Construire une méthodologie de forecast que votre CFO respectera vraiment

The CFO does not distrust your forecast because the number is wrong. They distrust it because they cannot see how it was built. The RevOps job is to make forecast construction transparent, defensible and reproducible. Here is the methodology we install with mid-market B2B CROs to earn CFO trust in one quarter.

The 3-lens forecast

A trustworthy forecast is not a single number. It is three numbers with different underlying assumptions, and a written narrative on what would need to change to move each.

1. Commit: what the rep would bet their compensation on. Historical accuracy: 92% within 5%.
2. Most Likely: the balanced expectation. Historical accuracy: 85% within 10%.
3. Best Case: everything reasonable goes right. Historical accuracy: 65% within 15%.

Each is a Forecast Category in Salesforce, and each is tracked separately over rolling quarters to build calibration.

The 5-input construction

Input 1: Rep-call forecast

Bottom-up, opportunity by opportunity, categorized by rep into Commit, Most Likely, Best Case. Never accept "the number" without the opps behind it.

Input 2: Weighted pipeline

Amount times stage-conversion-rate, using your last 4 quarters of Opportunity History. This is the objective counter-check on rep optimism.

Input 3: Einstein Forecasting

Systemic AI overlay on Opportunity data. Best used as a triangulation input, not the primary forecast, above 20 reps.

Input 4: Manager adjustment

Explicit up or down adjustment by each manager, with written reason. Tracked over time to calibrate manager judgment.

Input 5: Gap-to-plan

(Weighted forecast + Best Case pipeline) minus remaining quota. A negative number triggers pipeline generation actions, not more optimism in the forecast.

The weekly forecast call agenda

60 minutes, same time every week, non-negotiable. Structure:

• Minutes 0-10: variance to last week's commit, at team and manager level.
• Minutes 10-30: deal review on all deals over median size in Commit and Most Likely categories.
• Minutes 30-45: gap-to-plan review and pipeline generation triggers.
• Minutes 45-55: emerging risks (churn signals, competitive threats, deal slippage patterns).
• Minutes 55-60: written commit for next week, distributed within 24 hours.

What the CFO actually wants to see

Report	Cadence	What it answers
Commit vs Actual variance, rolling 4 quarters	Monthly	Are we becoming more accurate?
Rep-call vs Einstein variance	Weekly	Where is systemic optimism?
Manager adjustment accuracy	Quarterly	Which managers exercise good judgment?
Deal slippage by rep and by stage	Monthly	Where does the forecast break?

See our executive dashboards guide for the exact Salesforce report layouts.

Three CFO-trust mistakes to avoid

1. Reforecasting mid-quarter without narrative: if the number changed, the reason has to be written down.
2. Adjustments without an audit trail: manager overrides are fine, hidden overrides are not.
3. Skipping the deal review to hit the meeting time: the forecast is the deal review, not a spreadsheet.

Result: 90-day CFO trust arc

Weeks 1 to 4: install the methodology, weekly call, and rolling variance report. Weeks 5 to 8: publish variance narrative every Friday. Weeks 9 to 13: CFO stops second-guessing weekly, moves to quarterly deep review. See our forecast accuracy playbook for the full system.

Frequently asked questions

What if the CRO overrides the forecast?

Encouraged, if in writing with reason. Track override accuracy as a leading indicator of CRO judgment quality over time.

How do we handle expansion revenue?

Model separately with its own three lenses, then consolidate at the top level. Mixing new-logo and expansion in the same forecast hides both signals.

Does the CFO need Salesforce access?

Read-only on the executive dashboard, yes. Deep access is unnecessary and creates friction. The dashboard is the interface.

How do we transition from spreadsheet to Salesforce forecast?

Read the specific playbook, from spreadsheet to Einstein Forecasting.