A profitable month can be the most dangerous kind of reassurance. It says you made more than you spent on paper.
It doesn’t say whether your bank balance survives the Thursday when rent clears, the Friday when payroll clears, and the customer receipt that lands next Tuesday because “Net 30” was more of a mood than a calendar.
The missing control is a cash calendar: a direct-method view that sequences money by the week it clears.
Your budget tells you whether the business is pointed in the right direction. Your cash calendar tells you whether it reaches Friday.
Those aren’t competing spreadsheets. They’re two different instruments, and treating the monthly budget as both is how a profitable operation starts negotiating with a supplier at 4:40 p.m.
Two Clocks
The budget starts with sales, then builds costs against the volume those sales require, then tests cash. That order matters.
A company-wide growth percentage is ambition. A bottom-up sales build by rep, territory, service line, units, or hours is the number that can support crew-hours, materials, fuel, payroll, and inventory.
Your annual budget then becomes a fixed yardstick. Actual results are measured against it so you can see whether you hit the plan.
A rolling forecast is different. Each closed month becomes actual, a new month is added at the far end, and the forecast becomes your current best estimate of where the year lands. A lost renewal doesn’t become eleven months of repetitive commentary. It changes the remaining outlook.
Cash needs a faster clock. The near-term control window is conventionally 13 weeks, roughly 90 days, rebuilt weekly at daily or weekly grain. The longer view is typically 12 months, refreshed monthly.
The first catches a dated payroll or vendor obligation on a thin week. The second catches a slow quarter, a planned hire, or a loan payment that a weekly sheet would flatten into background noise.
Keep the two views reconciled where they overlap. Exit cash from week 13 should match the corresponding month in the 12-month view.
If it doesn’t, that disagreement isn’t a formatting issue. One set of assumptions is stale.
The Build
Start with a reconciled bank balance on a specific cutoff date, not the cash number sitting in your general ledger. Outstanding checks, in-transit deposits, processor holds, and bank fees mean the two can disagree.
Recorded cash, bank cash, and available cash are different quantities. An undrawn credit line is potential liquidity, not cash — show it separately with its expiry, covenant status, borrowing base, and lender discretion.
Then build the direct-method schedule in the order money actually moves:
- Pull receivables aging and place every open invoice in the week it is expected to clear. Use realized behaviour, not the term printed on the invoice. The corpus example is a Net 30 customer averaging 38 days across the last four quarters; forecast the receipt at 38 days.
- Pull payables aging and lay in payroll and payroll tax at fixed dates and amounts, then supplier runs, rent, subscriptions, debt service, and tax obligations when they will actually be released. A month is not a sequence. It is a bucket with better branding.
- Net receipts against disbursements by week, then show financing below operating cash. A credit-line draw may fund the position, but it must not disguise an operating gap as a healthy one.
That’s a direct-method forecast: expected cash receipts and disbursements by category, never back-calculated from net income. Use named invoices and payment dates for the 13-week window.
Past about 13 weeks, individual invoices aren’t issued or knowable enough for that precision. Switch the longer view to an indirect forecast built from forecast net income, noncash charges, and projected changes in receivables, payables, and inventory.
The Control Views
| Control view | Horizon and grain | Source data | Owner | Review cadence | Decision it supports |
|---|---|---|---|---|---|
| Direct cash forecast | 13 weeks; daily or weekly | Reconciled bank cash, receivables aging, payables aging, payroll, tax, supplier, rent, and debt dates | Forecast owner with the person who releases payments and makes borrowing decisions | Weekly | Whether dated obligations clear and whether financing is needed before the trough |
| Rolling annual view | Typically 12 months; monthly | Bottom-up sales build, cost assumptions, closed-month actuals, and revised operating assumptions | Budget owner with line-item and cost-center owners | Monthly | Whether the annual operating plan, hiring, slow-season exposure, and debt commitments remain viable |
Keep your owner-facing weekly number set short: cash position, receivables aging, revenue booked. A working dashboard commonly carries 5 to 8 elements and is designed to be read in under two minutes.
Gross margin, net margin, and the full P&L variance belong in the monthly close, when there’s a complete period to compare. A weekly review that carries every ratio isn’t more rigorous. It’s just harder to finish before somebody asks about payroll.
The Trough
Month-end cash is a photograph taken after the problem may already have happened.
Liquidity failure happens at the lowest available balance before an obligation settles. Record the daily or intraday minimum where payroll, tax, card settlements, or a supplier run produce sharp movements.
A month can open at $240,000, close at $260,000, and still go negative if $300,000 of payroll clears two days before a major customer receipt.
Compare that low point with the buffer you deliberately hold for it. Cash buffer days equal average daily cash balance divided by average daily cash outflow.
In the JPMorgan Chase Institute sample of roughly 597,000 small businesses, the median was 27 buffer days, with the 25th percentile at 13 and the 75th at 62. Restaurants had a median of 16 days and real estate businesses 47.
Those are context, not a universal target. A seasonal operator sizes the reserve against its own trough months from at least two prior cycles, before peak-season cash gets spent on inventory or a payroll ramp.
Where the Week Breaks
Use the symptom to find the stage that broke. Don’t begin with “cut costs” just because the checking account looks rude.
The month finishes above budget, but one week goes negative → sequencing is broken
The model has netted a month instead of placing receipts and disbursements on their actual clearing dates. Rebuild it by week from reconciled cash.
Week 1 is wrong before any assumption has had time to fail → the opening balance is broken
The forecast began from ledger cash rather than reconciled bank cash, or it omitted a hold, fee, unpresented payment, or restricted balance.
Receivables grow faster than revenue, or DSO climbs against your own terms → collections timing is broken
Age invoices by invoice date and contractual due date. Keep disputes, promised-payment dates, credit notes, unallocated receipts, invoice value, invoice count, and customer concentration visible. One customer funding half of next week’s receipts deserves a line of its own.
A cost line is favorably different every period after volume changed → the variance interpretation is broken
A static budget compares actual results with the original plan and blends volume with unit price or unit cost. Rebuild the line at actual activity volume with a flexible budget before you call the cost performance favorable. Repeated easy wins can also be budgetary slack, not operational genius.
Forecast error is no worse in week 13 than in week 1 → the rolling loop is broken
Near-term receipts are invoiced and nearer to settlement; later weeks are directional. A cited target is 90%+ accuracy in weeks 1–4, while weeks 9–13 aren’t a commitment. Flat error across the window means actual variances aren’t being rolled into the next forecast.
Ending cash is positive only after a draw, asset sale, or owner contribution → the operating bridge is broken
Separate cash generated by operations from cash supplied by lenders, owners, or asset sales. Positive ending cash funded by fresh debt doesn’t mean operations funded themselves.
The Variance Rule
The monthly review is where your budget earns its keep.
Calculate actual − budget in dollars and percentage for every material line, label revenue and cost signs correctly, and classify the exception by volume, price, timing, or one-off cause.
A positive revenue variance is favorable. A positive cost variance is unfavorable. The arithmetic is simple. The labels are where reports quietly lie.
Use a dual rule: a dollar floor and a percentage test. The corpus’s practitioner sources show policies that require both legs and policies where either leg fires, and that difference isn’t an invitation to improvise after results arrive.
State the house rule in advance, scale both dimensions to the account and the business, and apply it consistently. The cited working percentage range is 5–10%, with one cited practice using roughly 7% or $25,000.
SEC Staff Accounting Bulletin No. 99’s 5% rule of thumb concerns external financial-statement materiality, not an internal budget-variance standard.
Review detailed variances monthly, then step back quarterly to ask whether the plan itself is still right. A single miss is an event. The same direction for three periods is a pattern.
Show current-period, year-to-date, and prior-period comparisons so an invoice paid on August 1 rather than July 31 doesn’t get confused with a structural drift.
The Rebuild
Don’t overwrite the forecast to make it look smarter in retrospect. Keep original forecast, latest estimate, and actual columns.
Compare actuals with the forecast vintage that existed when somebody approved a hire, an inventory purchase, or a borrowing decision.
Score 1-week, 4-week, and 13-week views separately, and measure receipts and payments separately. A small net miss can hide two large errors that happen to cancel.
Classify each closed week’s variance as a timing, amount, or assumption error. Replace that week with bank-categorized actual activity, carry the correction through weeks 2–12, and append a new week 13.
That fixed loop is the control. The spreadsheet itself is just the clipboard.
A budget tells you whether the month should work. A cash calendar tells you whether the business survives the dates inside it.
Build the budget from sales, then cost, then cash. Run the cash forecast from reconciled bank balance, one week at a time. Protect the trough, not the month-end photograph.