Unit economics and finance
Scenarios and sensitivity
One structure, three cases. Then shock one driver at a time and see which fight is worth having.
Build one structure, three (or more) cases. The formulas stay the same. Only the drivers change.
| Case | What you change | Use |
|---|---|---|
| Base | Best honest estimate | Operating plan |
| Upside | Faster conversion, higher ARPA, lower churn | Investor ambition / stretch |
| Downside | Slower growth, hiring delayed or cut | Default-alive stress test |
Trade-off
Conservative base vs ambitious base. A base that only works if every deal closes trains the company to miss. A base so sandbagged that you always "beat plan" hides under-investment. Aim for a base you would defend in a board meeting without squirming — then show upside separately.
Walking through a tornado
Sensitivity: vary one driver at a time and watch ending cash and runway. A simple table is enough early.
Base case: ending cash in 18 months = $900k.
| Driver shock | Ending cash | Δ vs base |
|---|---|---|
| Win rate −20% relative | $420k | −$480k |
| Churn +1 pp (e.g. 2%→3% MoM) | $510k | −$390k |
| Hire plan −2 engineers | $1.35M | +$450k |
| ARPA +10% | $1.15M | +$250k |
| Paid ads −$20k/mo (with volume loss) | $1.05M | +$150k |
Here win rate and hiring dominate. That tells leadership where to hold the line: sales capacity and conversion quality before vanity channel experiments. If a COGS% shock dominated, you would staff FinOps and pricing before another AE.
Math check on churn: going from 2% to 3% monthly logo churn cuts expected lifetime from 50 months to ≈33 months (1/0.03). On the same CAC, LTV collapses by roughly a third.
Using the model
For fundraising:
- Show you know unit economics: CAC, payback, LTV inputs.
- Show use of proceeds: how many months of runway, what milestones the capital buys.
- Avoid claiming precision on year 5. Emphasize near-term drivers.
For internal decisions:
- Hire gate: does this hire keep us default alive under downside?
- Pricing change: impact on new ARR vs churn risk.
- Channel spend: payback months under base vs downside conversion.
Update monthly with actuals. A model that is not reconciled to the bank and the billing system becomes fiction within a quarter.
Common mistakes
- One scenario only.
- Collapsing operating plan and fundraising upside into one aggressive number.
- Opening the model only for fundraises. Rebuild drivers when GTM changes; otherwise refresh actuals monthly and rewrite scenarios quarterly.
Check yourself
- Win rate −20% drops ending cash by $480k; cutting two engineers adds $450k. Where do you hold the line? Conversion and hiring — those two drivers move cash the most in this tornado.
- Monthly churn goes from 2% to 3%. What happens to expected lifetime? It falls from 50 months to about 33 (
1/0.03).
Key takeaways
- One structure, three cases. Do not rebuild the sheet for each story.
- Sensitivity ranks the fights. Change one driver at a time.
- Defend a honest base; show upside separately.
- Reconcile to cash monthly or the model becomes a costume.
Further learning
Default Alive or Default Dead? (Paul Graham)Default Alive or Default Dead?paulgraham.com SaaS Metrics 2.0 (David Skok)“If you cannot measure it, you cannot improve it” – Lord Kelvin This article is a comprehensive and detailed look at the key metrics that are needed to understand and optimize a SaaS business. It is a completely updat…For Entrepreneurs
Back to the map: building a financial model.
Related
Building a startup financial model
A driver-based model: revenue drivers, cohorts, hiring plan and scenarios.
Driver-based models
A model is a machine that turns a few operational inputs into P&L, cash and runway. Drivers can be wrong. Plugs cannot be debugged.
Burn, runway and default alive
Calculating burn and runway, the default-alive test, and how to extend runway.