Who it's for: anyone in a SaaS or tech finance team who reports burn and runway, or builds or updates the weekly cash forecast.
What you'll get: clear definitions of gross burn, net burn and runway, a step-by-step method for a 13-week cash forecast, a full worked example, a weekly update routine, and a risks and controls table.
Reading time: about 13 minutes. Keep your bank balances and payment run dates to hand.
Why cash gets its own forecast
Profit tells you whether the business model works. Cash tells you whether the business is still here next quarter. In SaaS the two can drift a long way apart, because annual upfront billing, VAT, payroll timing and supplier runs move cash in lumps. A 13-week forecast shows those lumps week by week, so nobody is surprised by a low point that was visible a month ago.
Key terms
- Gross burn: total cash going out in a period, before any cash coming in.
- Net burn: cash out minus cash in. If cash in is bigger, the company is cash generative for that period.
- Runway: cash available divided by average monthly net burn. Shown in months.
- 13-week cash forecast: a week-by-week forecast of receipts and payments for the next quarter, built from known items (invoices, payroll, tax dates) rather than from the P&L.
- Direct method: forecasting actual receipts and payments line by line. Most 13-week forecasts use this.
- Indirect method: starting from profit and adjusting for non-cash items and working capital. More common for longer-range forecasts.
- Headroom: the gap between the forecast low point and the minimum cash level the business wants to hold.
- Variance to forecast: the difference between what was forecast for a week and what actually happened.
Burn and runway: getting the definitions straight
Many arguments about runway are really arguments about definitions. Write yours down and use them every month.
| Question | Common approach | Why it matters |
|---|---|---|
| Which cash counts? | Cash at bank and short-term deposits you can get at quickly. Exclude cash held as security (for example a rent deposit). | Locked-up cash can't fund payroll |
| Which burn figure? | Average monthly net burn over the last three months, or forecast burn for the next three to six | One month can be distorted by an annual renewal or a VAT payment |
| Include fundraising? | Exclude equity raised and loan drawdowns from burn; show them separately | Otherwise a funding round makes burn look like it fell |
| Include undrawn facilities? | Show runway with and without them | Lenders may set conditions that restrict access |
Backward-looking runway (using past burn) is simple and hard to argue with. Forward-looking runway (using forecast burn) is more useful when hiring plans or pricing are changing. Many boards ask for both.
Building a 13-week cash forecast, step by step
| Step | What "done" looks like | |
|---|---|---|
| ☐ | Set the weeks. Pick a week-ending day (Friday is common) and lay out 13 columns. | Week dates agree to the calendar, with bank holidays noted |
| ☐ | Put in opening cash from the bank reconciliation, not the ledger alone | Opening cash equals reconciled bank balances across all accounts |
| ☐ | Forecast customer receipts from aged debtors and the billing schedule | Each large debtor has an expected payment week; renewals due in the 13 weeks are included |
| ☐ | Apply realistic collection timing | Based on how customers actually pay (for example card payers on the day, invoice payers 30 to 45 days after invoice), not on payment terms alone |
| ☐ | Add payroll on pay dates | Net pay on payday; pension in the week it's collected |
| ☐ | Add PAYE and National Insurance to HMRC | In the week it must reach HMRC (see the dates below) |
| ☐ | Add VAT payments | Each VAT quarter's payment in the week it's due |
| ☐ | Add supplier payments from the purchase ledger and known commitments | Payment runs in the right weeks; annual renewals, rent and insurance included |
| ☐ | Add one-offs: corporation tax, loan repayments, capital spend, fundraising receipts | Each one-off listed with its source |
| ☐ | Calculate weekly net flow and closing cash | Closing cash each week = opening + receipts − payments |
| ☐ | Find the low point and compare it with your minimum cash level | Lowest week and headroom shown clearly at the top |
| ☐ | Write down the key assumptions | A short list a reader can challenge |
UK tax dates to build in (as at October 2026)
These are the common ones for a UK company. Check gov.uk for your own position, as dates depend on your payment frequency, VAT scheme and company size.
- PAYE and National Insurance: for monthly payers, due to HMRC by the 22nd of the next tax month if paying electronically (tax months run from the 6th to the 5th). Funds need to have reached HMRC by then, so allow clearing time when the 22nd falls on a weekend or bank holiday.
- VAT: for most businesses on standard quarterly returns, the return and payment are usually due one calendar month and seven days after the end of the VAT period. The payment needs to have cleared into HMRC's account by that date.
- Corporation tax: for companies that aren't paying in instalments, usually due nine months and one day after the end of the accounting period. Larger companies pay in quarterly instalments; gov.uk explains when this applies.
AI shortcut: give an approved AI tool your aged debtors report and the last six months of customer receipts, and ask it to work out each large customer's average days from invoice to payment, then suggest an expected payment week for every open invoice. Spot-check the slowest and largest customers yourself before using the dates.
Worked example: MonthEndClub Co Ltd, November to January
MonthEndClub Co Ltd is a fictional UK SaaS company. It has £2,950,000 of cash at 31 October. Payroll is paid on the 25th of each month, or the working day before if the 25th isn't a working day (so 24 December). Monthly PAYE and National Insurance is paid electronically: the November payment is made on Friday 20 November because the 22nd is a Sunday. It files VAT quarterly, with a quarter ending 31 October, so that payment is due on 7 December. Many customers renew annually in January.
The forecast (£k, weeks ending Friday).
| Wk | Week ending | Opening | Receipts | Payroll | HMRC PAYE/NI | VAT | Suppliers | Net flow | Closing |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 6 Nov | 2,950 | 85 | 0 | 0 | 0 | (80) | 5 | 2,955 |
| 2 | 13 Nov | 2,955 | 90 | 0 | 0 | 0 | (75) | 15 | 2,970 |
| 3 | 20 Nov | 2,970 | 85 | 0 | (100) | 0 | (85) | (100) | 2,870 |
| 4 | 27 Nov | 2,870 | 80 | (250) | 0 | 0 | (75) | (245) | 2,625 |
| 5 | 4 Dec | 2,625 | 95 | 0 | 0 | 0 | (100) | (5) | 2,620 |
| 6 | 11 Dec | 2,620 | 85 | 0 | 0 | (90) | (80) | (85) | 2,535 |
| 7 | 18 Dec | 2,535 | 80 | 0 | 0 | 0 | (75) | 5 | 2,540 |
| 8 | 25 Dec | 2,540 | 60 | (250) | (100) | 0 | (95) | (385) | 2,155 |
| 9 | 1 Jan | 2,155 | 35 | 0 | 0 | 0 | (45) | (10) | 2,145 |
| 10 | 8 Jan | 2,145 | 180 | 0 | 0 | 0 | (100) | 80 | 2,225 |
| 11 | 15 Jan | 2,225 | 260 | 0 | 0 | 0 | (85) | 175 | 2,400 |
| 12 | 22 Jan | 2,400 | 210 | 0 | (100) | 0 | (75) | 35 | 2,435 |
| 13 | 29 Jan | 2,435 | 120 | (250) | 0 | 0 | (85) | (215) | 2,220 |
| Total | 2,950 | 1,465 | (750) | (300) | (90) | (1,055) | (730) | 2,220 |
Check the totals: receipts £1,465k less payments of £2,195k (£750k + £300k + £90k + £1,055k) gives a net outflow of £730k. Opening £2,950k less £730k gives closing cash of £2,220k, which matches week 13.
What the forecast shows.
- The low point is week 9 (£2,145k), just after December payroll, December's PAYE and a slow holiday period for receipts. If the company's minimum cash level is £1,500k, headroom at the low point is £645k.
- January renewals bring cash back up in weeks 10 to 12. If those renewals slip by two weeks, the low point moves later and deeper. That's the main assumption to test.
- Net burn over the 13 weeks is £730k, about £243k a month (£730k ÷ 3).
- Forward runway at the end of January is about 9.1 months (£2,220k ÷ £243.3k).
A simple downside case. If half of the January renewal receipts in weeks 10 to 12 (£650k in total, so £325k) slip past week 13, closing cash at week 13 falls to £1,895k (£2,220k − £325k). Still above the £1,500k minimum, but headroom drops to £395k. Showing a downside line like this in the pack is a common approach.
AI shortcut: once the forecast is built, give an approved AI tool the table and your list of assumptions, and ask it to write three downside scenarios (for example late renewals, a large customer churning, a hiring plan brought forward) with the cash effect of each by week. Check every figure it produces against your own model before it goes anywhere.
Keeping it up to date: the weekly routine
| Task | What "done" looks like | |
|---|---|---|
| ☐ | Replace last week's forecast with actuals from the bank | Actual receipts and payments by line, agreed to bank statements |
| ☐ | Record the variance to forecast for each line | Variances above your threshold have a reason (timing or permanent) |
| ☐ | Roll forward: drop the past week, add a new week 13 | Always 13 weeks ahead |
| ☐ | Update receipts from the latest aged debtors | Paid invoices removed; new invoices added with expected weeks |
| ☐ | Update payments from the purchase ledger and payment run plan | Approved payment runs match the forecast |
| ☐ | Re-check the low point and headroom | Any fall below your minimum flagged to the finance lead the same day |
| ☐ | Save a dated copy | Each week's version kept, so accuracy can be tracked over time |
Tracking accuracy is what builds trust. A simple measure: for each week, compare the forecast made four weeks earlier with the actual. If receipts are always late, the collection assumptions need work.
Common mistakes and audit watch-outs
- Forecasting from the P&L. Revenue isn't cash. Annual invoices, deferred revenue and VAT make them very different in SaaS.
- Forgetting VAT on receipts. Customer receipts include VAT where charged; the VAT payment to HMRC comes later. Leaving one side out distorts both.
- Using payment terms instead of real behaviour. If customers on 30-day terms pay in 45, forecast 45.
- Missing annual and quarterly items. Insurance, software renewals, audit fees, corporation tax and rent quarters are easy to forget. Keep a list of known lumpy payments.
- Opening cash that doesn't match the bank. Always start from reconciled balances, including every account and currency.
- Counting a funding round as reduced burn. Show equity and loans separately.
- Unchecked AI scenarios. AI can produce neat but wrong arithmetic. Every AI-generated figure is checked against the model, and only company-approved AI tools are given bank or customer data.
If you report under US GAAP (ASC 606)
The cash forecast is the same under any framework, because cash is cash. What can differ is the revenue and deferred revenue figures used to explain why cash and profit differ. For standard SaaS subscriptions, IFRS 15 and ASC 606 usually give the same timing, so the explanations rarely change.
Tools that can help
Many teams build the 13-week forecast in Excel or Google Sheets, fed by exports from Xero, QuickBooks, Sage or NetSuite. Cash flow tools such as Float connect to the ledger and help lay out receipts and payments by week. Billing data from Stripe or Chargebee helps with renewal timing. A company-approved AI assistant can help with collection patterns and scenario drafts. None of these replace a weekly routine of comparing forecast to actual and fixing the assumptions.
Key risks and controls
| Risk | Control | How to evidence it |
|---|---|---|
| Opening cash is wrong | Opening cash taken from reconciled bank balances | Bank reconciliation reference on the forecast |
| Receipts forecast too early | Collection timing based on actual payment history | Forecast-to-actual variance log for receipts |
| Tax payment missed from forecast | Standing list of PAYE, VAT and corporation tax dates | Tax calendar attached to the forecast file |
| Low point not spotted in time | Minimum cash level with headroom check each week | Weekly email to the finance lead with low point and headroom |
| Runway calculated differently each month | Written definition of cash and burn | Definitions note in the board pack |
| AI errors or data leaks | Approved AI tools only; all AI output checked by a person | Reviewer sign-off on scenarios |
Rates, thresholds and rules in this guide
Figures were checked against the pages below in October 2026. They change regularly (often each April or at Budget time), and we can't guarantee they're still current or right for your situation. Always check the official page before relying on them.
- Pay employers' PAYE (GOV.UK): PAYE and National Insurance payment dates
- DMBM520055: PAYE payment frequencies and due dates (HMRC manual): tax months and cleared funds by the 22nd
- Sending a VAT Return: when to do a VAT Return (GOV.UK): VAT payment deadline and clearing
- Pay your Corporation Tax bill (GOV.UK): Corporation Tax due 9 months and 1 day after the period
- Pay Corporation Tax if you're a large company (GOV.UK): quarterly instalments
Key takeaways
- Write down your definitions of cash, net burn and runway, and use them every month.
- Build the 13-week forecast from known receipts and payments, not from the P&L.
- Start from reconciled bank balances and include PAYE, VAT and corporation tax dates.
- Find the low point, show headroom against a minimum cash level, and test the biggest assumption.
- Update weekly, compare to actuals and keep each version.
- Use AI for collection patterns and scenario drafts; a person checks every figure.
FAQs
Why 13 weeks?
It covers a full quarter, which catches a VAT payment, three payrolls and most annual renewals due soon, while still being short enough to forecast line by line with reasonable accuracy.
How is runway different from the 13-week forecast?
Runway is a single summary number: how many months cash would last at a given burn. The 13-week forecast shows the path week by week, including low points that a runway figure can hide.
Should runway include an undrawn loan facility?
A common approach is to show runway both with and without it, because facilities often come with conditions that could restrict access.
Can AI build the cash forecast?
It can help with parts of it: estimating when customers will pay, listing recurring supplier payments and drafting scenarios. The structure, assumptions and sign-off stay with a person, and only company-approved AI tools are used with bank and customer data.
How accurate should a 13-week forecast be?
There's no single standard. Many teams track the gap between forecast and actual closing cash four weeks out, and look for steady improvement rather than a fixed target.
Related guides
- Monthly Management Accounts and Board Pack Layout
- ARR vs MRR vs Revenue: Why They Don't Match and How to Reconcile Them
- Bank Reconciliations Done Properly
- Reforecasting Without the Pain
More in Metrics and Reporting guides.
From the Club
Runway is finite. Coffee, thankfully, is not. The Team Mug · The Dad Hat
Join the Club. New guides, early access and members-only batches. Join here.
This guide is general information about common practice. It isn't tailored to any business, and standards, tax rules and figures change. Check your own circumstances, the latest rules and, where needed, a qualified professional before acting.
Last reviewed: October 2026.