Cash flow forecasting for small businesses: how far ahead, and how accurate

Accuracy falls with distance because the inputs change in kind:

That is why a forecast is more useful as a range in its later weeks, and why the main question to ask of the last month is not “what will the balance be?” but “is there a week where it could fall below our minimum?”

Thirteen weeks is long enough to see a quarter’s payroll cycles, tax payments and supplier terms, and short enough that most of it rests on invoices and bills you already have. A 12-month forecast answers different questions: can we afford a hire, when will we need financing, what happens in our slow season.

• Opening cash. Start from the bank balance, not the book balance. Uncleared payments and deposits make every later day wrong by the same amount.
• Open accounts receivable. Place each invoice on the date the customer tends to pay, not the due date. Your days sales outstanding and an AR aging review show how far apart those are.
• Open accounts payable. Every approved bill, by its due date, plus committed purchase orders that will become bills. Bills stuck in approval or in someone’s inbox are the usual gap; see the AP month-end close checklist.
• Recurring payments. Rent, payroll, subscriptions, insurance and loan repayments. Quarterly and annual items, such as estimated tax payments and insurance renewals, are the ones most often missed.
• Timing assumptions. Which weekday payroll leaves, how long card receipts take to settle, whether you pay suppliers early to take a discount. Small timing errors move cash between weeks even when totals are right.

• Measure receipts and payments separately; a good closing balance can hide two large errors that cancel out.
• Save each week’s forecast before you update it. Comparing the forecast made four weeks ago with what actually happened shows how much your later weeks can be trusted.
• Record a reason for every large variance: timing (it moved to another week) or amount (it was a different size).
• Timing variances mean fixing a date assumption; amount variances mean fixing an estimate or a missing item.

Week 4 closes at $21,500, $3,500 under the buffer. Options include chasing a large overdue invoice, moving a supplier payment to its agreed due date instead of paying early, or arranging a short-term facility before it is needed.

At the end of week 1 the owner compares the forecast with the bank:

Payments were within 2%. Receipts were 20% short because one customer paid $8,400 nine days late. That is a timing variance: the money arrives later, so the owner moves the invoice to its new expected date, and from now on places that customer’s invoices about nine days after the due date. The rolled-forward forecast now starts from the actual $54,400.

1. 1

   Step 1

   Start from the actual cash balance in your bank accounts today, not the balance in your accounting system.
2. 2

   Step 2

   List known receipts: open customer invoices, placed on the date each customer actually tends to pay, not the due date.
3. 3

   Step 3

   List known payments: open bills by due date, committed purchase orders, payroll, rent, loan repayments and taxes.
4. 4

   Step 4

   Add recurring items you have not been billed for yet, such as subscriptions, insurance and utilities, from the last 12 months of history.
5. 5

   Step 5

   Add estimates for sales not yet invoiced in the later weeks, and mark them as estimates.
6. 6

   Step 6

   Set a minimum cash buffer and look for the first week the closing balance falls below it.
7. 7

   Step 7

   Each week, record actuals against the forecast, explain the biggest variances, fix the assumption behind them and roll the forecast forward.