How to Build a Financial Model for Your Startup
A financial model helps you plan, raise money, and make decisions. Here's how to build one that's actually useful.
A financial model is a spreadsheet that projects your company’s future financial performance. Done well, it helps you plan, make hiring decisions, decide when to raise, and communicate with investors.
Done poorly, it’s fiction that helps no one.
Why You Need a Model
Planning and Decision Making
A model forces you to think through:
•
How will we grow?
•
When do we need to hire?
•
How much runway do we have?
•
Can we afford this initiative?
Without a model, these decisions are guesswork.
Fundraising
Investors expect a financial model. It shows you understand your business economics and have a plan.
Accountability
Comparing actual results to the model reveals whether you’re on track. Variances prompt questions and adjustments.
The Structure
A SaaS financial model typically has these sections:
Revenue Model
Key drivers:
•
Number of customers (by segment if relevant)
•
Average revenue per customer (ARPU)
•
Growth rates (new customers, churn, expansion)
Build bottom-up:
•
Start with current customers
•
Add projected new customers each month
•
Subtract churn
•
Calculate revenue from the customer base
Cost of Goods Sold (COGS)
What to include:
•
Hosting/infrastructure
•
Third-party software costs that scale with usage
•
Customer support costs (sometimes)
•
Payment processing fees
COGS should be tied to revenue or customer count.
Operating Expenses
Categories:
•
Sales & Marketing (salaries, ads, tools)
•
Research & Development (engineering salaries, tools)
•
General & Administrative (office, legal, accounting)
For each category, model:
•
Headcount by month
•
Average salary by role
•
Non-personnel costs
Hiring Plan
Central to your model:
•
What roles will you hire?
•
When will they start?
•
What will you pay them?
This drives most of your expenses.
Cash Flow
Cash in:
•
Revenue collected
•
Funding raised
Cash out:
•
Operating expenses
•
Capital expenditures
Ending cash = Beginning cash + Cash in - Cash out
Track monthly. Know your runway.
Building the Model
Step 1: Historical Data
If you have history, start there:
•
Last 12-24 months of P&L
•
Customer count over time
•
Churn rates
•
Growth rates
History grounds your projections in reality.
Step 2: Key Assumptions
Document your assumptions explicitly:
•
Customer growth rate: X% monthly
•
Churn rate: Y% monthly
•
ARPU: $Z
•
CAC: $A
•
Payroll cost per engineer: $B/month
Keep assumptions in a separate tab so they’re easy to find and change.
Step 3: Revenue Projections
Build from assumptions:
New Customers (Month N) = Customers (N-1) × Growth Rate
Churned Customers = Customers (N-1) × Churn Rate
Ending Customers = Beginning + New - Churned
Revenue = Customers × ARPU
Model separately for different segments if economics differ.
Step 4: Expense Projections
Personnel:
Headcount (Role) = # of people in that role
Total Salary = Headcount × Avg Salary
Fully Loaded Cost = Salary × 1.25-1.35 (for benefits, taxes)
Non-personnel:
Marketing Spend = Revenue × Marketing % (or fixed + variable)
Infrastructure = Customers × Cost per Customer (or tiered)
Step 5: Calculate Key Metrics
Build calculations for:
•
Gross margin
•
Operating margin
•
Burn rate
•
Runway
•
CAC, LTV, LTV:CAC
•
MRR, ARR growth
Step 6: Scenario Analysis
Build at least two scenarios:
•
Base case: Realistic assumptions
•
Downside case: Lower growth, higher churn
Some add an upside case. Don’t build something you’d be embarrassed to share with investors.
Practical Tips
Use Monthly Granularity
For the first 18-24 months, model monthly. Beyond that, quarterly is fine.
Keep It Simple
Don’t model every detail. Focus on the drivers that matter most. A simpler model that you actually update is better than a complex one that becomes stale.
Tie Headcount to Milestones
Don’t hire randomly. Hire when you hit milestones:
•
“Hire engineer #3 when we hit 100 customers”
•
“Hire first salesperson when we have repeatable inbound”
Be Conservative
Underestimate revenue, overestimate costs. You’ll be wrong, and it’s better to be wrong in a way that leaves you with cash.
Update Regularly
Review monthly:
•
Actual vs. projected
•
Adjust forward projections
•
Revise assumptions as you learn
A model that’s never updated is useless.
Version Control
Save dated versions:
•
Model_v1_Jan2025
•
Model_v2_Feb2025
You’ll want to look back at how projections evolved.
Common Mistakes
Hockey stick without justification: Growth rates of 20% monthly forever. Be realistic about what drives growth and when it slows.
Forgetting cash timing: Revenue recognized isn’t cash received. Annual contracts paid upfront vs. monthly affect cash flow.
Underestimating hiring costs: Fully loaded cost is 25-35% above salary. Recruiting fees add up.
No scenario analysis: Single-point forecasts are always wrong. Show you’ve thought about risks.
Too much precision: Projecting revenue to the dollar 36 months out is false precision. Round to thousands.
Set it and forget it: Models must be living documents.
What Investors Look For
When reviewing your model, investors check:
Reasonableness: Do growth rates make sense? Are expenses realistic?
Unit economics: Does CAC/LTV work? Does it improve over time?
Efficiency: How are you spending money? Is the burn justified?
Path to milestones: Does the model show a clear path to next funding or profitability?
Understanding: Can you explain every assumption?
Tools
Google Sheets/Excel: Standard approach. Flexible, universal.
Causal: Modern alternative with built-in assumptions and scenarios.
Cube, Mosaic: More sophisticated FP&A tools for later stages.
For early stage, Google Sheets is fine.
The Minimum Model
At minimum, your model should answer:
1.
How much revenue will we generate?
2.
How much will we spend?
3.
How much cash do we have?
4.
When do we run out?
5.
How does that change if things go wrong?
If your model answers these questions, it’s serving its purpose.
Key Takeaways
•
Build a model even if numbers are uncertain—the process creates clarity
•
Structure: Revenue, COGS, Operating Expenses, Cash Flow
•
Document assumptions explicitly and keep them in one place
•
Build bottom-up from drivers (customers × ARPU)
•
Include multiple scenarios (base case, downside)
•
Update monthly by comparing actual to projected
•
Keep it simple—focus on what matters most
•
Be conservative—wrong in the safe direction
Put your context to work.
Keep your project’s decisions, next steps and proof together as your agents work.
Explore the agent workflow →