Why 13 weeks
Thirteen weeks is one quarter. Long enough to see a problem while you can still act on it, short enough that you can forecast at the invoice level rather than by trend.
That last part is the point. An annual budget forecasts revenue as a number. A 13-week forecasts cash as specific expected receipts and payments, each with a date. Those are different exercises, and profitable companies run out of cash because they only did the first one.
Profit is an opinion, cash is a fact
You can book revenue in March, collect it in June and pay for the labor that produced it in February. The P&L calls March a good month. The bank account disagrees. The 13-week is the report that reconciles those two views.
The structure
Thirteen columns, one per week, with rows grouped into four blocks.
- Opening cash: the actual bank balance at the start of the week
- Receipts: customer collections, financing draws, tax refunds, anything landing in the account
- Disbursements: payroll, rent, AP runs, debt service, taxes, card settlements
- Closing cash: opening plus receipts minus disbursements, which becomes next week opening
Building the first one
- Start from the bank balance, not the general ledger. This report is about the account, not the books.
- Pull the AR aging and place each open invoice in the week you actually expect it to land, using your own collection history rather than the stated terms.
- Pull the AP aging and place each bill in the week you intend to pay it. That is a decision, not a prediction.
- Layer in the fixed and known: payroll dates, rent, debt service, insurance, quarterly taxes.
- Add recurring variable costs as an average: cards, utilities, software, freight.
- Only then add new business, and be conservative. Unsigned revenue rarely arrives on schedule.
Running it every week
The forecast is worth little the first time you build it and a great deal by the sixth, because the value is in the variance rather than the projection.
Every Monday, replace last week forecast with what actually happened and roll a new week onto the end. Then ask why the difference exists. Collections slipped a week? Your assumption about days sales outstanding is wrong, and now you know by how much. That loop is what makes week 10 believable.
- The minimum closing balance across all 13 weeks, not the ending balance
- Any week dipping below your operating floor, which should be at least one payroll
- Collection variance, the single most common reason forecasts miss
- Disbursement timing you control, because it is your fastest lever
Common mistakes
- Forecasting from the P&L instead of the bank account
- Using invoice due dates rather than actual payment behaviour
- Forgetting the third payroll in months with five Fridays
- Leaving out quarterly and annual items: insurance renewals, estimated taxes, software renewals
- Building it once and never rolling it forward, which turns a process into a document