Cash Flow Forecasting for Startups, a Step-by-Step 13-Week Model

A 13-week cash flow forecast lists every expected cash receipt and payment week by week for the next quarter, starting from today's bank balance. It shows the lowest point your cash will reach, when that happens, and what you can still change.

Most startups track profit and loss every month, look at the bank balance every morning, and still get surprised by cash. That's not a discipline problem. A monthly P&L simply isn't built to answer the question a founder actually asks on a Tuesday afternoon: can we make payroll on the 28th if the big customer pays two weeks late? The 13-week model exists to answer exactly that question, and it's one of the oldest tools in finance for a reason.

Why thirteen weeks and not twelve months

Thirteen weeks is one quarter. It's long enough to cover a full payroll cycle several times over, a quarterly tax payment, an annual software renewal or two, and the slow-paying customers who always stretch to day 60. It's short enough that you can actually know most of the numbers instead of guessing them.

That's the core trade-off. A twelve-month forecast is a planning tool: it runs on assumptions about growth rates, hiring plans and churn. A 13-week forecast is a control tool: it runs on invoices you've sent, bills you've received, payroll you've already committed to and contracts that renew on known dates. You need both, and they answer different questions.

• The 12 to 18-month model answers "how long is our runway and when should we raise".
• The 13-week model answers "what will the bank balance be every Friday until the end of the quarter, and which week is the tight one".

Lenders and restructuring advisers have used the weekly format for decades because it's hard to fool. You can argue about an annual growth assumption. It's much harder to argue about whether a specific invoice is due on a specific Friday.

Direct method versus indirect method

There are two ways to build a cash forecast. The indirect method starts from net income and adjusts for non-cash items and changes in working capital. It's what you see on a formal cash flow statement, and it's fine for monthly reporting.

The 13-week forecast uses the direct method. You list actual cash in and actual cash out, line by line. Receipts from customers. Payroll. Rent. Each vendor that matters. Loan payments. Taxes. This is what makes it useful for decisions: when the low point shows up in week 9, you can see exactly which payments land that week and which receipts might move.

How to build the model step by step

Step 1, start from cash you can actually see

Your opening balance is the cash in your operating accounts today, not your accounting balance and not what you expect after a pending transfer. If you hold cash in several accounts or currencies, list them separately on a small table above the forecast. Exclude anything you can't spend this quarter, such as a deposit held as collateral.

Step 2, map receipts customer by customer

Pull your open accounts receivable and put each material invoice in the week you realistically expect the money, not the week it's due. Look at how each customer actually paid over the last few months. If a customer always pays 15 days late, forecast 15 days late.

Then add receipts that aren't invoiced yet: subscription renewals billed on a known date, card payments that settle within days, milestone payments in signed contracts. Be stingy with new sales. A pipeline deal that might close in week 5 shouldn't show up as cash in week 6 unless it has a very good track record behind it. Keep it on a separate "upside" line so it doesn't quietly inflate the base case.

Step 3, list every committed outflow

Payroll is usually the largest and the easiest to predict, so start there. Use the actual pay dates and include employer taxes and benefits in the same week. Then go through accounts payable and add each bill in the week you intend to pay it. After that, sweep the recurring items people forget: annual software renewals, insurance premiums, quarterly estimated taxes, sales tax remittances, loan interest, and credit card statements that settle many small expenses at once.

A good habit is to export the last 90 days of bank transactions and check every recurring payment against your list. You'll almost always find a few that nobody remembered.

Step 4, calculate closing cash and the minimum balance

For each week, closing cash equals opening cash plus receipts minus payments, and that closing number becomes next week's opening cash. Add one more row: the lowest closing balance across all thirteen weeks and the week it happens. That single number is what you'll talk about most.

Set a floor. Many companies pick a minimum cash buffer equal to one or two payroll cycles. If the forecast dips below that floor in any week, that's your early warning, and you have weeks rather than days to act.

Step 5, roll it forward every week

A 13-week forecast is only useful if it's alive. Every week, drop the week that just ended, add a new week 13, and replace the forecast numbers for the past week with what actually happened. Then compare: did customer receipts come in as expected? Did a payment go out early? That weekly variance check is where you learn how good your forecasting assumptions really are.

Most finance teams pick a fixed time, such as Monday morning after the bank feeds update, and keep the review to thirty minutes. Consistency matters more than precision.

Common mistakes that make the forecast lie

• Using invoice due dates instead of payment behavior. Customers pay when they pay. Forecast based on history.
• Mixing accrual numbers into a cash model. Revenue recognized this month isn't cash this month. Annual contracts paid upfront are the classic trap in both directions.
• Forgetting timing inside the month. Payroll on the 15th and the 30th, rent on the 1st, a big customer paying on the 25th: the monthly total might look fine while one week goes negative.
• Treating the pipeline as cash. Put unsigned deals on a separate line and keep them out of the base case.
• Never comparing forecast to actual. Without the weekly comparison you can't tell whether the model is improving or drifting.

Turning the forecast into decisions

Here's where the model pays for itself. Once the tight week is visible, you've got levers, and most of them are only available if you pull them early.

• Collections. Call the two or three customers whose payments matter most for the low week. A polite reminder ten days early works far better than a chase after the due date.
• Payment timing. Move discretionary vendor payments out of the tight week, within your agreed terms, and keep your critical suppliers paid on time.
• Annual prepayments. Offering a discount for annual upfront billing can move a lot of cash into the current quarter. It's a real trade-off on revenue, so model it before offering it.
• Hiring dates. Shifting a start date by a few weeks changes the weekly profile without changing the plan. The headcount planning view is the right place to see the full cost of each hire, including taxes and benefits.
• Financing. A credit line is much easier to arrange when you don't urgently need it. The 13-week view gives you the lead time to talk to a lender calmly.

It also helps to run two or three versions side by side. A base case, a case where your largest customer pays a month late, and a case where a key renewal slips. Those scenario planning comparisons turn "we might be fine" into a specific number of weeks of cushion.

How the 13-week model connects to runway

The weekly forecast and the monthly runway model should agree with each other. The first quarter of your 18-month model should roughly match the sum of the 13 weeks. When they don't, one of them is wrong, and finding out which one is a useful exercise in itself.

Runway, the number of months until cash runs out at the current burn, comes from the longer model. If you just want a quick read on that, a runway calculator gives you the months, the cash-out date and the break-even month from a handful of inputs. The 13-week forecast then tells you whether the next quarter inside that runway is smooth or bumpy.

Spreadsheet or software

You can absolutely build a 13-week forecast in a spreadsheet, and many companies start that way. The problem isn't building it. It's maintaining it: exporting bank transactions every week, re-keying receivables and payables, fixing broken formulas when someone inserts a row, and reconciling the weekly model with the monthly one.

Dedicated cash flow forecasting software connects to your accounting system and bank feed read-only, rolls the forecast forward automatically, pulls actuals in as they post, and flags the gap between forecast and actual for each line. Budgeter does this and connects the weekly view to your budget, headcount plan and scenarios, so a change in a hiring date shows up in the cash forecast immediately. You can see the plans and what each includes on the pricing page.

Whichever route you take, the habit matters more than the tool. A rough forecast reviewed every Monday beats a perfect one that was last updated in March.

A simple weekly routine

1. Update opening cash from the bank on the same day each week.
2. Replace last week's forecast with actuals and note the three biggest differences.
3. Update receivables and payables for new invoices and bills.
4. Add the new week 13.
5. Check the minimum balance and the week it happens against your cash floor.
6. Decide on one or two actions if the low point moved the wrong way.

Thirty minutes, once a week. That's the whole system, and it's the closest thing to an early warning radar a startup has.