CommitTrack · May 13, 2026 · 7 min read

How to Measure Forecast Accuracy (and Why 90% of AEs Have It Wrong)

Every AE is asked the same question on Thursday afternoon: How accurate is your forecast? Most answer with a percentage — their close rate for the quarter — and move on. That number is almost completely useless.

Here's why: tracking whether your deals closed sometime this quarter tells you almost nothing about whether you're actually good at forecasting. It conflates lucky timing with real predictive accuracy, hides date slippage, masks amount variance, and treats a deal that pushed into Q2 the same as a deal that never closed at all.

The AEs who earn credibility with their VPs — and whose numbers don't get adjusted every week — track something different. They measure whether what they said would happen, actually happened: at the right amount, on the right date, in the right category. There are four numbers that capture this. Here's how to calculate each one.

The metric most AEs are actually tracking

Ask an AE to define "forecast accuracy" and you'll typically hear: "What percentage of my Commit deals closed in the quarter?"

That's close rate. It is not forecast accuracy.

Consider two AEs, each with five Commit deals at $200K each ($1M total Commit):

AE A is the better forecaster. AE B is the better closer — this quarter. But if your VP is trying to commit a number to the board, AE A's pipeline is worth infinitely more. Every number AE A calls is precise. Every number AE B calls is a range with a ± attached.

Forecast accuracy and close rate measure different things. One measures execution. The other measures prediction. Both matter — but the metric that earns trust is the one almost nobody tracks.

The four metrics that actually matter

Forecasting accuracy breaks down into four distinct signals. Each one tells you something the others can't.

1. Commit-to-close hit rate

Did the deal you called as Commit actually close — at any amount, within the period? This is the binary signal. It answers: do the deals you say are going to close, close?

Track this at the deal level, not the revenue level. A $50K deal that closed and a $400K deal that closed each count as one hit. Top-quartile AEs hit above 47%. Median is closer to 31–35%. See our full benchmarks for cohort breakdowns by ARR band.

2. Amount delta %

For the deals that did close, how far was the actual amount from what you committed? This exposes discount pressure, scope creep, and negotiation patterns you might not see in CRM.

Lower is better. Top-quartile AEs land at under 8% amount delta on closed Commits. An amount delta above 20% means your Commit numbers aren't commitments — they're opening bids.

3. Slip days

For deals that closed, how many days after the committed close date did they actually close? This is the signal most ignored and most damaging to revenue planning. A deal that slips 30 days doesn't just miss the quarter — it misses the board update, the comp acceleration, and the VP's credibility with the CFO.

Measure this only on deals that closed — open slip is a separate problem. Top-quartile: under 14 days. Median: 31+ days. Anything over 45 days means committed close dates are decorative.

4. Forecast category accuracy

How often does a deal you called Commit actually close as Commit — versus slipping to Best Case or falling out entirely? This is the clearest indicator of whether your category definitions mean anything. See Article #8 for the full treatment of category misuse patterns.

If 60% of your Commits close in Best Case or later, your Commit threshold is set wrong. Target: above 70% for Commit-to-Commit close rate.

Why rolling 8-week windows beat quarterly retrospectives

The standard approach to forecast review is backward-looking: at the end of Q3, you compare what you called on July 1 against what actually closed by September 30. The problem isn't the comparison — it's the timing. By the time you have the data, you've already lived through the consequences.

A rolling 8-week window flips this. Instead of reviewing accuracy after the quarter closes, you calculate a rolling hit rate, amount delta, and slip-day average across the last 8 weeks of deals — updated every week. This creates a live signal rather than a post-mortem.

Teams using weekly pipeline velocity tracking achieve 87% forecast accuracy vs. 52% for teams that track irregularly. The gap isn't better models — it's more frequent measurement.

The math is the same. The difference is frequency. A quarterly retrospective shows you that your hit rate was 32%. A rolling 8-week window shows you that it was trending from 45% in weeks 1–4 to 28% in weeks 5–8 — which means you had a specific problem in a specific period you can actually do something about.

It also builds a credibility track record. If your VP can see 8 consecutive weeks where your Commits converted at 50%+, that track record is worth more than any verbal assertion that this quarter is "definitely solid." Evidence beats assertion. That's the entire philosophy behind CommitTrack's Forecast Checkup.

What good looks like

Based on data from enterprise AE cohorts (sourced from Gartner, Xactly, Ebsta, Klfuller, and Clari — full citations on our benchmarks page):

The 47% hit rate baseline might feel low. It is — intentionally. That number reflects what the best closers actually achieve, not what they claim during forecast calls. If you believe your Commit hit rate is above 60%, run the formula above on the last two quarters of your own deals. Bring receipts.

Most AEs who do this calculation for the first time discover they're in the 28–35% range. That's not failure — that's the baseline you need to know before you can improve.

The instrumentation problem

Here's the real reason 90% of AEs measure forecast accuracy wrong: they don't measure it at all. They think they do — but what they're looking at is CRM close rate, which is not the same thing.

The core problem is infrastructure. To calculate a real commit-to-close hit rate, you need to know:

• What category each deal was called at, at the start of the period
• The committed amount at the time of the call (not the current CRM value)
• The committed close date at the time of the call (not the current CRM date)
• What actually happened — close amount, close date, or "lost"

CRMs update fields in place. When you change a close date, the old date is gone. When you discount a deal from $240K to $190K, the previous committed amount disappears. When a Best Case becomes a Commit and then slips back, you've lost the audit trail entirely.

Spreadsheets don't help. They capture snapshots, not history. You can build a forecast in a spreadsheet. You cannot measure accuracy in one — not at the deal level, not with the four metrics above, not across a rolling window.

76% of CRM records are incomplete (Landbase, 2026). If close dates drift without updates and amounts are entered as round estimates, you literally cannot calculate forecast accuracy — because the inputs don't exist.

This is the problem the CommitTrack Forecast Accuracy Tracking dashboard was built to solve. It captures a snapshot of every deal in your Commit at the start of each week — amount, date, category — and compares it against what actually happened. Every week, automatically. No manual entry. No lost history.

The result is a rolling 8-week accuracy score broken down by all four metrics above, updated in real time, visible to both you and your VP. You stop arguing about whether your forecast is good and start showing the track record. That's the difference between "trust me" and "here's the data."

Use the deal review template to prep the deals behind your Commit, and the accuracy dashboard to prove the pattern over time. Together, they're the only two tools an enterprise AE needs to stop having their number adjusted every Thursday.