The 13-week cash forecast, demystified

If you've ever sat in a board meeting and heard the CFO talk about the "thirteen-week," you've watched the founders pretend they understood. The thirteen-week is one of the most useful tools in finance, and one of the most frequently bungled. This post is for the operators in the room.

What it is

A 13-week cash forecast is a week-by-week projection of your cash position over the next thirteen weeks. Not P&L. Not accruals. Cash. What's actually going to land in your bank account each week, and what's actually going to leave it.

The output is a single roll-forward table. Opening cash on Monday morning. Inflows over the week. Outflows over the week. Closing cash on Friday evening. Then that closing balance becomes the next week's opening balance.

Why thirteen

Thirteen weeks is roughly a quarter — long enough to see the seasonal shape of receivables, to anticipate the tax bills and the payroll cycles, and to give yourself runway for action if you spot a problem. Shorter than that and you're firefighting; longer than that and the projections become so uncertain they're worth less than the time spent on them.

What it tells you

Where the next pinch is. The week you'll dip below your minimum operating cash if you do nothing. Almost always solvable if you spot it three weeks out.
Whether your AR strategy is working. Stretches between the invoiced revenue and the cash arriving show up as cliffs in the forecast.
What headroom you have for a decision. Want to hire two engineers, buy a year's worth of inventory at a discount, pay down a credit line? The 13-week tells you whether the decision is comfortable, tight, or reckless.

The mistakes that ruin it

Mixing accruals into the cash forecast. The thirteen-week is cash, not P&L. If your "Q2 revenue" is a billed-but-not-collected number, it doesn't go in the inflows column on the day you billed.
Treating the forecast as a one-shot model. A 13-week is meant to be re-run weekly. Inputs change, outputs change, variance against last week's forecast is itself useful.
Optimism on collections. Founders consistently model their AR as if customers will pay net-30. Your actual data — which you have, and which most people don't bother to look at — will show median collections at 42 days.
Forgetting tax. VAT, payroll taxes, corporate tax instalments. They show up on predictable dates and they are not negotiable.
Underweighting one-offs. Annual software renewals, professional-services contracts that bill on completion, equipment purchases.
No scenario discipline. One forecast is fine for a calm quarter. For anything interesting, you need at least base case, downside, and stretch.

How we make it cheap to run

The reason most operators don't keep their 13-week current is that maintaining it is genuinely tedious in a spreadsheet. We're trying to remove that friction. In Finance Center, the forecast is a real-time view backed by your ledger: scheduled receivables show up automatically, invoiced AR pulls in with empirical collection patterns, recurring outflows like payroll and rent are configured once. Re-running the forecast is no work; it's just looking.

If you want to see how the forecast renders on your books, we'll walk you through it.

Further reading: the AICPA's treasury management resources and Aswath Damodaran's older working capital essays are good background.