Build a Downturn Scenario Plan:
The 13-Week Cash Flow Every Founder Should Run

Short answer

To stress-test a startup for a downturn, build three versions of a 13-week cash flow (base, best and worst case), then layer a severe combined shock onto the worst case: say a 20% sales drop, a 10% cost rise and customers paying 60 days late. The goal is to find your break-even week before you reach it.

Every founder has a runway number. Almost none of them know what that number becomes if sales fall a fifth and customers start paying two months late at the same time. Downturn scenario planning answers that question on paper, cheaply, before the market answers it for you. This is the exact 13-week cash-flow stress test to run: the assumptions to change, the shock to model, and the actions to pre-commit to at each trigger.

How do you stress-test a startup's finances for a downturn?

You model the same business under worse assumptions and watch what breaks first, which is usually cash, not profit. The method: take your 13-week cash-flow forecast, build three columns (base, best, worst), then push the worst case into genuinely severe territory by combining shocks rather than testing them one at a time. A single 10% sales dip is survivable. A 10% sales dip plus delayed payments plus a cost increase is where companies actually get into trouble, because shocks arrive together, not politely spaced out.

Why 13 weeks?

Thirteen weeks is one quarter: far enough to see a crisis coming, near enough that you can forecast receipts and payments with real accuracy. It's the horizon restructuring teams and CFOs reach for when cash gets tight, because it's built on actual invoice and payroll dates rather than smoothed monthly averages. Weekly granularity catches the thing monthly models hide: you can be cash-positive across a month and still miss payroll in week 6. Start from the mechanics in the 13-week cash flow forecast if you don't already run one.

The three scenarios: base, best, worst

Scenario planning fails when the cases are vibes instead of numbers. Define each by the specific assumptions you change. Keep base as your honest forecast, best as a realistic upside, and worst as a genuine stress case rather than a mild trim. Founders' base cases skew optimistic on their own, a predictable behavioural bias, so the worst case is where the real work happens.

The worst-case column is the one that matters. Notice it assumes no rescue: no round closes, no bridge appears. That is deliberate. A stress test that quietly assumes fresh capital isn't a stress test.

The severe combined shock: finding your break point

Here's the worked example the whole exercise builds toward. Take a startup with $600,000 in the bank, $200,000 in monthly revenue and $250,000 in monthly costs: a $50,000 net burn, about 12 months of runway on paper.

Now layer the combined shock. Sales fall 20%, so revenue drops to $160,000. Costs rise 10%, to $275,000. And customers stretch payment by 60 days, so two months of receipts arrive late. Net burn jumps from $50,000 toward $115,000 a month, and the payment delay punches a temporary hole worth roughly two months of sales in the early weeks, when collections briefly stop arriving while the receivables gap opens.

That 12-month runway is now closer to five months. And the tightest point isn't the end; it's the week the receivables gap is widest, maybe week 7 or 8, when a scheduled tax payment or quarterly bill lands on the same day. That week is your break point. The entire purpose of the model is to find it on a spreadsheet in a calm quarter, not in your bank balance during a bad one. (The numbers here are illustrative; the method is what transfers.)

Borrowing the regulators' method

You don't have to invent stress-testing from scratch; bank regulators have refined it for over a decade. Since the Dodd-Frank Act, the U.S. Federal Reserve has run annual supervisory stress tests that project a bank's losses and capital under hypothetical scenarios, chiefly a baseline and a "severely adverse" case featuring a deep recession and a sharp jump in unemployment toward 10% (Federal Reserve, 2026). Two principles transfer directly to a startup:

• The scenarios are hypothetical, not forecasts. The Fed isn't predicting a depression; it's checking whether the bank survives one. Your worst case isn't a prophecy either. Its job is to reveal the break point, not to be likely.
• Severity is the point. The regulator's severe scenario deliberately stacks bad conditions, because resilience only shows up under real strain. A gentle scenario tells you nothing you didn't already know.

Regulators drew the same lesson from the 2023 regional-bank failures: liquidity problems arrive fast and together (FDIC, 2024). Adapt the same shape for your own finances: one honest base case and one deliberately severe case built from combined shocks, run often enough to matter.

Downturn action triggers: what to do before you need to

A model that only produces a number is half-finished. The other half is deciding now, while calm, what you'll do when specific thresholds hit. Pre-committed triggers beat real-time panic every time.

Tie the triggers to your treasury setup so the cash is actually there when you need it. The tiered approach in startup treasury management keeps your reserve liquid, and the scenario mechanics extend naturally from scenario analysis in your financial model.

What this means for you

Run the three-case 13-week model once this quarter, before anything is wrong. Then update the worst case monthly and check it against your actual bank balance. The founders who navigate downturns aren't the ones who predicted them; they're the ones who'd already written down what they'd do at 9 months, at 6 months, at the break point. The model buys you the one thing a crisis destroys: time to decide well.