Cashflow forecasts are one of the most over-built and under-used artefacts in SME finance. A founder asks the accountant for one, the accountant produces a 24-column spreadsheet projecting three years out, it gets emailed, looked at once, and then closed for ever. By month two the assumptions are out of date, by month six the model is fiction, and by the time the business needs a forecast — usually because cash is suddenly tight — the model is no help at all.
The forecasts that actually get used share three properties. Here is what they look like.
Property 1: Short horizon, high cadence
Three years out is a strategic plan, not a cashflow forecast. The right horizon for operational cashflow is 13 weeks. That length is short enough that the assumptions are mostly known (open invoices on the receivables side, contracted commitments on the payables side) and long enough to spot real cash crunches before they hit.
Cadence: refresh weekly. Not "when we remember". Every Monday morning. The forecast only earns trust if the actuals line up against the forecast week after week — and that requires the forecast to be refreshed and the variance examined every single week.
Property 2: Built on commitments, not aspirations
The most common reason forecasts fail is that they are built on hoped-for sales rather than committed cashflow.
The right approach for the 13-week model:
- Receipts = open invoices (with realistic collection dates by debtor) + contracted recurring revenue + new sales only where there is a signed PO/contract and a known invoice date
- Payments = open supplier invoices + payroll commitments + tax liabilities (PAYE, VAT, corporation tax) + rent and recurring direct debits + planned capital expenditure
If a line in the forecast cannot be tied back to a specific document or commitment, it does not belong in the 13-week model. Aspirational sales go in the longer-horizon strategic plan instead.
Property 3: Variance is reviewed, not ignored
The weekly cadence has one purpose: comparing what actually happened to what the forecast said would happen. Cash variance is one of the strongest leading indicators in a business. A few patterns to watch:
- Receipts under forecast by more than 10% for three consecutive weeks — credit control issue. Not a sales issue.
- Payments over forecast consistently — usually means commitments are being made faster than they are being captured. Cost discipline issue.
- Both wrong, in opposite directions — the model itself is broken and needs a rebuild.
The variance review is a 15-minute weekly conversation, not a quarterly board paper. Keep it operational.
The trap to avoid: false precision
A 13-week forecast accurate to the pound is suspicious. Cashflow is inherently uncertain. We deliberately build models with sensitivity bands — best/expected/worst — rather than single-point estimates. If the worst case shows cash going below £0 in week 8, that triggers action this week, not in week 7.
The discipline of running three scenarios in parallel is more useful than running one in great detail. Founders make better decisions when the range is visible.
Beyond the 13 weeks: the longer view
The 13-week model is operational. Sitting on top of it should be a longer-horizon model — usually 12 to 18 months — that links to the budget and strategic plan. This longer model is updated monthly, not weekly, and is driven by P&L projections rather than transaction-level commitments. The two models share the same opening cash balance and feed into each other:
- Weeks 1 to 13: commitment-based 13-week model
- Months 4 to 18: P&L-driven medium-term model
- The handover is at week 13 / month 4, and the two should reconcile
Building the model: practical structure
The structure we use for most clients:
- Opening cash balance — by bank account
- Receipts — by debtor for the first 4 weeks, then by revenue stream
- Operating payments — payroll, rent, recurring DDs, supplier payments by major creditor
- Tax payments — separately identified by tax type and due date
- Capital expenditure — by project
- Financing flows — loan repayments, drawdowns, equity
- Closing cash balance — by week
- Headroom against facility limit — explicit, not derived
Linking it to decision-making
The forecast is only valuable if it changes behaviour. Three weekly questions that should come out of the review:
- Do we need to accelerate collections this week?
- Should we delay any payments — and which ones can be delayed without supplier damage?
- Is there a CAPEX decision pending that the cash picture says no to?
How we help
For our business advisory clients we build the model, set up the weekly data extraction (usually directly from Xero and the bank feeds), and run the first 8 weeks of variance reviews jointly with the founder until the rhythm is self-sustaining. For clients using our finance BPO, the model is maintained as part of the standard monthly close cycle.
If you currently have a forecast that nobody opens, we can review it and propose either a rebuild or a simpler replacement that gets used. Send us your current model and we will turn around a critique within a week.