The annual budget answers a strategic question: what do we intend to do this year? It cannot answer the operating question: will the cash be there on the 15th?
The rolling 13-week cash flow forecast fills that gap. Thirteen weeks is one quarter — long enough to act, short enough to be accurate.
Why weekly, why thirteen
- Weekly matches how cash actually moves. Payroll, rent, and major vendor runs are weekly or biweekly events. A monthly view hides a mid-month dip that a weekly view exposes.
- Thirteen weeks gives enough runway to draw on a credit line, delay a hire, or chase collections before the shortfall lands.
- Rolling means every week you drop the week just completed and add a new week thirteen out. The forecast never expires.
Structure
One tab. Columns are weeks. Rows are:
Opening cash — actual bank balance at the start of the week.
Receipts
- Collections on existing accounts receivable (by expected week, not invoice date)
- Collections on revenue not yet invoiced
- Other inflows: loan draws, tax refunds, asset sales
Disbursements
- Payroll and payroll taxes (exact dates)
- Rent and fixed contracts
- Vendor payments (by due date, adjusted for how you actually pay)
- Debt service
- Sales tax, income tax, and other statutory payments
- Capital expenditure
Net cash flow = receipts minus disbursements.
Closing cash = opening plus net. Becomes next week's opening.
Add a line for available credit so the forecast shows total liquidity, not just the bank balance.
Building it the first time
- Pull twelve months of bank transactions. Categorize them into the rows above. This gives you the real pattern of how money moves, not the accrual view.
- Age your receivables by expected collection week. Not by due date. If a customer always pays 20 days late, forecast it 20 days late.
- List every fixed disbursement with its exact date. Payroll, rent, loan payments, subscriptions. These are known and should be right.
- Estimate variable disbursements from the twelve-month history. Use averages, then adjust for known changes.
- Reconcile week one to the actual bank balance. If it does not tie, the model is wrong somewhere.
Keeping it accurate
Every Monday:
- Replace last week's forecast with actuals
- Record the variance by row
- Ask why for any variance over a set threshold (often 5% of the row or a fixed dollar amount)
- Roll a new week onto the end
The variance log is where the forecast earns its keep. After eight weeks you know which assumptions are reliable and which are wishful.
A worked example
A $5M staffing firm, weekly payroll of roughly $75,000, clients paying on 45-day terms. The annual budget showed positive cash every month. The 13-week forecast showed:
- Week 6: closing cash of $18,000, against a payroll of $75,000 in week 7
- Cause: two large clients' payments were forecast by due date but historically arrived 15 days late
- Fix: line of credit draw arranged in week 3, and collections calls started in week 2 rather than week 7
Without the weekly view, the shortfall would have appeared four days before payroll.
Common mistakes
- Forecasting receivables by invoice due date instead of actual customer behavior
- Omitting payroll tax remittances and sales tax, which are large and lumpy
- Letting the forecast go stale for three weeks and then rebuilding from scratch
- Treating it as a finance document rather than a leadership document — the CEO should see it weekly
What to do this quarter
- Build version one this week, even if it is rough. Accuracy comes from the Monday variance review, not from the first draft.
- Set a minimum cash threshold. When the forecast breaches it in any week, that is the trigger for action.
- Share the closing cash line with your leadership team every Monday. Cash discipline follows visibility.