Building a 13-week cash flow forecast that people actually use

Build it on receipts, not revenue

Cash forecasting fails when it starts from the profit and loss account. Start instead from the debtor ledger: take each open invoice, apply an expected payment date based on that customer's actual behaviour, and only then layer on new sales.

Historic average days to pay per customer is far more predictive than agreed terms. Most forecasts are optimistic because they assume contractual terms are met.

Model payments by category

Split outflows into payroll, taxes, suppliers, finance costs and discretionary spend. Payroll and tax dates are fixed and should be entered precisely. Supplier payments are the flexible element — the lever you can pull in a tight week.

Include VAT, PAYE and corporation tax as explicit lines with their true due dates. These are the payments that most often cause an unexpected shortfall.

• Payroll and pension: fixed dates, high certainty
• VAT and PAYE: fixed dates, calculated from actual liabilities
• Suppliers: variable timing, your main short-term lever
• Finance: loan repayments, interest, lease payments

Review variance every week

Each week, compare the previous week's forecast to actuals and explain the difference. This is the step that turns a spreadsheet into a forecasting capability, because it exposes systematic bias in your assumptions.

Then roll the model forward one week so the horizon always stays at thirteen. A forecast that is rebuilt monthly rather than rolled weekly loses its early-warning value.

Add scenarios, sparingly

Two downside scenarios are usually enough: a delay in your largest receipt, and a broad slowdown in collections. Model the effect on the lowest cash point rather than the closing balance.

The lowest point in the period, not the end of it, is what determines whether you need a facility.