1. Start with today, honestly
Your snapshot — MRR, customers, ARPU, growth, churn — is the base of the model. Round numbers belong in pitch decks, not in forecasts. Pull the real values from billing, not from memory.
2. Pick a realistic growth trend
Growth rarely stays flat. Stable means you compound at the current rate. Slowing assumes acquisition gets harder as you saturate. Accelerating assumes you unlock a new channel.
3. Layer planned changes
Price changes, new plans, upsell campaigns, and enterprise deals all shift the trajectory. Model them on the months they realistically ship, not the months you wish they did.
4. Account for risk
Every blocker you list — key-person dependence, market shifts, churn risk on a top customer — nudges projected growth down slightly. This is the model being honest on your behalf so the output stays usable.
5. Read the output as a roadmap
The biggest opportunity and the biggest risk are the two lines worth printing. Everything else is supporting evidence. Re-run the forecast monthly with fresh actuals and you'll quickly see which assumptions you're systematically getting wrong.
Frequently asked questions
How far ahead should I forecast a SaaS?
12 months for operating planning, 24–36 months for fundraising decks, 60 months as a directional valuation exercise. Beyond five years, assumptions dominate everything else.
How accurate is a SaaS forecast?
Month 1–3 should be within ~5%. Year 1 within ~15%. Beyond that, treat the forecast as scenarios, not predictions.
What's the difference between MRR and ARR forecast?
MRR is monthly recurring revenue; ARR is MRR × 12. Forecast MRR month-by-month — ARR falls out automatically.