Sales coaching conversations tend to center on the same things: call activity, discovery quality, deal velocity, objection handling. These are meaningful, but they're all lag indicators — post-mortems on deals that already happened. There's a different class of leading indicator that most managers have data on but rarely use systematically: how accurately each rep forecasts their own deals.
Rep forecast accuracy is not just a RevOps metric. It's a window into deal judgment. A rep who consistently over-forecasts deals by 40% isn't just creating a noise problem in the commit call — they're likely making optimism errors at the deal qualification level too. A rep who consistently under-forecasts by 30% might be sandbagging, or they might be working deals with higher uncertainty than average, or they might lack confidence in their pipeline for reasons worth understanding. The direction and magnitude of the bias are both diagnostic.
Measuring What "Forecast Accuracy" Actually Means
Before you can coach from forecast accuracy data, you need to agree on what you're measuring. The most common approach is comparing a rep's weekly commit call to their actual closed-won revenue for that quarter. But this conflates two different things: the accuracy of the quarter-level summary and the accuracy of individual deal-level assessments.
A more useful measurement framework tracks both. Quarter-level accuracy tells you whether the rep's portfolio of deals lands near their stated commit. Deal-level accuracy tells you whether the rep correctly assessed which specific deals would close and at what probability. A rep can have a high quarter-level accuracy by luck — two optimistic deals and two pessimistic deals that happen to average out — while having poor deal-level accuracy on every individual call. That's a very different coaching situation than a rep who is consistently calibrated at the deal level.
The metric we've found most useful for deal-level accuracy is a simplified calibration score: across all deals a rep classified as "commit" in a given quarter, what percentage actually closed? Across all deals classified as "best case," what percentage closed? If a rep's commit category closes at 45% instead of the expected 75-85%, their commit signal has low reliability and you should apply a systematic haircut to their quarterly call. If their best case category closes at 60%, they're actually better at recognizing upside than their conservatism suggests.
The Systematic Optimism Problem
In most B2B SaaS sales teams with 6-12 months of measurable history, about a third of reps are systematic optimists. Their deals consistently look better in the pipeline view than they eventually turn out. Their close dates slip regularly. Their commit totals exceed actual closed revenue by 20-40% in a typical quarter.
The coaching conversation with a systematic optimist is often uncomfortable because the rep has usually convinced themselves — and sometimes their manager — that each miss was attributable to external factors: procurement delays, champion changes, timing issues. Some of those explanations are true, but the pattern across quarters reveals that the optimism is intrinsic to how the rep evaluates deals, not situational bad luck.
What changes when you have the data is the nature of the conversation. Instead of saying "I think you tend to be optimistic about deals," you can say: "Over the last three quarters, your commit category has closed at 52% while the team median is 78%. The gap isn't coming from your deal size — your deal size distribution looks similar to the team's. Let's look at the five specific deals where you called commit and didn't close, and see what they have in common." That's a diagnostic exercise, not a judgment call. It's much easier for a rep to engage with data than with a manager's opinion.
What Systematic Under-Forecasting Actually Signals
The under-forecasting case is often handled as if it's only a sandbagging problem — reps who hide good deals to manage expectations or protect their quota for the next quarter. Sandbagging is real and worth addressing. But systematic under-forecasting also appears in two other patterns that require different coaching approaches.
The first is pipeline quality anxiety. A rep who consistently builds a large pipeline and forecasts conservatively might be doing so because they've learned that their deals have a higher-than-average slip rate in the final stages, and they're compensating by under-calling. The right question here is: why are their deals slipping at higher rates? Is it a discovery problem? A qualification problem? A champion access problem in the accounts they're working? The under-forecasting is a symptom, and the treatment isn't to tell the rep to forecast higher — it's to address the underlying pattern that's creating the uncertainty.
The second pattern is territory or segment mismatch. Reps covering accounts with more complex procurement environments, longer decision cycles, or less clear champion structures will tend to have naturally higher deal-level uncertainty than reps with simpler territory profiles. If you're seeing systematic under-forecasting from a rep in a specific segment or account type, it might be telling you something about whether the go-to-market motion matches the complexity of that segment — not just something about the rep's calibration.
Using Accuracy Scores Without Creating Sandbagging Incentives
This is the tension that makes most sales managers cautious about making forecast accuracy visible: if reps know they're being measured on accuracy, they'll sandbag. They'll only call commit on deals that are essentially signed, which makes the commit call useless as a planning instrument and pushes all the uncertainty into "best case," which everyone then ignores.
We're not saying this risk doesn't exist — it's real. But it's a function of how accuracy is measured and what it's tied to. If accuracy is measured as "commit deals that close as a percentage of commit total," and it's used as a performance evaluation metric, reps will defend their accuracy by shrinking their commit category. That's the wrong framing.
The framing that works better is calibration rather than accuracy. Calibration means your probability estimates correspond to reality at the stated confidence level — not that you only forecast things you're certain about. A rep with good calibration can say "these five deals are in my commit at 80% probability" and have roughly four of them close. That's a calibrated forecast even if one deal slipped. Compare that to a rep who only calls commit on deals that are 99% certain and never misses, but has only called commit on three deals all quarter. The first rep is more useful for planning purposes even though their raw miss rate is higher.
When accuracy is framed as calibration, the coaching conversation is about quality of judgment rather than outcomes. Managers can celebrate good calibration and coach around miscalibration without creating incentives to sandbag or over-conserve.
Building a Rep Accuracy Baseline
To use forecast accuracy for coaching, you need enough historical data to distinguish signal from noise. At minimum, four to six quarters of weekly commit calls and actual outcomes per rep. With less than that, individual quarters can be misleading — a rep might have had one large deal slip for an entirely legitimate reason, and calling them an optimist on the basis of that one quarter is wrong.
With four-plus quarters of data, patterns become stable. The reps who are systematically biased in one direction will show it clearly. More importantly, you can start to decompose the accuracy score by deal characteristics: does the rep's accuracy differ by deal size? By sales cycle length? By whether the deal involved an inbound vs. outbound motion? By segment? These decompositions often point to specific skill gaps that are much more actionable than aggregate "you over-forecast" feedback.
One calibration check worth running: compare a rep's forecasted probability on a given deal to the model's probability for the same deal. When a rep calls commit on a deal that the model says is a 45% close probability, you want to understand whether the rep has genuine information the model doesn't have access to (a conversation they had last week, a champion signal they picked up in a call) or whether they're simply more optimistic than the evidence supports. That comparison gives managers a specific, deal-level conversation starter rather than a quarterly average.
Forecast accuracy data doesn't replace the coaching conversation — it changes its quality. Moving from "I think you might be too optimistic" to "here's three quarters of data showing your commit conversion rate and where it diverges from the team" gives both the manager and the rep a shared starting point that's harder to dispute and easier to build from.