed-lost deals have a loss reason other than 'price'. If price dominates, that does not mean you are expensive. It means the real reasons were spoken aloud in calls and never written down anywhere.
Key takeaway
A forecast is only as good as the freshness of the data beneath it. If stages change a week after the call that moved them, you are forecasting the past.