13-Week Cash Flow Model Template: The Exact Layout, and How to Build It Today
By Vishesh Khurana — Oct 14, 2026
A 13-week cash flow model is a forecast of the cash actually moving in and out of your bank account, week by week, for the next 13 weeks. It is the standard liquidity document in a fundraising process, in a lending conversation, and in any situation where an investor or a board wants to know whether you can meet next month's payroll.
It is a different instrument from a runway model, and confusing the two is the most common mistake founders make with it. Your runway tells you how long you have. A 13-week cash flow tells you whether the money arrives before the obligation does, which is a question about timing rather than totals — and it is the question that kills companies which look solvent on a monthly report.
Why 13 weeks
Thirteen weeks is a quarter plus a week: long enough to cover every payment cycle in a normal business — a monthly payroll, a quarterly VAT or sales tax remittance, a quarterly software renewal, a quarterly board meeting — and short enough that a week-by-week forecast is genuinely defensible rather than invented. It is the horizon lenders use for covenant monitoring and the horizon advisers use in a restructuring, and it is short enough that you are forecasting collections you can name rather than guessing at a growth curve.
The rolling version is the useful one. Every Monday you add a new thirteenth week at the right of the sheet, delete the week that has just finished, and replace that week's forecast with what actually happened. Fifteen minutes a week buys you a forecast that is never more than a week stale.
The layout
The sheet is 13 columns wide, one per week, labelled with the Monday date and the week number. The rows below are where the model lives, and the order matters because each block feeds the next.
Opening cash. Week 1 is the balance across every bank account on the forecast start date. Weeks 2 to 13 take the closing balance from the week before.
Cash in. Split into lines you can actually date: customer collections from invoices already issued, cash from new bookings, financing received, tax refunds, and any asset sales. Keep collections that are already invoiced separate from collections you are hoping for, because they carry completely different probability.
Cash out. The lines that matter are payroll and payroll taxes (on their real pay dates, not as a monthly total), contractor and vendor payments, rent, software and subscriptions, marketing and paid acquisition, hosting and cost of goods, taxes due, debt service, capital expenditure, and one-off items such as a legal bill or a signing bonus.
Net movement. Total cash in minus total cash out for the week.
Closing cash. Opening cash plus net movement. This is the only row a lender reads first, and the number that becomes next week's opening.
Add two lines at the bottom for committed undrawn facilities and closing cash including the facility. The gap between those two rows is the thing you are trying to keep from becoming the thing you have to explain.
That is the whole structure. It is deliberately unglamorous: no discounting, no growth model, no ARR. Every cell is a cash amount with a date attached, which is why it survives contact with a lender.
The four inputs that decide whether it is right
Starting cash is the bank balance, not the cash line on your balance sheet. Include every account, including the ones you forget about and any payment processor balance not yet swept. If a cheque you wrote on Friday has not cleared, it has still left the business.
Collection timing is where forecasts go wrong. An invoice dated the 1st on 30-day terms, from a customer who pays in 45 days, is cash in week 6 or 7 rather than week 4. Use your actual payment history by customer rather than your stated terms, because terms describe what should happen and history describes what does.
Payroll goes on its real dates. A monthly payroll of $180,000 that lands in the third week of the month is not $45,000 a week. Getting this wrong is how founders discover a problem two weeks later than they could have.
Separate committed from hoped. A signed contract is not cash. A verbal commitment is not cash. An invoice is not cash. Put the things you control in the main block and the things you are hoping for on a clearly labelled line of their own, so you can see the shape of the business without them.
Building it in half an hour
Export three months of bank transactions as CSV and bucket them into recurring categories. You are not building a chart of accounts; you are finding the fifteen-odd lines that account for nearly all of the money.
List your known dated cash events for the next 13 weeks: every payroll date, every rent date, every tax remittance, every subscription renewal, and every invoice you have already issued with an expected payment date.
Enter them into their weeks, then write the closing formula once and copy it across all 13 columns. Opening plus net movement, every week, nothing clever.
Sanity-check the shape, not the total. Does the closing balance go low enough to matter, and in which week? That week is the finding, and it is usually a collection you assumed would arrive earlier.
Then set a recurring 30-minute slot on the same weekday every week and roll the window. The model is only worth having if it is current; a 13-week forecast that is three weeks old is a history lesson.
The mistakes that make one useless
Smoothing the cash in. Spreading expected collections evenly across weeks to make the line look tidy hides exactly the week you needed to see.
Forgetting the taxes that are not income tax. VAT, sales tax and payroll taxes sit in your account for weeks looking like available cash until the remittance date arrives.
Treating an annual renewal as a monthly cost. A $40,000 annual contract is a single week's outflow, and a model that spreads it will miss that week by $37,000.
Ignoring the customer who is always late. If your largest customer pays in 60 days rather than 30, your whole forecast is wrong by one customer's invoice — and the effect compounds, because you spend against collections you have not received.
Calling a recurring cost a one-off. A cost that appears in three consecutive quarters is a run-rate cost, and it belongs in the recurring block.
Updating the forecast to match reality. Rolling the window forward is right; editing last month's forecast so the variance disappears is how you lose the ability to trust your own numbers.
What it does not tell you
A 13-week cash flow model will not tell you whether to hire, when to raise, or how long you have. It is a liquidity instrument with a three-month horizon and it is silent on everything past week 13. Use it to answer whether the cash lands before the obligation; use a runway model that produces a distribution rather than a single line to answer how long you can keep going and what happens if the raise slips a quarter.
That division of labour is worth being explicit about with your board, because the two documents are asked for by different people for different reasons and are routinely confused in the same meeting.
Doing it without the spreadsheet
The spreadsheet version works, and there is nothing wrong with it for the first year. It stops working when the numbers need refreshing more often than you want to re-key them, and when the question shifts from whether payroll clears in week 9 to what the next two quarters look like if revenue grows 6% a month instead of 3% — which is a distribution question, not a spreadsheet one.
FounderConsole connects your accounting, payment and banking data so the cash position is current rather than hand-entered, and models decisions directly: delay a hire by a month, cut paid acquisition by 40%, and see what happens to the runway distribution. Metrics we cannot compute honestly from your data are marked as estimates or shown as N/A rather than filled in with a plausible-looking default.
To see the shape of the product without an account, the free runway calculator needs no signup, and the default-alive check answers whether you reach profitability on the cash you have. And if you mainly wanted the layout above, you already have it — paste it into a sheet and start.