A template for a B2B SaaS sales forecasting model should build ARR projections from three components: new business pipeline (stage-weighted), expansion ARR from existing accounts, and gross churn — because a forecast that only models new business will systematically overestimate ARR by 15–25% in mid-market and enterprise SaaS where expansion and churn are material.
Most early-stage SaaS teams build revenue forecasts as "deals expected to close × ACV." This works at Seed. It fails at Series A+ when existing customers represent 30%+ of ARR and expansion becomes a material growth lever.
Model Structure
Component 1: New Business Pipeline Forecast
Stage-weighted pipeline model:
| Pipeline Stage | Close Probability | Average Sales Cycle | |---|---|---| | Marketing Qualified Lead (MQL) | 5–8% | 90+ days | | Sales Qualified Lead (SQL) | 15–20% | 60–90 days | | Discovery Completed | 25–35% | 45–60 days | | Demo Completed | 40–50% | 30–45 days | | Proposal Sent | 60–70% | 14–21 days | | Verbal Commit | 85–90% | 5–10 days | | Contract Sent | 95% | 2–5 days |
Quarterly new business forecast formula: Sum(ACV × Stage Probability) for all opportunities expected to close in the quarter.
Calibration: After each quarter, compare forecasted vs. actual closed ARR per stage. If your Discovery Completed stage consistently closes at 20% (not 30%), update your stage probability.
According to Lenny Rachitsky's writing on B2B SaaS metrics, the most common forecasting error is using industry-average stage probabilities rather than your own historical win rates — your win rate from demo to close may be 35% or 55% depending on your competitive position, and using the wrong number creates systematic forecast error.
Component 2: Expansion ARR Forecast
Expansion ARR sources:
- Seat expansion (existing accounts adding users)
- Tier upgrades (accounts moving to a higher plan)
- Usage overages (for usage-based pricing)
- New use case expansion (second team or department)
Expansion ARR formula: Beginning-of-period ARR × Expected NRR (Net Revenue Retention) - Beginning ARR = Expansion ARR
If your NRR is 115% and beginning ARR is $5M, you expect $5M × 1.15 = $5.75M at year-end from existing accounts, adding $750K in expansion ARR before any new business.
Component 3: Gross Churn Forecast
Churn calculation: Gross churn = ARR at beginning of period × Expected churn rate
Churn rate benchmarks by segment:
| Segment | Annual Gross Churn Benchmark | |---|---| | SMB (<10 seats) | 10–15% | | Mid-market (10–100 seats) | 5–8% | | Enterprise (100+ seats) | 2–5% |
Leading indicators of churn (to improve churn forecast accuracy):
- Product health scores at renewal date (accounts below 60/100 health score churn at 4x base rate)
- Support ticket volume in last 90 days (accounts with 3+ escalated tickets in 90 days churn at 3x base rate)
- Last login date (accounts with no login in 30 days churn at 5x base rate)
ARR Bridge Model
Beginning ARR: $X
+ New Business ARR: $Y [Stage-weighted pipeline × close rates]
+ Expansion ARR: $Z [NRR model from existing accounts]
- Gross Churn: -$W [Historical churn rate × renewal ARR at risk]
= Ending ARR: $X+Y+Z-W
According to Shreyas Doshi on Lenny's Podcast, the ARR bridge model is the single most important artifact that B2B SaaS CEOs should review monthly — it separates growth into its three levers (new business, expansion, retention) and shows immediately which lever is dragging or accelerating growth.
Confidence Intervals
Best-practice forecasting reports three numbers per quarter, not one:
| Scenario | Description | Use | |---|---|---| | Commit | 90%+ confidence deals only | Board committed number | | Best Case | All commit + high-probability pipeline | Upside scenario | | Worst Case | Commit minus 10% slip risk | Downside planning |
FAQ
Q: What should a B2B SaaS sales forecasting model template include? A: Three components: stage-weighted new business pipeline forecast, expansion ARR forecast based on NRR from existing accounts, and gross churn forecast based on health scores and historical churn rates, combined in an ARR bridge model.
Q: What are typical pipeline stage close probabilities for B2B SaaS? A: MQL at 5-8%, SQL at 15-20%, Discovery Completed at 25-35%, Demo Completed at 40-50%, Proposal Sent at 60-70%, Verbal Commit at 85-90%, Contract Sent at 95%. Calibrate with your own historical win rates — industry averages create systematic forecast error.
Q: What is an ARR bridge model? A: A model that decomposes ARR growth into beginning ARR plus new business ARR plus expansion ARR minus gross churn, showing the three levers (new business, expansion, retention) separately so you can identify which is driving or dragging growth.
Q: What are the leading indicators of churn in a B2B SaaS forecasting model? A: Product health score below 60 (4x base churn rate), 3 or more escalated support tickets in 90 days (3x base churn rate), and no login in 30 days (5x base churn rate). These leading indicators improve churn forecast accuracy 30-60 days before the renewal date.
Q: How do you calculate expansion ARR in a B2B SaaS forecast? A: Multiply beginning-of-period ARR by your expected NRR rate. If NRR is 115% on $5M ARR, you expect $5.75M from existing accounts at year-end, generating $750K in expansion ARR before any new business contribution.
HowTo: Build a B2B SaaS Sales Forecasting Model
- Establish historical win rates for each pipeline stage from your CRM data rather than using industry averages, updating them quarterly as your competitive position and sales team evolve
- Build a stage-weighted new business pipeline forecast by multiplying each open opportunity's ACV by the close probability for its current pipeline stage
- Model expansion ARR using your trailing 4-quarter NRR applied to beginning-of-period ARR, segmented by account size and cohort
- Model gross churn using your historical churn rate by segment plus a risk adjustment for accounts with health scores below 60, 3 or more escalated tickets, or no login in 30 days
- Combine the three components in an ARR bridge and report three scenarios each quarter: Commit (90 percent confidence), Best Case (high-probability pipeline included), and Worst Case (10 percent slip discount applied)