How to Measure Forecast Accuracy (and What Good Actually Looks Like)

You cannot improve what you do not measure. Most revenue teams treat forecast accuracy as a feeling rather than a metric.

The three metrics to track

Commit accuracy

Compare commit number at the start of the period to actual closed revenue. Formula: 1 - |Commit - Actual| / Actual. Best-in-class: above 90%. Median: 70-80%.

Weighted pipeline accuracy

Multiply each deal by its stage probability. Compare to actual closed revenue. If weighted pipeline consistently overstates by 30%+, your stage probabilities are wrong.

Category movement

Track how much revenue moves between forecast categories during the quarter. Healthy teams see less than 20% of commit-level deals change category.

Setting up the measurement

Create a snapshot process that captures the forecast state at a fixed point each period. In Salesforce, use Collaborative Forecasting snapshots or build a custom snapshot object. Without snapshots, you cannot compare "what we thought" to "what happened."

Using the data

Review accuracy metrics each quarter. Look for patterns: are specific reps consistently over-forecasting? Are certain stages leaking? Is accuracy improving over time?

The Forecasting Model Simulator lets you run five models against a closed quarter to see which told the truth.

Forecast accuracy is one of four criteria in the Forecasting Trust dimension of the AI-Ready RevOps Framework.

Try it free →Forecasting Model Simulator

Frequently asked questions

How is forecast accuracy calculated?

The most common formula: 1 - |Forecast - Actual| / Actual. Track at commit, best-case, and total pipeline levels.

What is a good forecast accuracy for B2B SaaS?

Best-in-class: 90-95% at the commit level. Median: 70-80%. Below 70% means pipeline data and stage definitions need attention.

---

← All articles