Do you know what your bank balance will be in 90 days?
You check your bank balance more often than you'd admit. So do I.
At $3M to $5M in revenue, though, the bank balance has stopped being the most useful number. You can have $210k sitting there and be three weeks from a really bad Friday. The balance tells you where you've been, not what happens in the next twenty-one days: three payroll runs, the material buy for the job starting next month, the insurance audit premium balance.
Cash flow problems at this size announce themselves late, on a Tuesday night when you're trying to figure out if you can make Friday’s payroll from the operating account or need to draw from the LOC.
The fix is a weekly forecast.
Your P&L won't warn you
You can close a month at a healthy gross margin and still run out of cash in week six. The income statement records revenue when you earn it. The cash shows up whenever the customer sends it. Between those two dates sits everything that strains a growing company:
Work in progress you've paid for in labor and materials but haven't invoiced yet
Retainage — 5% or 10% held back per contract, sitting on your balance sheet, not spendable
Customers on net 30 who pay in 47 days because that's how their AP department works
Material and equipment bought in advance of a job start
A three-payroll month
Every one of those makes your cash position worse than your P&L suggests, and none of them appear on the P&L at all.
What a 13-week forecast is
Thirteen weeks is one quarter, laid out week by week rather than month by month.
Monthly forecasting hides exactly the problem you're trying to see: a month can net positive even though week three is $40,000 short. Payroll lands on the 10th and the 24th. Sales tax is due on the 20th. A monthly view averages all of it into a total that hides the nuances of the week by week view.
The forecast itself is a grid of thirteen columns, one per week. Starting cash on top, everything coming in, everything going out, and an ending balance that becomes next week's starting balance. That's all the math there is. The value is in drilling down on the timing.
The factors that drive the balance
Payroll, on the actual calendar. The specific dates, not "monthly payroll." If you run biweekly, two or three months a year have three payroll runs, and those are the months that surprise people. Put the tax deposits on their own line right behind each payroll run.
AR by customer, at their real pay behavior. Use the pay history. If a general contractor has paid you in 45 days for eleven straight months, forecast 45 days. Forecasting 30 because the contract says 30 is an easy way to make a forecast go wrong.
Retainage, held separately. It's earned, it's on your balance sheet, but it isn't cash until closeout. Give it its own line with a realistic release date, or leave it out entirely. Don't leave it buried in your AR assumptions where it looks collectible.
Committed purchases, including the ones with no bill yet. The PO you signed for $129k of inventory is a cash event whether or not the invoice has hit AP. Schedule out the down payment on your forecast according to when it’s required.
Annual and quarterly items. Estimated tax installments, GE or sales tax filings, insurance renewals, equipment notes and debt service, and the workers' comp or GL audit true-up, which will be bigger than last year's if you've grown payroll.
Your line of credit as a separate row. Available credit is not cash. Keep it below the ending balance line so you can see which weeks you'd be dipping into it and by how much.
What this looks like with real numbers
An illustration: a $4.8M specialty contractor on biweekly payroll.
Starting balance is $210,000. Feels comfortable. Then the quarter shapes up like this:
Payroll and taxes run about $62,000 per cycle, and this quarter has one month with three cycles
A $95,000 material purchase for a job starting in week seven, payable on delivery
A $28,000 estimated tax installment in week five
A $19,000 workers' comp audit true-up, invoiced in week six
Two receivables ($140,000 and $86,000) both on "net 30," both historically 45 to 50 days
Laid out weekly, the cash balance dips to roughly negative $23,000 in week six. The company is profitable the whole time. Three payroll runs, a tax installment, and an audit true-up simply clear before either large receivable arrives.
See that in week one and you have options: follow up with both customers now instead of in week five, push the insurance payment out two weeks, make a draw on the LOC. See it in week six and you're scrambling to make sure paychecks don’t bounce.
How to build a forecast
Start from what you already have:
AR aging, adjusted to how each customer pays
AP and open POs, dated to when you intend to pay them
The payroll calendar for the next thirteen weeks, including payroll taxes
Fixed monthly outflows — rent, insurance, software, loan payments
Known one-time expenses — annual renewals, deposits, audit true-ups
Follow two rules. Be conservative on income; a forecast that surprises you upward is great, one that surprises you downward defeats the purpose. And update it weekly against actuals, replacing projections with what happened. That weekly update is what turns it from a document into a control.
If you want to see the mechanics before committing to a process, I built a free browser-based version you can enter your own numbers into.
A forecast is only as good as the books underneath it
If your AR aging includes invoices that were paid in March, if AP isn't entered until someone gets a past-due call, if retainage has never been split out, if the last three months closed on cash basis with no accruals, the forecast will still produce a number. That number will be fiction!
That’s why cash forecasting is the last thing we build for a client. First comes an accrual close with real cutoff, a schedule behind every balance sheet account, and AR and AP that reflect reality. Then the forecast means something, because every input is one you can defend.
The rolling 13-week forecast in our Full Service Accounting engagement is a number maintained weekly on top of books that are closed.
Book a discovery call and we'll look at your last three months and tell you whether your books can support a forecast yet.
