The first time I ran a forecast call, I asked every rep the same question: "How's it looking for this quarter?" I got four confident yeses and missed the number by 40%. Nobody lied to me. They just answered the question I actually asked, which was the wrong one.
Open-ended forecast questions get open-ended answers, and reps fill that space with the number that keeps the meeting short, not the number that's true. That's not dishonesty, it's a predictable response to a badly designed conversation. Fix the questions and the numbers get more honest without anyone changing their behavior.
Why "how's it looking" produces a fake number
Two failure modes show up in almost every forecast call, and they pull in opposite directions. Sandbagging happens when a rep who's already hit quota quietly holds a deal out of this week's number so it can rescue a worse week later. Happy ears happen when a rep genuinely believes a deal is closing because the buyer was friendly on the last call, even though nothing in the deal has actually moved. Both produce a confident-sounding number. Neither produces an accurate one.
A generic status question can't tell these apart, because it never asks for evidence, only a feeling. The fix is to stop asking how a deal is going and start asking what specifically has to happen, by when, for it to close, and who besides the rep has confirmed it.
The script, deal by deal
Run this on every deal a rep has called this quarter, in this order, before you let them give you a category like "commit" or "best case." The category comes last, not first.
- "What's the single next action, and who owns it?" Not "following up" or "waiting to hear back." A named action with a named owner. If the rep can't answer this in one sentence, the deal isn't as far along as they think.
- "What has the champion done, not said, in the last seven days?" A forwarded contract, a scheduled procurement call, an intro to legal all count. "They said they're excited" doesn't. This question alone kills most happy-ears deals, because it forces the rep to produce evidence instead of a vibe.
- "Who signs, and have you talked to them directly?" If the rep has never spoken to the actual signer, treat the deal as one stage earlier than it's marked, regardless of what the CRM says.
- "What would have to be true for this to slip a quarter?" This is the sandbagging check, run in reverse. Reps who are sitting on a deal usually answer fast and specifically, because they've already thought about it. Reps who are guessing struggle to name a real risk, which is its own signal.
- "Give me a number, not a category." Ask for a percentage likelihood, not commit or best-case. Categories let a rep round a 30% deal up to "commit" because it feels close. A number forces the same rounding decision out into the open, where you can push back on it.
What a real exchange sounds like
Manager: "Next action on Acme, and who owns it?" Rep: "I'm sending the redlined contract back today." Manager: "What's the champion done in the last week, not said?" Rep: "...they replied to my email confirming budget is approved." Manager: "Confirmed by whom, and did you see it in writing?" Rep: "By their VP, over email, yeah." Manager: "Who signs?" Rep: "The VP, I think, I haven't confirmed that directly." Manager: "Then it's not a 90% deal, it's a 60% deal until you confirm the signer. Move it down, keep the redline moving, tell me next week."
Notice what happened: the rep wasn't caught lying, they were caught assuming, and the number moved down without an argument because it was grounded in a specific gap, not a manager's gut feeling versus the rep's.
Turning the answers into a weighted number
Once every deal has a next action, evidence of buyer activity, a confirmed or unconfirmed signer, and a rep-given percentage, don't just sum the percentages, they're still self-reported. Instead, apply one manager-level discount: any deal where the signer hasn't been directly confirmed gets capped at 50%, no matter what number the rep gave you. That single rule removes most of the optimism bias without requiring you to litigate every deal individually. For the weighted-pipeline math itself, the method we've laid out in how to forecast sales with no historical data still applies here, this script just fixes the inputs going into it.
It's worth running this discipline even when the accuracy gain feels small, because the cost of skipping it isn't abstract. We've covered what a bad sales forecast actually costs in hiring and spending decisions elsewhere, and a five-question script is a lot cheaper than that mistake.
Run this Monday
Before your next forecast call, write the five questions above on an index card and use them verbatim on your three biggest open deals. Don't ask for the category until the end. You'll likely see at least one deal move down in confidence in the first call, not because the rep was hiding something, but because nobody had asked precisely enough to find the gap before.
Frequently asked questions
How long should a weekly forecast call take with this script?
About three to five minutes per deal once reps get used to the format, so a rep with six open deals should take 20 to 30 minutes. It's slower than a status round the first few weeks and faster once reps learn to come prepared with the evidence instead of a feeling.
Will reps feel interrogated by this many questions?
Some do at first, because it's a real change from a status update. Framing it as "help me not surprise the board" rather than "prove you're not lying" defuses most of that, and reps who are actually on top of their deals tend to like it, since it's their fastest path to a manager backing off.
What if a rep genuinely doesn't know the answer to one of these questions?
That's the point, not a failure of the script. "I don't know who signs" is more useful information than a confident guess, because it tells you exactly what needs to happen before next week's call, and it's the reason that deal's number should come down today.
Does this replace CRM-based forecasting tools?
No, it feeds them. A forecasting tool can only weight the data it's given, and this script is how you make sure the stage, close date, and confidence a rep enters are grounded in evidence instead of optimism before they ever hit the dashboard.