Sales forecasts are so often wrong because deals go into commit on what the rep believes will happen, and nobody checks whether the buyer has done anything that supports that belief.
The short version
A forecast is a list of deals, each with a rep's judgment attached. When the judgment rests on buyer actions (the signer has heard the case, a buyer named the cost of the problem, the close date is tied to something on the buyer's calendar), it tends to hold. When it rests on a good call last week and a rep's hope, it slips.
Most explanations of forecast misses list several causes. Avoma names three: fuzzy stage exits where "commit" means something different to every rep, incentives that reward sandbagging or hero calls, and no tracking of accuracy at all (Avoma). Salesloft adds reps who rate their deals healthier than the buyer signals support, and incomplete CRM data (Salesloft). Each one is real. They meet in the same place: a deal sits in commit without the evidence commit should need, and the process lets it stay.
Why the evidence goes unchecked
Reps are often wrong about their own deals, and they do not know it. Anova Consulting matched reps' stated reasons for losing deals against what buyers told an independent interviewer. Reps had the full and accurate picture 40% of the time, part of the story 28% of the time, and were completely wrong 32% of the time (Anova Consulting). If a rep misreads why a lost deal was lost, the same rep can misread whether an open deal will close.
The CRM does not catch it, because the rep fills in the CRM. And a weekly forecast call that asks "is it still on track?" invites a yes. The question that catches the gap asks for evidence: who signs, when did they last hear the case, and what did they say?
Managers know this, and the constraint is time. A manager with eight reps and 60 open deals cannot test the evidence on every deal every week, so the testing happens on the big deals and the rest ride on belief.
An illustrative slip
This example is illustrative. A rep commits a $90,000 deal in week 4. The champion is keen, the demo went well, and the rep has a verbal "we're aiming for end of quarter." In week 11 the deal slips to next quarter. The reason written up afterwards was that the CFO who signs had never heard the case, and when she did, she asked for a payback figure nobody had prepared.
- Week 4Committed at $90,000
- The CFO who signs never on a call
- No buyer cost on the problem
- The date came from the champion
- Week 11Slips to next quarter
- The CFO asks for a payback figure
All of it was visible in week 4: the CFO had not been on a call, no buyer had put a cost on the problem, and the date came from the champion. Asked for evidence that week, the rep would have moved the deal to best case and had seven weeks to get the CFO on a call. What the rep needed was a 20-minute coaching call that week to talk the deal through and rehearse the ask to the champion.
The weekly check that catches it
The check is four questions, asked of every deal in commit, every week, with the answer judged by what the buyer has done. It takes two or three minutes a deal once reps know the questions are coming.
| Question | An answer that holds | An answer that means best case |
|---|---|---|
| Who signs, and when did they last hear the case? | "The CFO, on the call on the 9th. She asked about the rollout plan." | "Our champion is taking it to her." |
| What does the problem cost, in the buyer's words? | "Their ops lead said about $40k a month in overtime, and the CFO has seen it." | "They know it's expensive." |
| What is the date tied to? | "Their new depot opens on 1 March and they need it live before." | "They said end of quarter." |
| What happens between yes and a signature? | "Legal review, then procurement. Named people, about two weeks." | "Should be quick once they say yes." |
One answer from the right-hand column moves the deal out of commit until the missing piece is in. The rep learns which conversation to have this week, and the manager gets a forecast that changes for a reason. The answers in the table are illustrative.
The deal in the example fails the first three questions in week 4. The coaching that follows is narrow: who can get the CFO onto a call, what the rep asks the champion, and what the CFO will want to see when she joins.
Running it on every deal
The four questions are simple. Asking them of every committed deal, every week, is the part that falls over when a manager has eight reps. Our AI sales experts take the rep's side of it: the rep talks a deal through with Caty on a call, Caty asks for the evidence the team's method needs at that stage, and the rep leaves with the ask for the next buyer conversation. Bryce, who coaches the managers, shows which commits are missing which evidence across the team before the forecast call. If one of your committed deals has a signer who has never been on a call, talk to us while the quarter still has weeks left in it.
Fixing the next forecast
- How to improve sales forecast accuracy: define commit by buyer evidence and trace every slip.
- Deal risk guide: what each stage should require.
- How to run a forecast call: the five evidence questions for every committed deal.
FAQ
Do reps sandbag or over-forecast more?
Both happen, and a team total can hide them by cancelling out. Avoma describes positive bias as hero commits that miss and negative bias as under-calling then over-delivering. Measure bias by rep over several quarters to see which pattern each rep has, then coach to it.
Does better CRM data fix forecast accuracy?
Better CRM data helps, but it does not close the gap on its own. Cleaner fields tell a manager what the rep recorded. The forecast still misses if the deal lacks the buyer evidence commit needs, such as a signer who has heard the case. Check the evidence, then fix the data.
Why do deals slip at the end of the quarter?
Most late slips trace to something missing weeks earlier: a close date the buyer never owned, a signer who had not heard the case, or approval steps nobody mapped. The slip shows up late because that is when the gap meets the buyer's real process.