Why

An org-level forecast hides who is growing. Team-level forecasts allow targeted action. A VP who sees “cloud will cost $5.6M next quarter” cannot act; a VP who sees “your team’s RDS spend is projected to grow 18% due to Data Lake Phase 2” can. Integration with business planning prevents cloud spend from existing as a disconnected number.

What

Produce the forecast at team/BU level, update it at least monthly, measure forecast accuracy, and feed the cloud forecast into the broader business financial planning cycle.

How

Set Granularity

Forecast at the level you allocate costs. If cost allocation is by BU and service, forecast at that level. Produce roll-ups for executive reporting.

Establish Cadence

Cadence	Activity	Output To
Monthly	Refresh forecast with latest actuals. Incorporate planning changes.	Finance, Eng leads, exec dashboard
Quarterly	Full rebuild. Align with demand intake. Recalibrate baseline.	Budget setting, commitment purchases, board reporting
Continuous	Monitor Infracost PR estimates vs forecast. Flag if infra changes outpace plan.	Early warning system

Track Forecast Accuracy

Forecast Accuracy = 1 − |Actual − Forecast| / Actual

Accuracy Range	What It Means
95–100%	Excellent. Spend is stable or planning inputs are strong
85–95%	Good. Normal for active roadmaps
70–85%	Needs work. Likely missing planning inputs
Below 70%	Forecasting is not functioning. Rebuild.

Every month, answer three questions: What was the variance? Why? What changes?

Integrate with Business Planning

Feed the cloud forecast into the broader financial planning cycle. The cloud number should appear alongside other IT spend in the CFO’s planning model, not exist as a separate, disconnected artefact.

Deliverable Checklist

Forecast produced at team/BU level (not just org-wide)
Monthly refresh cadence operational
Quarterly full rebuild aligned with demand intake
Forecast accuracy measured and tracked monthly
Cloud forecast integrated into business financial planning