How do I build a 13-week cash flow forecast?

The basic structure: thirteen weekly columns

The forecast is a simple grid: one row for opening cash, rows for each category of receipts, rows for each category of disbursements, and a closing cash row, repeated across thirteen weekly columns. Each week's closing cash becomes the next week's opening cash, so the whole model rolls forward automatically once the formulas are set up correctly. Thirteen weeks is the standard window because it covers roughly one quarter, long enough to see a seasonal dip or a large payment coming, short enough that the weekly detail still means something.

Some businesses extend or shorten the window depending on how far ahead they need visibility, but thirteen weeks is the starting point most lenders and advisors expect to see when this tool comes up.

A common early mistake is building the model with monthly totals instead of weekly ones. Monthly buckets hide exactly the kind of short, sharp cash squeeze the tool is meant to catch, a week where payroll, rent, and a large supplier payment all land within a few days of each other, even though the month as a whole looks perfectly healthy on paper.

What belongs on the receipts side

Receipts should be broken out by customer wherever a handful of customers make up most of the revenue, using each customer's actual payment history rather than the invoice due date. A customer who reliably pays thirty days late should be forecast that way, not on terms nobody expects them to meet. For businesses with many small customers, receipts can be grouped by typical collection timing instead, current, 30 days, 60 days, and beyond, based on the accounts receivable aging.

New sales not yet invoiced should only go in if there is real confidence they will close and be collected inside the thirteen-week window; optimistic revenue is the single most common way a cash flow forecast ends up wrong.

It also helps to keep a small contingency line for receipts that do not fit neatly into either category, one-off sales, tax refunds, or an asset sale, rather than forcing them into the regular customer rows where they distort the pattern the model is otherwise meant to show clearly.

What belongs on the disbursements side

Disbursements should be broken into categories that behave differently, not lumped into one number:

• Payroll, including source deductions, on the actual pay dates, not spread evenly across the month
• HST remittances, due on the filing schedule the business is actually on, monthly, quarterly, or annually
• Corporate tax instalments, where applicable, on the CRA's instalment due dates
• Debt payments, principal and interest, on the lender's actual schedule
• Rent and other fixed costs, which rarely move week to week
• Supplier payments, timed to when the business actually intends to pay, not when the invoice arrived

Breaking these out separately matters because a week that looks fine on total cash can still be tight the day payroll and an HST remittance land in the same few days.

It is worth marking which disbursements are truly fixed and which have some flexibility in timing. Rent and loan payments rarely move, but some supplier terms have room to negotiate a few days either way if a specific week is tight, and knowing that flexibility exists ahead of time is far more useful than discovering it under pressure.

Assumptions and the weekly variance review

Every forecast rests on assumptions, when a specific customer will pay, when a specific supplier will be paid, whether a tax instalment date is confirmed, and those assumptions should be written down somewhere next to the model, not left implicit. The forecast is not finished once it is built; the real value comes from comparing each week's actual receipts and disbursements against what was forecast, then asking why the gap happened and adjusting the next few weeks accordingly.

A forecast that is never checked against actuals tends to drift further from reality every week it goes unchecked, which is often exactly how a business ends up profitable on paper but perpetually short on cash. A short weekly review, even fifteen minutes, keeps the model honest and catches a looming cash gap early enough to actually do something about it.

Tools: spreadsheet or dedicated software

Many businesses run this in a spreadsheet, which works well and is easy to customize once someone has built the formulas correctly. Purpose-built tools like Float or Fathom connect directly to accounting software and update some of the actuals automatically, which saves time once a business outgrows manual spreadsheet updates. Neither approach is inherently better; the spreadsheet route costs less to start, and the dedicated tools save time as the business and the forecast get more complex.

Seasonal businesses need a different lens

A business with a strong season and a slow season needs the forecast to reflect that pattern explicitly rather than smoothing it out, since the slow-season weeks are exactly where a cash gap tends to show up. Our post on managing cash flow in a seasonal business covers how to build a reserve and time major payments around a predictable seasonal pattern.

How we build these with clients

As part of our CFO advisory work, we typically build the first thirteen-week forecast together with a client using their actual receivables aging and payment history, then hand over a version they can update weekly on their own, with us reviewing the variance each month. The goal is a tool the business actually keeps using, not a one-time report.