Financial Model

A financial model is a spreadsheet or planning system that projects a company's revenue, expenses, cash flow, headcount, and key metrics into the future. The model serves as the operating planning document (drives hiring decisions, budget allocation, runway analysis), the capital-raising document (investors review it in diligence), and the board reporting document (actuals are compared against it each month). Model quality is a meaningful signal about how rigorously the company runs its operations. It is the document where many decisions ultimately get tested before they're made.

The core components of a financial model:

Revenue model:

• Revenue drivers: what specifically generates revenue? Number of customers x price, or volume x rate, or seats x ARPU.
• Customer acquisition: new customers per period, by channel/segment.
• Customer retention/churn: what percentage of customers leave each period?
• Expansion: do existing customers grow over time? Net revenue retention.
• Pricing assumptions: average revenue per customer/user, pricing changes over time.

Expense model:

• Headcount: each role's hire date, salary, benefits multiplier, tied to hiring plan.
• Variable costs: cost of goods sold (COGS), payment processing, hosting, etc.
• Marketing/sales spend: CAC assumptions, growth investment.
• G&A: rent, software, legal, accounting, miscellaneous.

Output statements:

• P&L (income statement): revenue, costs, gross margin, operating expenses, EBITDA, net income.
• Cash flow: cash in, cash out, ending cash position.
• Balance sheet (sometimes): assets, liabilities, equity. Less critical at early-stage startups.
• Key metrics: burn rate, runway, ARR, customer count, gross margin, CAC, LTV.

Scenarios:

• Base case: most-likely outcome based on current trajectory.
• Upside case: things go better than expected.
• Downside case: things go worse than expected. Critical for understanding risk.

The quality bar for financial models:

Good models are:

• Driver-based: connected via formulas to underlying drivers (e.g., revenue = customers x ARPU, not hard-coded).
• Bottoms-up: built from specific assumptions, not top-down ("we'll grow 200% per year because that's what tech companies do").
• Internally consistent: changes to one assumption flow through to all dependent calculations.
• Sensitivity-able: scenarios easy to model by changing key drivers.
• Documented: assumptions and rationale recorded, not just numbers.

Bad models are:

• Hard-coded: numbers entered directly without underlying drivers. Changes require manual updates everywhere.
• Top-down: "we'll get 1% market share" with no specifics about how.
• Internally inconsistent: changing one number doesn't update dependent numbers.
• Single-scenario: only the optimistic case modeled; no sensitivity analysis.
• Undocumented: numbers without assumptions; can't understand or evaluate the model.

Common modeling tools:

• Excel/Google Sheets: the historical default. Flexible, universal, can become unwieldy at scale.
• Mosaic, Pry, Cube: purpose-built financial planning platforms. Better for scaling teams but require investment.
• Causal, Vena: modern alternatives with better scenario handling.
• In-house systems: large companies sometimes build custom planning systems.

The fundraising context:

• Investors expect to see a financial model in Series A+ diligence.
• The model's quality signals operational rigor.
• Investors typically question key assumptions; founders should have rationale ready.
• Model should be defensible but not so detailed that it implies false precision.

Ryan's Take

Financial model is one of those documents that gets dismissed as "spreadsheet work" until you realize it's actually how the company runs. The model captures every operational assumption: who you'll hire, how much you'll spend, what customers will pay, how fast you'll grow. When the model is good, decisions get made against it ("if we hire 3 more engineers, runway drops to 14 months; that's our threshold"). When the model is bad, decisions get made on vibes ("we feel good about hiring more"). The right discipline: invest in a real driver-based model at Series A or before. Connect every operating decision to the model. Update monthly with actuals vs plan. Use scenarios for stress-testing major decisions. The model is the company's operating brain made explicit; treat it that way.

What founders get wrong: Treating the financial model as a fundraising artifact rather than an operating tool, then never updating it after the round closes. The right discipline: build a driver-based model at Series A or earlier, connect every operating decision to it (hiring, marketing spend, product investment), update monthly with actuals vs plan, and use scenarios to stress-test major decisions. The model is the company's operating brain made explicit, not a deck appendix.

Related: [Financial Projections] · [Cash Flow] · [Revenue Model] · [Burn Rate] · [Runway]

FAQ

What is a financial model? A spreadsheet (or purpose-built system) projecting a company's revenue, expenses, cash flow, headcount, and key metrics into the future, typically monthly for 12-24 months and annually for 3-5 years. Used for operating planning, capital raising, and board reporting.

What should be in a financial model? Revenue model (drivers, acquisition, retention, pricing), expense model (headcount tied to hiring plan, variable costs, marketing/sales, G&A), output statements (P&L, cash flow, sometimes balance sheet), key metrics (burn, runway, ARR, gross margin), and scenarios (base, upside, downside).

What makes a good financial model? Driver-based (formulas connecting to underlying drivers, not hard-coded numbers), bottoms-up (built from specific assumptions, not top-down growth targets), internally consistent (changes flow through), sensitivity-able (scenarios easy to model), and documented (assumptions recorded with rationale).