Forecasting Guide
Learn how forecasting works, the available models, and how to configure them.
By Sebastian StreiffertPublished Jan 10, 2026Updated May 29, 20265 min read
Why Forecasting Matters
Revenue forecasting isn't just about predicting numbers. It's about making better decisions. When you know what's likely to close and when, you can staff appropriately, manage cash flow, and spot pipeline problems before they become quarterly disasters.
The challenge? Most forecasts are either wildly optimistic (reps inflate their confidence) or so conservative they're useless for planning. Lumenbase addresses this by combining your deal data with historical performance to generate calibrated, risk-adjusted projections.
Multiple Models
Historical Calibration
Risk Adjustments
Velocity Tracking
Choosing Your Forecast Model
Lumenbase offers two forecasting models. Choosing the right one depends on how your business generates revenue.
Dynamic (Deal-Driven) Model
Best for transactional businesses where most revenue flows through discrete deals such as software sales, consulting engagements, or equipment purchases. If your revenue is primarily "new business" rather than recurring subscriptions, this is your model.
Every open deal contributes to the forecast based on its value, stage probability, and health status. Deals with overdue tasks or stale activity are automatically discounted. See Deals guidefor details on deal health tracking.
Inertia + Delta Model
Built for subscription and recurring revenue businesses. Instead of starting from zero each period, this model begins with your baseline recurring revenue and adds (or subtracts) expected changes.
"Inertia" is your expected recurring revenue based on historical invoicing patterns. "Delta" captures new deals closing, churn, and expansion. This model shines when most of your revenue is predictable renewals.
Understanding Calibration
Here's an uncomfortable truth: sales reps are optimistic. That $100k deal at 80% probability? History might show your team closes only 35% of deals at that stage. Historical calibration adjusts for this gap.
If your actual win rate is 30%, the calibration factor becomes 0.6, effectively reducing forecast values by 40%. If you're closing 70% of deals, the factor rises to 1.4, boosting projections. This prevents both over-optimism and excessive pessimism.
Calibration Requirements
- Minimum 5 closed deals in the lookback period
- Lookback period configurable: 6, 12, or 24 months
- Can be disabled entirely if you prefer raw probabilities
- Updates automatically as new deals close
Deal Health and Risk Adjustments
Not all open deals are equally likely to close. A deal with recent activity, engaged stakeholders, and a stable close date deserves more forecast weight than one that's been sitting untouched for three weeks with the close date pushed twice.
Health Status Multipliers
| Status | Multiplier | Effect |
|---|---|---|
| Good | 1.0 (100%) | Full weighted value included |
| Warning | 0.85 (85%) | 15% reduction applied |
| At Risk | 0.6 (60%) | 40% reduction applied |
What Triggers Health Warnings
| Factor | Warning Threshold | Critical Threshold |
|---|---|---|
| Staleness (no activity) | 7 days | 14 days |
| Close Date Pushes | 2 times | 4 times |
| Overdue Tasks | 1 task | 3+ tasks |
| Low Engagement | Below average | Minimal activity |
Each factor has a configurable weight determining its impact on overall health. If close date changes are a bigger red flag in your business than task completion, adjust the weights accordingly.
Sales Velocity Adjustments
Velocity measures how quickly deals progress through your pipeline compared to historical averages. A deal moving faster than typical is probably more likely to close, and close sooner. One that's stalling? Less optimistic.
| Velocity Status | Adjustment | Criteria |
|---|---|---|
| Accelerating | +5% boost | More stages faster than average |
| Stable | No change | Mixed or average velocity |
| Slowing | -10% penalty | More stages slower than average |
Velocity adjustments are applied after deal health multipliers. A healthy, fast-moving deal gets the full boost. An at-risk, slowing deal compounds its penalties.
Revenue at Risk (RAR)
Revenue at Risk captures potential losses that don't show up in your deal pipeline, such as contract non-renewals, service reductions, and at-risk accounts that might churn. These entries create explicit downside adjustments to your forecast.
RAR Behavior by Model
| Model | RAR Effect |
|---|---|
| Dynamic (confidence mode) | Widens confidence bands only |
| Dynamic (direct mode) | Deducts from forecast (capped at 50%) |
| Inertia + Delta | Always deducts from forecast |
Even if you're not sure a customer will churn, logging it as RAR with a 30% likelihood creates appropriate forecast conservatism. Better to plan for potential losses than be blindsided.
Confidence Bands and Scenarios
Single-point forecasts create false precision. A $500k forecast could realistically land anywhere from $400k to $600k depending on which deals close and which slip. Confidence bands acknowledge this uncertainty.
Confidence Level Determination
| Level | Criteria |
|---|---|
| High | >50% healthy deals, win rate ≥40%, no RAR entries |
| Medium | Default when neither extreme applies |
| Low | >30% at-risk deals, OR win rate <20%, OR 3+ RAR entries |
Scenario Planning
Three built-in scenarios help with planning across different risk appetites:
| Scenario | Adjustment | Use Case |
|---|---|---|
| Optimistic | +10% | Aggressive hiring or investment decisions |
| Baseline | 0% | Standard operating plan |
| Conservative | -15% | Cash flow planning, runway calculations |
Practical Configuration Tips
For New Teams
- Start with the Dynamic model: it's simpler and requires less historical data
- Disable calibration until you have 20+ closed deals
- Use default health weights initially; tune based on your deal patterns
- Focus on getting close dates and probabilities accurate before worrying about advanced settings
For Mature Teams
- Enable calibration with 12-month lookback for balanced historical weighting
- Adjust health weights based on which factors actually predict losses in your business
- Consider Inertia + Delta if recurring revenue exceeds 50% of total
- Use RAR proactively: log any account showing warning signs
Interpreting Your Forecast
A forecast is a tool, not a truth. The weighted forecast is your most likely outcome. The risk-adjusted version accounts for deal health. The confidence bands show the realistic range.
- Base Pipeline: What you'd close if every deal came in, which almost never happens
- Weighted Forecast: Probability-adjusted expectation
- Risk-Adjusted: Weighted forecast minus health and staleness penalties
- Conservative Scenario: Use this for financial planning and commitments
If your risk-adjusted forecast is significantly lower than your weighted forecast, that's a signal. You have deals that look good on paper but show warning signs. Dig into those specific opportunities.
Was this article helpful?