A 13-week cash flow forecast is a rolling, weekly projection of cash receipts and payments across one quarter, refreshed every week so the completed week's actuals replace the forecast and a new week 13 is added. To implement one, you build a direct-method model from your bank balance, AR and AP ledgers, payroll schedule and tax calendar — then sustain it with a weekly variance review. The weekly rhythm, not the spreadsheet, is what makes it work.

This guide covers what the forecast is for, how to build it step by step, what accuracy to expect, why forecasts fail, and when a spreadsheet stops being enough.

What is a 13-week cash flow forecast?

A 13-week cash flow forecast is a direct-method liquidity model: it projects actual cash movements — customer receipts, supplier payments, payroll, tax — week by week, rather than starting from accounting profit. Each week's closing balance becomes the next week's opening balance, and the whole window rolls forward every week so the business always sees 13 weeks ahead.

It is not the statutory cash flow statement, and it is not the annual budget divided by 52. It sits beneath your longer-range monthly forecast as the tactical layer: the monthly model answers strategic questions about growth, headcount and funding over 12 months or more; the 13-week model answers exactly when money clears the bank, and which week is tightest. The Association of Corporate Treasurers (ACT) describes the next 13 weeks as the normal territory of an operational cash forecast, with medium-term monthly forecasting serving funding and investment decisions beyond it — and the convention was already documented in ACT's journal The Treasurer as far back as January 2000.

Why 13 weeks?

Thirteen weeks is one fiscal quarter — roughly 91 days. That horizon works for three reasons:

It matches how liquidity actually moves. A monthly forecast can show positive cash for the quarter while hiding the one week where payroll, a VAT payment and a large supplier run all land in the same seven days. Only weekly buckets expose the crunch weeks.

It is long enough to act. A problem visible eight weeks out leaves time to chase collections, reschedule a payment run, or arrange a facility draw. A four-week view often shows the problem too late to change it.

It is short enough to stay credible. Beyond roughly three months, week-level timing becomes guesswork — new orders, changing terms and pipeline uncertainty degrade the detail. That longer horizon belongs to your monthly forecast.

The number itself is a convention, not a rule. The horizon in most tools — Float included — is adjustable; 13 weeks is simply the default operating cadence most finance teams settle on because it tiles neatly against quarterly board cycles, VAT/GST quarters and (in Australia) BAS dates.

A note on the format's history. The 13-week forecast has deep roots in restructuring and turnaround work, where lenders and courts require it, and that heritage sometimes makes people assume running one signals trouble. The opposite is true in practice: ICAEW recommends a detailed rolling 13-week forecast for scale-ups as standard discipline, and turnaround practitioners themselves draw the distinction — a forecast produced on demand signals concern; a forecast that already exists, with a variance track record, signals control. Growing companies run one for the same reason large treasury teams do: liquidity moves weekly, so it should be managed weekly.

What does a 13-week forecast help you decide?

For a finance team of three or more in a growing business, the forecast informs a specific set of weekly decisions:

• Payroll confidence — every payroll date in the window covered, confirmed before anyone asks.
• Tax readiness — VAT/GST, PAYE and corporation or provisional tax on their real statutory dates, not smoothed into monthly averages. In all four of the UK, US, Australia and New Zealand, the statutory cash calendar is fixed and knowable 13 weeks out.
• Collections focus — which receipts are material to the lowest cash point, which have slipped, and who owns the chase this week.
• Supplier and payment-run timing — which payments to make, schedule or renegotiate.
• Hiring and discretionary spend — whether a new hire, a capital purchase or a campaign is cash-supportable within the quarter.
• Facility use — when to draw on an overdraft or revolving facility, and when a draw can be avoided.

The output that matters is not the thirteen columns — it is the lowest projected cash point, the week it lands in, and the actions attached to it.

How to build a 13-week cash flow forecast

The build is mechanically simple; the discipline is in the data quality and the weekly rhythm that follows.

1. Define the perimeter. Document which entities, bank accounts and currencies are in scope, the week-ending day, and who owns the model. Exclude restricted cash from available balances.

2. Start from reconciled cash. Week 0 is your reconciled bank balance at a stated cut-off — not the ledger balance. A stale opening figure contaminates all thirteen weeks.

3. Build the receipts side. Take the aged debtors ledger and time each expected receipt by when cash actually clears the bank, not the invoice due date. Use each customer's real payment behaviour for your largest accounts — Xero Small Business Insights data shows UK small businesses waiting roughly 29–30 days on average to be paid, with invoices running around 7–8 days late against terms in recent quarters (Xero XSBI, 2024–25) — so terms-based timing systematically flatters the forecast. Keep committed receipts (invoiced, confirmed) separate from expected pipeline.

4. Build the payments side. Payroll and payroll taxes at exact amounts on exact dates — this is the anchor of the model. Then supplier payment runs from the AP ledger, rent, subscriptions, debt service, and tax remittances on their statutory dates. Itemise capital expenditure rather than averaging it.

5. Show financing separately. Facility draws and repayments sit beneath operating cash flow, so the model shows both the unfunded position and the position after planned financing.

6. Keep categories disciplined. Enough rows to assign a variance to a cause; few enough that the weekly refresh stays under a couple of hours. For most businesses of this size that means roughly 15–30 meaningful rows. Flag one-off items separately so they do not contaminate the recurring pattern.

7. Add a variance layer from day one. A forecast-versus-actual comparison for every closed week is what turns the model from a report into a learning loop. Many free templates omit this entirely — it is the single most important thing to add.

What is the weekly rhythm that keeps it alive?

The forecast survives on cadence, not construction. The teams that sustain one run the same cycle every week, most commonly on Monday morning:

1. Replace the completed week with actuals from the bank, categorised against the model.

2. Review the variance line by line and assign a cause: timing (the cash moved, in a different week), amount (the event happened at a different value), or assumption (the model was simply wrong). Each has a different fix.

3. Re-forecast weeks 2–12 using what the variance taught you — a customer who paid late this week will probably pay the next invoice late too.

4. Add a new week 13 with rolled-forward payroll dates, supplier terms, recurring overheads and any new commitments.

5. Hold a short cash conversation — 20 to 30 minutes between the finance manager or controller who owns the model and the FD or CFO who makes the payment, hiring and drawdown calls. The essential feature is that the forecast review and the payment decisions happen in the same conversation. The CEO typically consumes a one-page summary: closing balances by week, the lowest cash point, and the headline variances.

After ten to twelve weekly cycles, teams know which lines they systematically over- or under-estimate, and the model becomes markedly more reliable. The improvement mechanism is the variance discipline itself.

How accurate should a 13-week forecast be?

Honestly stated: there is no independently published accuracy benchmark for 13-week forecasts at this company size. Specific percentages circulate online — week-band accuracy targets, adoption statistics — but they trace back to vendor content rather than primary research, and the definitions behind them vary too much to compare. What the independent evidence does establish is that this is genuinely hard: in the AFP's 2025 Treasury Benchmarking Survey, more than 60% of treasury professionals identified cash and liquidity forecasting as a leading challenge.

The realistic expectation is a confidence gradient, not a single number:

• Weeks 1–2 are dominated by known items — payroll, tax, approved payments, invoiced receipts — so misses here usually indicate a process problem (stale AR data, uncategorised transactions) rather than forecasting difficulty.
• Weeks 3–6 are still actionable, but collections timing starts to dominate the error.
• Weeks 7–13 are directional: useful for spotting the shape and timing of a low point, not for predicting a specific week's balance precisely.

Measure your own accuracy rather than chasing an industry number: lock each week's forecast before actuals land, compare weekly (never monthly — monthly averages away exactly the timing misses the forecast exists to catch), score receipts and payments separately so offsetting errors cannot hide, and track repeated same-direction misses as assumptions to fix. Set tolerance thresholds against your own cash buffer, tighter in weeks 1–4 than at the far end of the window.

Why do 13-week forecasts fail?

The failure modes are consistent across the treasury and practitioner literature, and they are behavioural more than technical:

• No variance review — the most consistently named killer. The forecast becomes a reporting ritual: produced, circulated, and never compared with reality, so the same errors repeat indefinitely.
• Due-date optimism — forecasting collections at stated terms when actual payment runs a week or more later. The most common single source of error.
• Stale AR data — remittances unlogged and disputes untreated, so the model treats every open invoice as collectable on schedule.
• Pipeline counted as cash — "hoped" receipts mixed in with committed ones. The documented pattern is a forecast that looks fine in week 1 and collapses by week 5.
• Cadence slipping — weekly becomes fortnightly becomes monthly, at which point the rolling window and the learning loop both break.
• Category sprawl — the refresh grows past its time budget and starts getting skipped.
• Key-person dependency — a hand-built spreadsheet with undocumented assumptions, maintained by one person in a small finance team, stops updating the week that person is away.

Spreadsheet or software?

Most teams start in Excel, from a downloaded template or an adviser-supplied model, and for some businesses that is genuinely sufficient: a single entity, a single currency, transaction volume low enough that the weekly refresh stays under an hour or two, one disciplined owner, and a variance log that is actually maintained.

Template users consistently hit the same walls, though. The first fill is easy; the weekly refresh is where adoption dies — the manual pull of bank transactions, AR and AP each week, the fiddly mechanics of rolling the window forward without breaking formulae, and the absence of any variance layer in most free templates. There is also a quieter risk: decades of research led by Raymond Panko found errors in the overwhelming majority of operational spreadsheets audited, with roughly 1–5% of formula cells containing errors — and a 13-week model is a grid of several hundred linked formulae edited by hand every week.

The balance tips towards connected software when any of the following arrive:

• Multiple entities or currencies — consolidation by hand is where spreadsheet models break first, and many finance-led teams at this size run more than one entity.
• Rising AR volume — many customers, disputes and part-payments make manual timing judgements the dominant time cost.
• The refresh repeatedly slipping past its time budget.
• More than one pair of hands — version control and auditability start to matter the moment a second person edits the file, or a board or lender wants a regular pack.

What changes with software connected to your accounting platform is the plumbing, not the method: actuals, AR and AP flow in automatically, the rolling window and variance comparison are maintained by the system, and the formula-error class of risk is engineered out. The weekly update becomes a review of the output rather than a data-assembly exercise. The judgement — when customers will really pay, which payments to move — stays with you.

How Float fits

Float builds the rolling forecast directly from your accounting platform, with live connections to Xero and QuickBooks Online (Sage Intacct support is in development — join the waitlist). Your invoices, bills and bank balances flow in automatically, so the receipts and payments sides of the model stay current without weekly re-keying.

The forecast horizon is adjustable — run the 13-week operating view alongside a longer-range monthly view in the same model. Scenario planning lets you hold best-case, expected and downside versions side by side, so the gap between "paid to terms" and "largest customer slips 30 days" is visible before it happens, and budgets versus actuals are tracked continuously so the variance discipline described above happens in the product rather than in a separate log.

Frequently asked questions

What is a 13-week cash flow forecast?

A 13-week cash flow forecast is a rolling, weekly, direct-method projection of cash receipts and payments across one quarter. Each week's closing balance becomes the next week's opening balance, and every week the completed week is replaced with actuals and a new week 13 is added, so the business always sees 13 weeks ahead.

Why is it 13 weeks rather than 12 months?

Thirteen weeks is one fiscal quarter — long enough to see a problem forming and act on it, short enough that week-level timing stays credible. Longer horizons are better served by a monthly forecast; the two run alongside each other rather than competing.

Does running a 13-week forecast mean a business is in trouble?

No. The format has restructuring heritage, but ICAEW recommends a rolling 13-week forecast for healthy scale-ups as standard practice. The distinction practitioners draw is that a forecast produced on a lender's demand signals concern, while one that already exists with a variance track record signals control.

How accurate should a 13-week cash flow forecast be?

There is no independently published benchmark at this company size, and circulating percentage targets are vendor-sourced. The realistic pattern is a gradient: weeks 1–2 should be tight because they are dominated by known items, weeks 3–6 carry growing collections-timing error, and weeks 7–13 are directional. Measure your own accuracy weekly with a variance log rather than chasing an industry number.

How long does the weekly update take?

For a single-entity business with a disciplined process, practitioners report the established weekly refresh as a one-to-two-hour job in a spreadsheet. Connected software removes most of the data-assembly portion, leaving the review and decisions.

Can I build a 13-week forecast in Excel?

Yes, and many teams start there. A spreadsheet is genuinely sufficient for a single entity and currency with modest transaction volume, one disciplined owner and a maintained variance log. Spreadsheet research led by Raymond Panko found errors in the overwhelming majority of operational spreadsheets audited, so keep a hand-built model simple, stable and checked.

When should we move from a spreadsheet to software?

The common tipping points are multiple entities or currencies, rising AR volume, the weekly refresh repeatedly slipping past its time budget, and more than one person needing to work in the model. Any one of these is usually enough.

Does Float support 13-week cash flow forecasting?

Yes. Float's rolling forecast runs the 13-week weekly view as an operating cadence, with an adjustable horizon so you can look further out in the same model. Data flows in from your accounting platform automatically, and scenarios let you hold expected and downside versions side by side.

Which accounting platforms does Float connect to?

Float connects live to Xero and QuickBooks Online. Sage Intacct support is in development — you can join the waitlist on our integrations page.

Who can see the forecast in Float?

Access is permission-based: you control who on the team can view or edit the forecast, so the weekly cash view can be shared with the FD, CEO or board without opening up the underlying accounting platform. Data is transferred and stored encrypted.

How much does Float cost?

Float's current plans and prices are on our pricing page at floatapp.com/pricing, including what is included at each tier.

See your next 13 weeks in Float. Connect Xero or QuickBooks Online and get a rolling cash flow forecast built from your live data. Start a free trial · Book a demo