A founder I work with once opened his banking app in the middle of a board meeting, stared at the balance for a solid ten seconds, and said: "I have no idea if we can make payroll in seven weeks."
His company was doing fine. Revenue was up. He had eleven people, two big clients, a product people actually liked. And he was flying blind on the one number that decides whether any of that matters. That moment is why startup cash flow forecasting is the skill I push hardest on founders — harder than fundraising decks, harder than growth experiments.
Because a cash forecast isn't a finance artifact. It's the difference between steering and hoping.
Key Takeaways
- You don't need a finance degree to build a working cash forecast — you need a weekly grid and honest assumptions.
- A 13-week rolling forecast is the workhorse; a 12-month view is what you show investors and use for planning hires.
- Cash timing beats margin math. A profitable contract that pays in 90 days can still sink you.
- Pre-revenue companies forecast from assumptions, not history — sales cycles and payment delays are the whole game.
- Update the forecast every week. A forecast you built once and never touched is worse than no forecast, because you'll trust it.
Why startup cash flow forecasting decides whether you survive
Here's a pattern I've watched play out too many times: a startup runs out of money not because it was unprofitable, but because its spending rhythm and its collection rhythm drifted apart.
The mechanism is almost boring. You hire ahead of revenue. You sign a lease with two months' deposit. You win a big client who pays on 60-day terms and then pays on 75 because their AP team is slow. Each of those is a reasonable decision. Stacked together, they create a cash gap measured in months.
The burn rate you think you have vs. the one you actually have
Most founders quote a burn number that's a rough average of the last three months of bank statements. I did the same thing early on with a previous project, and I was off by roughly 30%. Why? I'd excluded the quarterly software renewals, the corp tax payment, and a chunk of contractor invoices that landed in arrears.
The burn rate that matters is net cash out — every euro or dollar that leaves the account across a quarter, divided by the weeks in it. Not the tidy figure you memorized.
What runway really means
Runway is cash in the bank divided by weekly net burn. If you have €240,000 and burn €18,000 a week, you have about 13 weeks. That's it. Simple arithmetic, brutal implications.
- The number itself is easy.
- The hard part is that your burn changes the moment you hire, cut, raise prices, or lose a client.
- And the forecast is only useful if you can see the week the number crosses the danger line, not just the month.
So build the thing weekly. Daily precision is noise. Monthly is too late. Weekly hits the sweet spot where you can still act.
How to build a cash flow forecast, step by step
After a lot of trial and error — and a spreadsheet I rebuilt from scratch four times — here's the structure I now use for every company I advise.
The 13-week rolling forecast
Thirteen weeks is one quarter. It's far enough to catch trouble, close enough that your assumptions are grounded in real conversations with customers.
- Opening balance: actual cash across every account today.
- Collections: listed week by week, by customer, at the date you expect the money to land — not the invoice date. If a client pays on 60-day terms, it goes in week 9, not week 1.
- Fixed outflows: payroll, rent, subscriptions, loan repayments. These are near-certain, so they anchor the model.
- Variable outflows: contractors, ad spend, inventory, travel. Estimate by week, rounded up.
- Taxes and one-offs: VAT/sales tax remittances, quarterly filings, insurance premiums, equipment purchases. Founders forget these relentlessly.
- Closing balance: opening plus collections minus outflows. Repeat for all thirteen weeks.
The 12-month view
The yearly model runs on the same logic but at monthly granularity, and it's the one you use for hiring plans and board conversations. I keep both open side by side — the 13-week model as the operational truth, the 12-month as the strategic picture.
| Feature | 13-week forecast | 12-month forecast |
|---|---|---|
| Granularity | Weekly | Monthly |
| Primary use | Operational decisions, payroll safety | Hiring, fundraising, planning |
| Accuracy of assumptions | High — based on real invoices | Lower — trend extrapolation |
| Update cadence | Weekly | Monthly |
| Who reads it | You and your co-founder | Investors, board |
Build the weekly one first. If you nail the thirteen weeks, the yearly model writes itself.
Pre-revenue forecasting: how to model without history
Every template I ever downloaded assumed I had historical data. In month one of a new company, you have none. So what do you put in the sheet?
You put in your assumptions, and you label them as such. That means: expected number of deals per month, average contract value, sales cycle length, and payment terms. Then you stress-test each one.
- Sales cycle: if you think it's 30 days but it's actually 75, what happens to your runway?
- Payment terms: a signed contract is not cash until it clears.
- Churn: assume you lose some customers — it's the rare startup that doesn't.
- Concentration risk: if the top client is 40% of revenue, model the week they leave.
The assumptions tab is the forecast
Here's the thing: the numbers in your cells aren't the forecast. The assumptions behind those numbers are the forecast. If you change the sales cycle from 30 to 60 days and the model breaks, you've learned something far more valuable than any single cash figure — you've learned your company is fragile to a single variable.
I once built a pre-revenue forecast with a colleague where every single assumption was optimistic. We finished, looked at the runway, and it said 18 months. We rebuilt it with each assumption at the pessimistic end and got 4 months. The truth, once we started selling, landed closer to 7. That gap is the whole point of forecasting — not prediction, but preparation.
The mistakes founders keep making
These aren't theoretical. Each one is a mistake I've made, or watched someone make close enough that it stung.
Confusing cash with profit
This is the big one. Your P&L can look healthy while your bank account drains. A signed contract shows up as revenue on the income statement the moment you deliver, but the cash might arrive weeks later — or never, if the client goes quiet.
Two numbers need to live in your head separately: what you earned and what you collected. The second one pays salaries.
Forgetting taxes and fee timing
VAT, sales tax, payroll contributions, quarterly estimated taxes — these hit in lumpy waves that have nothing to do with your revenue rhythm. The first time I did a weekly forecast, I left out a quarterly tax payment of roughly €14,000 and the forecast showed a healthy balance in week 12 that simply didn't exist.
The developer/freelancer payment delay
You hire a contractor, they invoice, and the payment clears in 45 days. If your forecast assumes 30, you have a two-week gap you didn't budget for. Multiply by three contractors and it becomes the whole quarter.
How fundraising changes your forecast
A round of funding is not a magic reset. It's a big cash inflow with a big new burn attached — because you'll hire, you'll spend on growth, and your burn rate will climb. Plenty of founders raise money and then discover their runway is exactly the same as before, just with a bigger number in the account.
When you model a raise, include three scenarios in the same sheet: raise closes in month X, raise closes six weeks late, and raise doesn't close at all. That third column is the one most founders refuse to build. Build it anyway. It tells you exactly which costs you'd cut and when, which is precisely the conversation you want to have now rather than at 2 a.m. in a panic.
The catch? Fundraising timelines slip. I've seen a round that was "basically closed" in March land in June. If your forecast only works with the money in the bank, it doesn't work.
Frequently asked questions
How often should I update my cash flow forecast?
Once a week, with actual numbers from the bank account replacing the previous week's estimate. Roll the thirteen-week window forward by one week each time. The update takes ten minutes once the structure is built, and it gives you a live view of your runway rather than a snapshot from six weeks ago.
Do I need an accountant to do this?
No. Cash forecasting is arithmetic and discipline. Spreadsheets work fine, and there are plenty of lightweight tools if you'd rather not build one. An accountant helps you get the tax and collection assumptions right, but the weekly discipline is yours. The moment you outsource the forecast, you stop understanding your own runway.
What's the difference between burn rate and runway?
Burn rate is how fast you're spending — cash out minus cash in, per week or month. Runway is how long your remaining cash lasts given that burn. If you burn €15,000 a week and have €180,000, your runway is 12 weeks. They're two views of the same reality: one is a speed, the other is a timer.
The forecast you actually use
Cash forecasting is not a finance ritual. It's the habit of asking "can we survive the next three months if the worst happens?" — and having a specific, weekly answer.
So here's a test: open your spreadsheet, or your tool, and answer one question right now. Which week does my cash get uncomfortable? If you can point to a specific week on a calendar, you're already better off than the founder I described at the start. If you can't, that's the work.
And if your answer is six months out and comfortable, then look harder. The worst forecasts I've seen weren't wrong about the numbers — they were right about the numbers and blind to the assumption that changed in month four.
