Profit and cash are not the same thing. You can close a strong month on paper and still be short on the 15th because a client pays net-45 and payroll does not. A cash flow forecast fixes that blind spot by showing you, week by week, what your bank balance is likely to be — early enough that you still have cheap options. The version that survives is thirteen weeks long, lives in one file, and takes about twenty minutes every Friday.
Why Thirteen Weeks Beats a Twelve-Month Budget
Thirteen weeks is one quarter. That matters because most of the payments that hurt are quarterly: estimated taxes, insurance renewals, annual software billed on a cycle, sales tax filings. A quarter-long window catches all of them at least once.
It is also the shortest horizon that still gives you room to act. If a client pays on net-30 and you have to chase them, the distance between „invoice sent“ and „money in the account“ is often six to eight weeks. A four-week view tells you about the problem after your options have narrowed to borrowing or delaying payroll. Thirteen weeks tells you while you can still bill earlier, shift a purchase, or make three collection calls.
Go much further out and the weekly numbers stop being real. Week 30 is a guess dressed up as a forecast, and pretending otherwise is exactly how people lose trust in their own model.
Use weeks, not months. A monthly view averages away the trough. If payroll lands on the 15th, a big vendor invoice on the 20th, and your largest client pays on the 28th, the month looks fine and the third week is negative. That third week is the one that puts fees on your account.
What Goes in Each Column of a Cash Flow Forecast
Set up thirteen columns, one per week, dated by week ending Friday. Rows go in a fixed order and do not change mid-quarter.
- Opening balance — last week’s closing balance, carried forward automatically.
- Inflows — split into confirmed customer payments (listed invoice by invoice, not as a lump), recurring or subscription revenue, and other cash in such as a tax refund, loan draw, or owner contribution.
- Outflows — payroll and contractors, payroll tax deposits and estimated tax, fixed obligations (rent, insurance, loan payments), software and subscriptions, direct delivery costs such as materials, subcontractors and ad spend, and owner draw as its own line.
- Net movement — inflows minus outflows.
- Closing balance — opening plus net movement.
- Minimum operating balance — the floor you refuse to go below.
- Cushion — closing balance minus the floor. This is the row you actually read.
Pick your floor before you need it. A common approach is one full payroll run plus one month of fixed costs, but set it at a number that lets you sleep, then treat it as a hard line rather than a suggestion.
A worked week looks like this. Week 5 opens at $18,400. Expected inflows are $9,200. Known outflows are $21,700, because payroll and the quarterly insurance premium land together. Closing balance is $5,900 against a floor of $12,000 — a $6,100 gap. You are seeing that in week 1, with a month to fix it.
How to Treat Unpaid Invoices Honestly
This is where most forecasts quietly become fiction. The temptation is to enter every open invoice on its due date at full value. Clients do not read your terms that closely.
List every open invoice on its own row with the client, amount, invoice date, terms, and — the important column — the expected payment week based on how that client has actually paid you before, not on what the invoice says. If a client has taken 42 days three times running, forecast 42 days.
Then apply a confidence tier instead of pretending you know:
- Confirmed — client has told you the payment date or the ACH is scheduled. Forecast at full value.
- Reliable — consistent on-time history. Forecast at full value in their normal week.
- Slow but paying — past due under 30 days, still responsive. Forecast at a discount, or push it two weeks out. Pick one method and stay consistent.
- At risk — past due 30 days or more, or gone quiet. Remove it from the forecast entirely and move it to a collections list. It reappears only when a payment date is confirmed.
Three rules keep this clean. Never forecast an invoice you have not actually sent. Never include unsigned proposals — put them on a separate pipeline line that never touches the closing balance. And add a „times slipped“ counter next to each invoice.
That counter is the honest part. When an invoice slips for the second time, it is no longer a forecasting question, it is a collections question. Stop moving the row and make the call, send the statement, or pause new work for that client. Moving a number sideways every Friday feels productive and changes nothing.
The Friday Routine That Keeps It Current
Forecasts get abandoned because nobody decided when they get updated. Put a recurring twenty-minute block on your calendar for Friday afternoon and run the same six steps every time.
- Pull the actual bank balance as of Friday. From the bank, not your bookkeeping software, which may be days behind.
- Log the variance. Compare what you forecast for this week against what happened, for inflows and outflows separately. Write the two numbers down. Do not fix anything yet.
- Roll the window. Drop the completed week into a history tab and add a new week 13 at the far end.
- Update receivables. Walk each open invoice, confirm or move its expected week, bump the slip counter where needed, and assign one action to anything that moved.
- Update known outflows. Payroll dates, quarterly items, anything you committed to this week that was not already in the model.
- Read the cushion row. Find the first week where the closing balance drops below your floor. That week is your decision deadline.
Do it Friday, not Monday. On Friday you still remember which client promised what. By Monday it is archaeology.
The variance log is the part that pays off. After six or eight weeks you will see your own bias in writing — most people overestimate inflows and forget irregular outflows — and you can correct for it. The closing balance is also one of the few numbers worth putting on a weekly dashboard your whole team can see, because it turns „we should be careful“ into a specific week and a specific dollar figure.
What to Do When the Forecast Turns Red
A shortfall six weeks out and a shortfall next Tuesday are different problems. Match the lever to the lead time.
Six weeks or more: use the cheap levers. Call the three largest open invoices. Move a discretionary purchase to the next quarter. Bill mid-month instead of at month end. Ask for a deposit on new work before it starts.
Three to five weeks: pull billing forward to milestones rather than completion. Ask one specific client for early payment with a specific reason, which works far better than a general reminder. Talk to a vendor about extended terms now, while you are still current — vendors accept a plan and resent a surprise.
One to two weeks: sequence payments deliberately. Payroll and tax deposits first. Contact anyone you cannot pay in full before the due date, not after, and offer a partial payment with a date for the rest.
One more rule. If your closing balance dips below the floor in three or more of the thirteen weeks, the timing is not the problem. That is a pricing or cost structure issue, and no amount of forecasting will fix it — repeating a one-time rescue every quarter just hides it for longer.
Keeping the Model Alive Past Week Three
Freeze your row categories for the full quarter. Every time you restructure the sheet you lose comparability and add ten minutes to the routine. Keep one file with one owner. Never overwrite completed weeks — the history is what teaches you your own payment patterns. And do not merge this with your profit and loss statement; a cash flow forecast only cares about money moving, not about revenue recognition or depreciation.
If you want the structure already built, the 13-Week Cash Flow & Receivables Control Center ships as an 8-page start-here guide plus an Excel model with 11 sheets and 517 formulas, including the receivables tiers and variance log described above. If you would rather see what else fits your week, the full workbook library is a reasonable place to start.