--- name: rolling-forecast description: Maintain continuously updated forecasts with monthly refreshes, rolling forward the outlook window. Use when updating forecasts with latest actuals, revising assumptions based on new information, refreshing scenario projections, tracking forecast accuracy, or presenting updated outlook to leadership. Triggers on phrases like "rolling forecast", "forecast refresh", "reforecast", "forecast vs actual", "continuous forecasting", "outlook update", "monthly forecast refresh". --- # Rolling Forecast Updates Maintain a continuously refreshed financial outlook that evolves with actual results and changing conditions. ## Workflow ### Monthly Forecast Refresh Cycle Trigger at month-end alongside the close process (typically Day 5–12 post-close): 1. **Load finalized actuals**: Pull P&L, balance sheet, cash flow from GL; import operational KPIs (headcount, customers, orders); validate against preliminary close figures. 2. **Variance analysis vs. prior forecast**: Calculate actual vs. forecast by line item; flag variances exceeding thresholds ($10K or 5%, whichever is lower); categorize as structural, timing, or one-time. 3. **Update key assumptions**: Volume/demand, pricing realization, cost inflation, headcount/productivity, market/competitive shifts — each tied to specific data sources. 4. **Refresh pipeline inputs**: Update weighted pipeline by stage; revise win rates from recent close data; adjust deal cycle assumptions; add new opportunities. 5. **Auto-recalculation**: Re-run forecast models with updated inputs; propagate changes through linked drivers; validate consistency and balance. 6. **Scenario refresh**: Update base/upside/downside cases; re-run sensitivity tables; refresh cash flow projections for each scenario. 7. **Management commentary**: Draft variance explanations for material items; highlight strategic decisions impacting outlook; flag emerging risks. 8. **Version control & distribution**: Archive prior version; tag changes; distribute to stakeholders; update executive dashboards. ### Rolling Forecast Calendar ``` MONTHLY ROLLING FOREST REFRESH — 12-DAY CYCLE =============================================== Day 1–3 (Post-Close Data Window): └─ Preliminary actuals available (Day 1–2) └─ Finance team begins preliminary variance review └─ Department heads alerted to submit commentary by Day 5 └─ Pipeline snapshot pulled from CRM Day 4–6 (Data Integration & Validation): ├─ Finalized actuals loaded into planning system ├─ Operational KPIs updated (customers, headcount, orders) ├─ Pipeline refreshed with latest deal stages and values ├─ Market intelligence reviewed (competitor moves, economic data) └─ Variance report generated (auto-flagged exceptions) Day 7–8 (Assumption Updates): ├─ FP&A analysts review flagged variances with department owners ├─ Assumptions updated: pricing, volume, costs, headcount ├─ Pipeline conversion rates recalibrated based on recent close data ├─ External inputs updated (FX rates, commodity prices, interest rates) └─ New strategic initiatives added (product launches, market entry) Day 9–10 (Recalculation & Validation): ├─ Models re-run with updated assumptions ├─ Scenario variants refreshed (base, upside, downside) ├─ Cash flow projections recalculated ├─ Balance sheet balances validated ├─ Key ratios recalculated (margins, leverage, coverage) └─ Consistency check: revenue → COGS → gross margin → opex → EBITDA → FCF Day 11 (Management Review): ├─ FP&A presents updated forecast to CFO (30-minute review) ├─ Key changes discussed: > 5% variances from prior forecast ├─ Assumption challenges addressed ├─ Commentary approved └─ Final forecast version locked Day 12 (Distribution & Archival): ├─ Forecast published to executive dashboard (Tableau/Power BI) ├─ Summary memo distributed to department heads and executives ├─ Prior forecast version archived with timestamp and version number ├─ Board deck updated if material changes (for upcoming meetings) └─ Cycle documentation completed; lessons logged for process improvement ``` ## Rolling vs. Static Budget Comparison ### Approach Comparison ``` ROLLING FORECAST vs. STATIC BUDGET ==================================== Dimension | Rolling Forecast | Static Budget -----------------------|--------------------------|------------------- Update frequency | Monthly (or weekly) | Annual (once/year) Horizon | 12–18 months forward | 12 months (annual) Assumptions | Current, data-driven | May be stale by Q2 Variance tracking | Forecast vs. Actual | Budget vs. Actual Decision relevance | High (current reality) | Decreases over time Resource intensity | Moderate (monthly effort) | High (annual cycle) Scenario flexibility | Easy to adjust | Requires formal amendment Forecast accuracy | Higher (closer to actual)| Lower over time Board reporting | Always current | "Budget" + "Forecast" Budget amendments | Not needed | Frequent mid-year Best practices: - Use static budget for annual planning and accountability - Use rolling forecast for decision-making and resource allocation - Both coexist: Budget = annual target; Forecast = current expectation - Budget variance → explains past performance - Forecast variance → predicts future performance - Most mature companies: Annual budget + quarterly forecast refresh - Highest maturity: Annual budget + monthly rolling forecast + weekly cash ``` ### Forecast Horizon and Granularity ``` FORECAST HORIZON FRAMEWORK ============================ Near-Term (0–3 months): - Granularity: Weekly or daily - Accuracy target: > 95% - Focus: Cash flow, order intake, immediate resource needs - Update frequency: Weekly (cash), monthly (P&L) - Key drivers: Committed orders, pipeline stage 4–5 deals, known costs - Use case: Working capital management, staffing, vendor payments Medium-Term (4–12 months): - Granularity: Monthly - Accuracy target: > 85% - Focus: Revenue trends, hiring plans, capex timing, headcount - Update frequency: Monthly - Key drivers: Full pipeline, product roadmap, market intelligence - Use case: Hiring decisions, product investments, budget reallocation Long-Term (13–24 months): - Granularity: Quarterly - Accuracy target: > 70% - Focus: Strategic investments, M&A, market entry, capacity planning - Update frequency: Quarterly - Key drivers: Market sizing, strategic initiatives, capital plans - Use case: Board strategy sessions, fundraising, capacity expansion Rolling window behavior: - As actuals come in, forecast rolls forward by one period - Month 1 actual becomes part of history; Month 13 added as new forecast - Accuracy naturally improves as forecast horizon shortens - Track " Month-1 forecast accuracy" as primary process metric ``` ## Variance Analysis Framework ### Variance Classification and Response ``` VARIANCE ANALYSIS WORKFLOW ============================ Step 1: Calculate Variance Variance = Actual − Forecast Variance % = (Actual − Forecast) / Forecast × 100 Materiality threshold: > $10,000 OR > 5% (whichever is lower) Step 2: Classify Variance Structural variance: - Permanent change in business dynamics - Example: Customer migration to lower-priced tier - Action: Update baseline forecast permanently Timing variance: - Revenue/cost shifted between periods - Example: Large deal closed in Q2 instead of Q1 - Action: No forecast change needed; note in commentary One-time variance: - Non-recurring event - Example: Laws settlement, insurance recovery, one-time bonus - Action: Exclude from run-rate forecast; disclose separately Assumption variance: - Original assumption proved wrong - Example: Win rate was 25% vs. 35% assumed - Action: Recalibrate assumption for remaining periods Step 3: Root Cause Analysis Revenue variances: - New logos: Too many? Too few? Why? (pipeline quality, competitive win rates) - Expansion: Upsell/cross-sell performing as planned? (product adoption, sales motion) - Churn: Higher/lower than expected? (customer satisfaction, competitive threat) - Pricing: Realization vs. list price? (discounting, promotions, mix shift) Expense variances: - Headcount: Faster/slower hiring? Higher/lower attrition? - Compensation: Merit increases above/below plan? Bonus payout variance? - Technology: License utilization? Contract renewals at different rates? - Professional services: More/less consulting needed? Project delays? - Marketing: Campaign effectiveness? Channel mix shift? CAC variance? Step 4: Update Forecast - Adjust remaining months based on new information - Recalculate scenarios - Update confidence intervals - Document all changes in assumption log Step 5: Commentary - Executive summary: 3–5 bullet points on key changes - Detailed: Line-item explanations for all material variances - Forward-looking: Impact on full-year outlook - Action items: Corrective actions identified and assigned ``` ## Forecast Accuracy Management ### Accuracy Metrics and Improvement ``` FORECAST ACCURACY MEASUREMENT ============================== Primary Metrics: MAPE (Mean Absolute Percentage Error): Formula: (1/n) × Σ(|Actual − Forecast| / Actual) × 100 Target: < 5% for revenue, < 8% for expenses, < 10% for cash flow Interpretation: Lower is better; < 5% = excellent, 5–10% = good, 10–15% = fair, > 15% = needs improvement RMSE (Root Mean Square Error): Formula: √((1/n) × Σ(Actual − Forecast)²) Use: Penalizes large errors more heavily than MAPE Target: Track trend; decreasing RMSE = improving process Bias (Directional Accuracy): Formula: (1/n) × Σ(Actual − Forecast) / Actual × 100 Positive bias: Systematically over-forecasting (conservative) Negative bias: Systematically under-forecasting (optimistic) Target: Bias within ±1% (balanced) Directional Accuracy: Formula: % of periods where forecast direction matches actual direction Target: > 80% Use: Important for strategic decisions (growing vs. declining) Accuracy by Time Horizon (typical): Month-1 forecast (issued for next month): > 95% Month-3 forecast (issued 3 months out): > 85% Month-6 forecast (issued 6 months out): > 75% Full-year forecast (issued at start of year): > 80% Improvement Strategies: 1. Break forecasts into components (new logos, expansion, churn) — forecast each separately 2. Use leading indicators (pipeline, leads, website traffic) instead of lagging (revenue) 3. Involve operational owners (sales, marketing, product) in forecast assumptions 4. Automate data feeds — reduce manual entry errors 5. Benchmark against industry peers 6. Post-mortem analysis on large misses (> 15% variance) 7. Continuous model improvement — test multiple models, select best-performing ``` ## Scenario and Sensitivity Management ### Rolling Scenario Updates ``` SCENARIO REFRESH PROCESS ========================== Upon each monthly refresh, update all three scenarios: Base Case (60% probability): - Reflects current trajectory and most likely conditions - Assumptions grounded in latest actuals and verified data - Updated pipeline, confirmed deals, known cost changes - Primary scenario for operational decisions Upside Case (20% probability): - Assumes favorable conditions materialize - Key favorable factors: * Win rates 10–15% above current average * Major pipeline deals close early * Market tailwinds (economic expansion, competitor weakness) * Product launch exceeds adoption targets - Use for: Capacity planning, opportunistic investments Downside Case (20% probability): - Assumes adverse conditions materialize - Key adverse factors: * Win rates 10–15% below current average * Key customers delayed or cancelled * Economic slowdown in key markets * Competitive pricing pressure - Use for: Stress testing, contingency planning, risk mitigation Contingency Triggers: - Revenue < 85% of base: Activate cost reduction plan (Tier 1) - Revenue < 75% of base: Activate cost reduction plan (Tier 2) + hiring freeze - Revenue < 65% of base: Activate crisis plan + Board notification - Cash runway < 6 months: Emergency fundraising plan - Cash runway < 3 months: Immediate cost cuts + emergency measures Sensitivity Analysis (updated monthly): - Revenue impact of ±10% change in: win rate, ACV, churn, expansion rate - EBITDA impact of ±10% change in: headcount, CAC, marketing spend - Cash flow impact of ±10% change in: DSO, DPO, inventory days - Present as tornado chart: variables ranked by impact magnitude ``` ## Edge Cases - **Post-closing forecast (first 10 days of month)**: - Use "preliminary forecast" with Day 1–10 actuals + projected remainder - Accuracy lower than full-month forecast (±10–15% vs. ±5%) - Critical for: Early revenue recognition, cash position, urgent decisions - Method: Daily actuals × (working days remaining / working days elapsed) - Validate against: Year-over-year daily patterns for reasonableness - **Major M&A during forecast period**: - Separate organic and inorganic components in forecast - Day 1: Add acquired company's forecast (based on acquisition agreement) - Integration costs: $5M–$50M one-time (typically 3–12% of deal value) - Synergy realization: Revenue synergies 5–15% in Year 1; cost synergies 10–25% over 2–3 years - Risk: Integration disruption may cause 2–5% revenue dip in first 2 quarters - Model: Three cases — synergies achieved on plan, half achieved, not achieved - **Economic shock events** (pandemic, financial crisis, trade war): - Immediate action: Re-run stress scenarios within 48 hours - Revenue impact: Model 20–50% decline in worst case (varies by industry) - Cost reduction: Pre-approved tiers of cuts (10%, 20%, 30% of opex) - Cash preservation: Extend payables, reduce capex, suspend dividends - Communication: Weekly cash updates to leadership; daily during crisis - Government support: Model grant/loan availability (PPP, grants, tax relief) - Reopening/recovery: Multiple recovery shapes (V, U, W, L) with probability weights - **First-time forecast for new product/line**: - No historical data — use analogous product launch data - Conservative assumption: 50–70% of marketing projections - Milestone-based: Tie revenue to go-to-market milestones (launch, first 10 customers, 100 customers) - Scenario range wide: 25–150% of base case - Monthly re-forecast: Actual adoption data rapidly replaces assumptions - Kill switch: Define go/no-go criteria at 3, 6, 12 months post-launch - **Forecast with significant one-time events**: - Separate recurring (run-rate) from non-recurring components - Examples: Large custom deal ($5M+), insurance recovery, lawsuit settlement - Impact: Can distort forecast accuracy if not isolated - Method: Report both "reported" and "adjusted" forecast - Disclosure: Material one-time events disclosed in forecast commentary - GAAP consideration: One-time items included in statutory forecast; excluded from operational forecast - **Multi-entity rolling forecast**: - All entities refresh on same 12-day cycle - Intercompany: Both sides of IC transactions must reconcile - Currency: FX impact modeled at closing rates + forward assumptions - Consolidation: Entity-level forecasts rolled up with eliminations - Timing: Allow 2 extra days for entities in different time zones - Standardization: Common templates, definitions, and assumptions across entities ## Integration Points - **ERP/GL**: NetSuite, SAP S/4HANA, Oracle ERP Cloud — actual results, trial balance, GL transactions - **Planning Platforms**: Adaptive Insights (Workday), Anaplan, Oracle Hyperion EPM — forecast models, scenario management - **CRM**: Salesforce, HubSpot — pipeline data, deal stages, weighted revenue, win rate history - **BI/Dashboards**: Tableau, Power BI, Looker — forecast vs. actual dashboards, automated distribution - **Data Warehouse**: Snowflake, BigQuery, Redshift — historical data, actuals integration, ML model training - **HRIS**: Workday HCM, BambooHR, ADP — headcount actuals, planned hires, attrition rates - **Expense Systems**: Concur, Expensify, Rippling — actual spending by category and department - **Revenue Tools**: Zuora, Chargebee, RevOps platforms — MRR/ARR actuals, churn, expansion - **Communication**: Slack, Microsoft Teams — automated alerts, forecast distribution, approval workflows - **Document Management**: SharePoint, Confluence — forecast memos, assumption logs, version history