Always Current Forecasts
Rolling Forecast Assistant
Driver-based rolling forecasts with scenarios, built by AI
Get the SkillHow to Upload a Claude Skill
Rolling Forecast Assistant in Action
Features List
3
Forecast Structures
9
Driver Model Types
3
Scenarios Built In
Why FP&A Teams Use This Skill
Days to Hours, Not Days to Weeks
Building a rolling forecast from scratch — scoping the drivers, setting up assumption tabs, wiring formulas, creating three scenarios, and structuring the workbook — takes days. This skill guides you through the full build in hours: scope it, provide your data, confirm assumptions, and download the 7-tab Excel model with all three scenarios working. Updates take even less time.
Driver Models Selected Per Line Item
Revenue gets the right model for the business type — Units times Price for product revenue, MRR waterfall for SaaS, Headcount times Utilisation times Rate for services, stage-weighted pipeline for deal-driven businesses. Expenses get headcount builds for salaries, benefit rate links, commission tier structures, and T&E per FTE. Nine driver templates, each applied where it fits.
Seasonality Detected Automatically
The skill calculates seasonal indices from your historical data — identifying which months and quarters run above or below the annual average, validating statistical significance, and flagging outlier periods that could distort the pattern. You confirm or adjust the detected pattern before it's applied. Forecasts reflect how your business actually behaves across the year, not a flat monthly spread.
Three Scenarios, One Assumptions Tab
Base, Upside, and Downside scenarios are built from a single Assumptions tab with differentiated driver inputs. Revenue growth 10% base, 15% upside, 5% downside. Win rate 25%, 30%, 20%. Churn 5%, 3%, 8%. Change one assumption and both alternative scenarios update. A scenario toggle in the Summary tab shows the comparison across all periods in one view.
Roll Forward in Minutes
When actuals arrive for a closed period, the skill guides you through locking those periods, replacing forecast with actuals, analysing the variance against prior forecast, updating driver assumptions based on actual performance, and extending the forecast horizon. What used to be a rebuild becomes a structured update. The model stays current without starting from scratch each cycle.
What You Get
-
Claude Custom Skill (.skill file)
The core skill file for uploading to Claude. Contains the full 7-step rolling forecast workflow, seasonality detection methodology, driver model selection guide, scenario framework, Excel workbook structure with finance-standard formatting rules, roll-forward process, and 3-statement model extension for Balance Sheet and Cash Flow.
-
Flexible Input Handling
Works with whatever data you have: 12-24 months historical actuals (used for seasonality detection and driver ratios), current year budget (used as baseline for comparison), prior forecast (rolled forward and updated), or assumptions only (builds from stated drivers). The skill notes which analyses require additional data and adapts its methodology to what's available.
-
Automatic Seasonality Detection
For each line item with sufficient history: calculates trailing 12-month average by period, computes seasonal index as Actual divided by TTM average, averages indices across available years, and normalises so indices sum to 12 (monthly) or 4 (quarterly). Validates statistical significance (CV greater than 10%), flags outlier periods, and prompts you to confirm the pattern before applying it.
-
9 Driver Model Templates
Product Revenue (Units times Price with growth and seasonality), Subscription/SaaS (MRR waterfall — beginning plus new logos minus gross churn plus expansion equals ending MRR), Services Revenue (billable headcount times utilisation times bill rate times working days), Pipeline-based Revenue (stage-weighted pipeline times win rate), Salaries (headcount build by department with hires, attrition, and merit timing), Benefits (% of salary with itemised healthcare, 401k, and payroll tax), Commissions (tiered rates by quota achievement), T&E (headcount times cost per FTE with seasonality), and Marketing (% of revenue or campaign-based fixed plan).
-
Base / Upside / Downside Scenario Framework
Three scenarios with differentiated driver assumptions across every key input. Upside typically 10-20% above base on key drivers (favorable deal close rates, lower churn, faster hiring). Downside typically 10-20% below (slower sales, higher costs, attrition). Optional probability-weighted expected value: Base 50%, Upside 25%, Downside 25%, with weights adjustable to your confidence level.
-
7-Tab Excel Workbook
Cover tab (model name, version, date, source data reference), Assumptions tab (all inputs in one place, blue font, clearly labelled by category with scenario columns for Base/Upside/Downside), Historical tab (source data for reference and seasonality calculation), Forecast Base tab (full P&L or 3-statement with Excel formulas linked to Assumptions), Forecast Upside tab, Forecast Downside tab, and Summary tab (scenario comparison across all periods, key metrics dashboard).
-
Roll Forward Process
Step-by-step guidance for updating the model each forecast cycle: lock closed periods by replacing forecast with actuals, analyse variance between prior forecast and actuals for closed periods (including a Forecast vs. Prior Forecast tracking table), update driver assumptions based on actual performance and revised outlook, extend the horizon by adding new periods, and let formulas recalculate the remaining forecast periods automatically.
-
3-Statement Extension
For full 3-statement models: Balance Sheet drivers (Accounts Receivable from Revenue times DSO over 365, Inventory from COGS times DIO over 365, Accounts Payable from COGS times DPO over 365, PP&E from opening plus CapEx minus depreciation, Debt from opening plus borrowings minus repayments), Cash Flow integration (Operating CF, Investing CF, Financing CF netting to ending cash), and Balance Sheet balancing using cash or revolver as the plug.
-
Uber Q1 2026 – Q2 2027 Sample Output
A complete 7-tab rolling forecast built from Uber's Q4 2025 earnings supplemental data (5 quarters of actuals, Q4 2024 through Q4 2025). Shows 6-quarter Base/Upside/Downside forecast with differentiated assumptions: MAPCs growth at 17% / 20% / 13%, Mobility Gross Bookings growth at 18% / 22% / 13%, Delivery Gross Bookings growth at 24% / 29% / 18%. Base case Gross Bookings: $46.8B (Q1 2026) through $59.0B (Q4 2026).
-
Two Walkthrough Videos
Video 1: How to Upload a Claude Skill — the standard installation walkthrough. Video 2: Full screen recording of the rolling forecast being built from the Uber Q4 2025 earnings data — showing assumption setup, seasonality detection, driver selection, scenario differentiation, and the completed 7-tab workbook, so viewers see exactly what their output will look like before they start.
When to Use This Skill
Building a rolling forecast for the first time
Your team is moving from an annual budget to a rolling model and needs a structured approach. The skill scopes the forecast, helps you select driver models for each line item, detects seasonality from your historical data, and builds the full workbook with three scenarios — in hours, not the week it would take starting from a blank Excel file.
Monthly reforecast after actuals close
The month-end close is done and it's time to update the rolling forecast. The skill guides you through locking closed periods with actuals, calculating the variance against prior forecast, updating driver assumptions based on actual performance, extending the horizon, and recalculating. A structured update process that takes hours instead of rebuilding from scratch.
Board presentation needs Base / Upside / Downside
The board wants to see three scenarios with clear assumption differences and a comparison table. The Assumptions tab has differentiated driver inputs per scenario, the Forecast tabs show the full P&L for each, and the Summary tab shows the scenario comparison across all periods. All three scenarios are structurally identical — the only differences are the assumption inputs, which are clearly visible side by side.
SaaS business needs an MRR waterfall forecast
You're running a subscription business and need a forecast that starts from beginning MRR, adds new logos, subtracts gross churn, adds expansion MRR, and flows to ending MRR and ARR each month. The skill selects the SaaS MRR waterfall driver model, sets up the Assumptions tab with gross churn rate, NRR, and new logo MRR inputs, and builds the workbook so changing churn assumptions flows through all three scenarios immediately.
New CFO wants to understand the forecast model
A new CFO has joined and the existing model is a black box. The skill builds a new rolling forecast with clear assumption separation — every driver in one tab, blue font on inputs, black on formulas, green on cross-sheet links, and a Cover tab documenting the model version and source data. The forecast becomes traceable: any number can be followed back to its assumption in seconds.
Nine Driver Models — One Selected Per Line Item
| Driver Model | Formula Structure | Best For | Key Assumptions |
|---|---|---|---|
| Product Revenue | Units × Avg Selling Price × Seasonal Index | Manufactured goods, e-commerce, physical products | Unit growth rate, price increase %, seasonality by month |
| Subscription / SaaS | MRR[t-1] + New MRR – Gross Churn + Expansion = MRR[t] | SaaS, recurring revenue, subscription businesses | Gross churn %, NRR %, new logo MRR per month, expansion rate |
| Services Revenue | Billable HC × Utilisation % × Bill Rate × Working Days | Consulting, staffing, professional services | Billable FTE plan, target utilisation, average bill rate |
| Pipeline-based | Qualified Pipeline × Win Rate × Avg Deal Size × Seasonal Index | Enterprise sales, deal-driven revenue, long cycles | Stage-weighted pipeline, win rate %, avg deal size, cycle length |
| Salaries | Headcount[dept] × Avg Salary × (1 + Merit) / 12 | All P&L structures — largest OpEx driver for most | Opening HC, monthly hires by dept, attrition rate, merit timing |
| Benefits | Salaries × Benefit Rate (or HC × Healthcare + Salary × 401k%) | Linked to salary forecast; all businesses | Benefit rate %, healthcare, 401k match, FICA |
| Commissions | Commissionable Revenue × Commission Rate (tiered by quota %) | Sales-driven businesses with variable comp structures | Commission rates by quota achievement tier |
| T&E | Headcount × T&E per FTE × Seasonal Index (high Q1/Q4, low summer) | Any business with employee travel, field sales | T&E per FTE per month, seasonal adjustment by quarter |
| Marketing | Revenue × Marketing % (or fixed campaign calendar spend) | Brand/awareness spend or campaign-specific allocation | Marketing as % of revenue, or fixed monthly spend by campaign |
Common Questions
What data do I need to get started?
What does 'driver-based' mean and why does it matter?
How does seasonality detection work?
What does the Uber sample output show?
Can I use this for a full 3-statement model, not just a P&L?
How long does the roll-forward update take each month?
Your Forecast Should Always Be Current. Now It Will Be.
Driver-based rolling forecasts with seasonality, three scenarios, and a roll-forward process. Updated for 2026.