There is a specific moment when a business discovers it needs a thirteen week cash flow. A lender asks for one, a covenant is getting close, a large customer goes quiet, or the board realises nobody can answer how much cash there will be in six weeks.
It is the most useful single document a finance function produces in a difficult period, and it is routinely built badly.
What it is not
It is not the cash flow statement from your accounts. That is a historical reconciliation derived from profit, and it is the wrong tool entirely.
It is not a monthly forecast divided by four. Monthly buckets hide exactly the thing you are trying to see. A month can end comfortably and still contain a Tuesday where payroll, a quarterly rent and a tax payment coincide and you are over your facility by Wednesday.
It is a direct forecast of money in and money out, week by week, built from when cash actually moves.
How to build it
Start with committed outflows, because they are the most certain and the least forgiving. Payroll, with the actual dates. Tax payments, with the actual dates. Rent and other quarterly charges. Loan repayments and interest. These are known weeks in advance and they do not negotiate.
Then receipts, from the debtor ledger rather than from revenue. Take each invoice outstanding, apply the date that customer actually pays rather than the terms on the invoice, and put it in the week it will land. A customer on thirty day terms who reliably pays at fifty five days goes in at fifty five. This single discipline is the difference between a forecast that works and one that does not.
Then supplier payments, from the creditor ledger, in your payment runs rather than on due dates.
Then everything else: new sales expected to be invoiced and collected inside the window, capital expenditure, anything unusual.
The output is an opening balance, receipts, payments and a closing balance for each of the thirteen weeks, with the facility headroom shown underneath. The headroom line is what anybody reading it looks at first.
Make the bad weeks visible
The purpose is not to prove you will be fine. It is to find the weeks where you will not be, early enough to do something.
Show the low point in every week, not just the closing balance, if your receipts and payments cluster. Show the facility limit as a line on the chart. Flag the weeks where headroom falls below whatever level makes your board uncomfortable, and say so plainly rather than hoping nobody adds up the column.
A forecast that shows a problem in week nine is doing its job. You have eight weeks to chase debtors, phase a payment run, delay capital spend or have a conversation with the bank while you still have options. A forecast that shows no problem and is wrong leaves you with none of that.
Roll it weekly
This is where most attempts fail. The forecast gets built for the lender, used once, and left.
It should roll every week: drop the week that has passed, add a new week thirteen, and update everything in between for what actually happened. The update takes an hour or two once the structure exists.
Keep the variance. Each week, record what you forecast against what occurred, and keep the history. After six or eight weeks you will know your own bias, and it is almost always the same one: receipts assumed earlier than they arrive. Once you can see the pattern you can correct for it, and the forecast becomes genuinely predictive rather than aspirational.
A few scenarios, not twenty
Three cases is usually right. Base, which is your honest expectation. Downside, where the largest customer pays a month late and new sales come in below plan. And a stress case that answers one question: what has to go wrong before we breach.
Knowing the answer to that last question is worth the whole exercise. It converts an anxiety into a specific, monitorable condition, and it is the thing a lender most wants to see you have thought about.
Who should own it
The person who can chase a debtor and move a payment run. A forecast owned by someone who only reports it is a reporting exercise. Owned by someone who can act on it, it becomes the weekly agenda for managing cash, which is what it is for.