Ask a rep about a deal and you get a story: the champion loves it, legal has the paper, the CFO is “aligned.” Ask the same rep what share of deals at this stage closed last year, and you usually get silence. That silence is most of the forecasting problem.
This paper is about the judgment itself. The incentive side, where reps enter data they know is wrong because the system rewards it, is covered in why good reps enter bad data. Here the rep is trying to call it straight and still missing.
Why are sales reps bad at forecasting?
Because a deal-level forecast asks for a probability, and reps are positioned to see stories, not probabilities. They know the buyer and not the base rate. In Xactly’s 2024 Sales Forecasting Benchmark Report, 52% of sales leaders said their forecasts are off by 10% or more, while about two-thirds of finance leaders (66%) said they are usually off by less than 9%.
Finance does not know more about the deals. Finance forecasts from history and rates; sales forecasts from conversations. Both methods have a blind spot, and the one that comes with conversations is bigger, because every conversation is designed, by both sides, to sound encouraging.
Sales leaders know it. Gartner’s State of Sales Operations survey, published in February 2020, found only 45% of sales leaders and sellers had high confidence in their organization’s forecasting accuracy. Fewer than half of the people producing the number had high confidence in it.
The inside view: why every deal looks like the exception
Psychologists Daniel Kahneman and Amos Tversky named it the planning fallacy: people predicting their own projects lean on the specifics of the case and discount how similar cases turned out. A rep in the middle of a deal is the purest example. Every detail that makes this deal feel winnable is vivid. The deals at the same stage that died last quarter are not in the room.
The corrective is what forecasters call the outside view: start from the base rate, then adjust for the specifics. For a rep, the base rate is their own history. Of the deals that reached this stage in the last year, what share closed, and how long did it take? Almost no CRM puts that number in front of the rep at the moment they set a close date or pick a category. The default stage probability is the closest thing, and in HubSpot it is a generic ladder from 20% to 90% that has nothing to do with the rep or the company. We cover that in why your HubSpot forecast is always wrong.
Experience does not cure this on its own. A veteran rep has more stories, not more base rates, unless someone shows them their numbers. The best forecasters we have worked with were not the most optimistic or the most cautious. They were the ones who knew their own conversion rates by stage and quoted them unprompted.
Happy ears and sandbagging are the same error, pointed in opposite directions
Happy ears is hearing buying signals that are not there and forecasting a deal too high. Sandbagging is holding back deals you expect to win and forecasting too low. Both are the rep managing the forecast instead of reporting it: one to stay hopeful, one to stay safe. Both leave fingerprints in CRM data that a manager can see without asking the rep a single question.
| Behavior | What the rep is protecting | What it looks like in the CRM |
|---|---|---|
| Happy ears | Hope, and a pipeline that looks covered | Late-stage deals with no buyer activity in weeks; close dates pushed again and again; single-threaded deals in commit |
| Sandbagging | A safe number and an upside story for later | Deals created and closed in the same period; stage jumps in the last weeks of the quarter; commits that run under what closes |
| Stale pipeline | Avoiding the close-it-lost conversation | Open deals with past close dates, no next step, no activity in months |
| Late entry | Time, and privacy from inspection | Deals created at proposal stage; recorded sales cycles far shorter than real ones |
Each behavior is rational somewhere. Sandbagging is rational when the comp plan and the culture punish a miss harder than they reward an accurate call. Happy ears is rational when pipeline coverage is the metric a rep gets reviewed on. That is why lecturing about either rarely lasts, and why changing what gets inspected usually does. Precise definitions of each term are in our forecasting glossary.
Why don’t reps update the CRM before the forecast call?
Because CRM updates are a chore done in a batch, usually right before the meeting that inspects them. Between batches the deal moves and the record does not, so the forecast you review on Monday describes what the rep remembered on Sunday night. Fewer required fields, updates tied to buyer events instead of calendar deadlines, and automatic activity capture all shorten that lag.
The gap matters more than it looks, because memory is where optimism does its best work. A rep who logs the next step while it is fresh records what the buyer committed to. A rep reconstructing a week of calls records what they hoped the buyer meant.
Every field you add to the update makes the batch longer and the lag worse. Reps already lose most of their week to work that is not selling, as we covered in why sellers lose 60% of their week to everything but selling. A deal record with forty required fields guarantees they update the minimum, the night before. Cutting fields is a forecasting decision as much as an admin one.
What can sales managers do about inaccurate rep forecasts?
Four moves make rep forecasts measurably better: inspect buyer evidence instead of rep confidence, show every rep their own stage-to-close history, define forecast categories in writing with the evidence each requires, and track each rep’s forecast bias over time so coaching targets direction. None needs new software. All four need a manager willing to run the forecast call differently.
1. Inspect evidence, not adjectives
“How do you feel about it?” invites a story. Ask for facts the buyer created.
- Forecast call questions
- What did the buyer do since our last call, and on what date?
- Who signs, and have you spoken with them directly?
- What date did the buyer give, and what happens on that date?
- What would have to be true for this deal to slip, and is any of it true yet?
- If this deal moves out of the period, what replaces it?
2. Give reps the base rate
Show each rep, each quarter, what share of their own deals closed from each stage and how long it took. A rep who sees that 22% of their demos closed last year will hesitate before calling a demo-stage deal commit. The number does the coaching.
3. Write the category definitions down
Commit should require evidence, for example a buyer-confirmed date, an engaged signer and paperwork in motion, not a feeling. Write the bar for commit and best case into the forecast process, and hold reps to the definition, not to the outcome. Our glossary has a starting version.
4. Track bias by rep, and coach the direction
A rep who over-calls by 15% every quarter needs a different conversation from one who swings 15% either way at random. Track each rep’s signed forecast error for four or more periods, using the formula in our sales forecast accuracy guide. Then make accuracy, not optimism, the thing that gets noticed.
The data under the judgment
Even a well-calibrated rep cannot call a deal from a record that is wrong. Duplicate contacts split a buyer’s activity across two records, deals without contacts hide who is engaged, and stale close dates distort the base rate itself. Salesforce’s State of Sales, 7th edition, published February 2026, found high performers prioritize data hygiene at 79%, against 54% of underperformers.
That survey covered 4,050 sales professionals across 22 countries, and the gap is a correlation, not proof. It fits what I’ve found coaching reps: judgment work pays off fastest on a CRM where the records can be trusted, because the rep and the manager are finally looking at the same deal.
If you want a measured read of the records your reps forecast from, the AeolusGTM CRM Diagnostic runs a read-only scan of HubSpot, Salesforce or Pipedrive and shows where open pipeline has no recorded evidence behind it: missing next steps, missing activity and single-threaded deals. HubSpot teams can start with the free HubSpot audit.
You can keep asking reps how they feel about their deals. Or you can give them the base rate, an evidence bar and a clean record, and ask them for a probability. Only one of those conversations gets better with practice.