The Report Nobody Builds Until They Need It
There is a pattern we see often enough to set a watch by. A business trades happily for three or four years without ever forecasting cash. Then something changes, usually a bank asking for projections, a big contract that needs funding up front, or a month where the wages run went through and the balance afterwards was smaller than anyone expected. Only then does somebody open a spreadsheet.
The forecast that gets built at that point is nearly always too late to be useful and, more importantly, nearly always wrong. Not wrong in the sense of a bad guess about sales, which is forgivable, but wrong in its construction. It is usually a budget with the word "cash" typed at the top: revenue in the month it was earned, costs in the month they were incurred, no VAT anywhere, no tax bill, and a smooth line where the real bank account has cliffs.
Between June 2025 and May 2026, one in every 196 companies on the Companies House effective register entered insolvency, and the overwhelming majority of those were creditors' voluntary liquidations, which is the route a director takes when the company cannot pay what it owes as it falls due. Very few of those companies were unprofitable on the day they stopped. They ran out of cash on a specific date, and in most cases that date was visible six weeks earlier to anyone who had looked.
This guide covers how to build a forecast properly: what belongs in it, what must never go in it, how to get the timing right rather than guessing, and how to read the output. It also covers the part almost everyone skips, which is checking the forecast against what actually happened and using the difference.
What a Cash Flow Forecast Actually Is
A cash flow forecast is a schedule of every pound you expect to enter and leave your bank account, listed on the date the money is expected to move. That is the whole definition, and the important word in it is "date".
It is worth being precise about how this differs from the two documents it gets confused with.
A cash flow statement is historic. It reports the cash that moved during a period that has already finished, grouped into operating, investing and financing activities, and it sits in your statutory accounts. It tells you what happened. A forecast tells you what is going to happen, and it is a management document with no statutory format at all, which is both its strength and the reason so many are badly built.
A budget is a forecast of profit. It runs on the profit and loss calendar, where income is recognised when earned and costs when incurred, regardless of when the money moves. A budget and a cash flow forecast for the same business, over the same year, should produce two different numbers and two completely different shapes. If yours look similar, one of them is built wrong.
The distinction that trips people up is the accruals concept. Almost every number in your accounts has been adjusted so it lands in the right period rather than the period the money moved. A cash flow forecast reverses all of that. You are not interested in the period the cost relates to. You are interested in the day it leaves the bank.
What Goes In, and What Never Should
The rows in a cash flow forecast are not the rows in your accounts. Some things appear that have never once shown up in your profit figure, and some things that dominate your accounts do not appear at all.
Include these, even though they are not costs in your accounts:
- VAT. The VAT you collect is not income and the VAT you pay on purchases is not a cost, so neither appears in your profit and loss account. Both move through your bank, and the quarterly net payment is often one of the largest single amounts a small business pays all year. Leaving VAT out of a forecast is the single most common construction error we see.
- Corporation tax. It is a cash payment on a fixed date, and it relates to a year that finished nine months earlier.
- Loan capital repayments. Only the interest touches your profit and loss account. The capital element empties the bank all the same.
- Asset purchases in full. If you buy a £24,000 van, the forecast shows £24,000 (or the deposit and the monthly instalments, depending on how it is financed) on the date it is paid. Your accounts will show four or five years of depreciation instead.
- Dividends and drawings. A dividend is not an expense, so it never reduces profit, but it does reduce the balance. So does a director's loan account repayment.
- Personal tax. For sole traders and for directors with a self assessment liability, the January and July payments on account come out of the same bank account for many owner-managed businesses.
Exclude these, even though they are large numbers in your accounts:
- Depreciation and amortisation. No money moves. Ever. If depreciation appears in your cash flow forecast, delete the row.
- Accruals, prepayments and provisions. These exist purely to move a cost between accounting periods. The forecast only cares about the payment date.
- Bad debt provisions. You do not add a provision as a cash outflow. You remove the receipt you no longer expect.
A Worked Example
Take a commercial joinery firm turning over roughly £900,000 a year net of VAT, with a 31 March year end and VAT quarters ending in February, May, August and November. It is profitable, it has no overdraft facility, and it opens September with £12,000 in the bank.
The owner builds a six-month monthly forecast covering September to February. Receipts lag sales by about two months, materials and subcontractors are paid in the month following the work, and payroll and overheads run at a steady £24,000 a month. The VAT for the quarter ending 30 November is due by 7 January, and the corporation tax for the year ended 31 March 2026 is due on 1 January 2027.
| Sep | Oct | Nov | Dec | Jan | Feb | |
|---|---|---|---|---|---|---|
| Opening balance | 12,000 | 20,000 | 20,000 | 30,000 | 46,000 | 19,000 |
| Receipts from customers | 84,000 | 78,000 | 90,000 | 90,000 | 90,000 | 72,000 |
| Materials and subcontractors | (52,000) | (54,000) | (56,000) | (50,000) | (44,000) | (52,000) |
| Payroll and overheads | (24,000) | (24,000) | (24,000) | (24,000) | (24,000) | (24,000) |
| VAT | (18,000) | |||||
| Corporation tax | (31,000) | |||||
| Closing balance | 20,000 | 20,000 | 30,000 | 46,000 | 19,000 | 15,000 |
Read that table the way an owner reads it and the conclusion is comfortable. Six months, never a negative month, and the worst point is a £15,000 balance at the end of February. January absorbs £49,000 of tax and still closes at £19,000. Tight, but fine.
It is not fine, and the reason is that a month is far too long a unit of time to be useful.
The Number That Matters Is the Trough
Every monthly forecast contains a hidden assumption that nobody states out loud: that within each month, the money arrives before it leaves. It usually does not.
Here is the same January, broken into weeks. Corporation tax goes out on the 1st. VAT goes out by the 7th. The quarterly rent invoice, dated on the Christmas quarter day, gets paid in the second week. Customer receipts in the first fortnight of January are close to nothing, because everybody's accounts department was shut for a fortnight and the payment runs restart late.
| 1-7 Jan | 8-14 Jan | 15-21 Jan | 22-28 Jan | 29-31 Jan | |
|---|---|---|---|---|---|
| Opening balance | 46,000 | (8,000) | (15,000) | (7,000) | 13,000 |
| Receipts from customers | 6,000 | 14,000 | 18,000 | 46,000 | 6,000 |
| Materials and subcontractors | (11,000) | (12,000) | (10,000) | (11,000) | |
| Payroll, PAYE and rent | (9,000) | (15,000) | |||
| VAT | (18,000) | ||||
| Corporation tax | (31,000) | ||||
| Closing balance | (8,000) | (15,000) | (7,000) | 13,000 | 19,000 |
Same month. Same total receipts, same total payments, same £19,000 closing balance. The business is £15,000 overdrawn on 14 January, it has no facility, and it stays in the red for the best part of three weeks before the month-end payment run rescues it.
This is the point we would make more forcefully than anything else in this guide. The output of a cash flow forecast is not the closing balance. It is the lowest balance in the period, and the date it occurs. Owners read the right-hand end of the row because that is where the eye goes and because that is the number the spreadsheet bolds. The number that determines whether the business survives the period is somewhere in the middle, and on a monthly grid it is invisible by construction.
There is a practical consequence. The trough is also the size of the facility you need to arrange, and the date of the trough tells you when to arrange it by. A bank will lend against a £15,000 trough in November, when your last filed accounts are recent, your bank conduct is clean and you have a forecast that shows you saw it coming. The same bank is a great deal less helpful on 12 January.
Timing: Use Your Own Debtor Days, Not Your Payment Terms
Almost every wrong forecast we are asked to look at is wrong in the same direction, and the cause is nearly always the same. The receipts have been placed using the payment terms printed on the invoice rather than the dates customers actually pay.
Your terms are a statement of intent. Your debtor days are a measurement. In construction, average actual payment times run around 61 days against typical 30-day terms, and across the wider economy the Small Business Commissioner's research puts the average small business owed roughly £22,000 in overdue invoices at any one time. If you forecast on 30 days and get paid on 61, you have pulled a month of receipts forward through the entire model, and the error compounds down the row rather than cancelling out.
The fix takes about ten minutes. Debtor days = (trade receivables ÷ turnover) × 365. Take the trade receivables figure from your last balance sheet and the turnover for the same year.
There is a wrinkle in that formula that is worth knowing, because it makes almost every published debtor days figure too high. Trade receivables include VAT, because that is what the customer owes you. Turnover excludes it. Divide one by the other and you are comparing a gross number with a net one.
Take our joinery firm with £172,000 of trade receivables and £900,000 of turnover. The standard formula gives 69.8 days. Gross the turnover up to £1,080,000 to put both figures on the same footing and you get 58.1 days. That is nearly twelve days of difference on identical facts, and twelve days is £34,000 of receipts sitting in the wrong column of your forecast. Use the grossed-up version, or strip the VAT out of receivables, but do not mix the two.
Do the same on the other side. Your trade payables days tell you when your own money actually leaves, which is often later than your suppliers' terms too. And do it by customer where you can, because a single large client paying at 75 days will drag a blended average around and hide the fact that everyone else pays on time.
The Lumps, and Why Averages Hide Them
The second structural failure is smoothing. Somebody takes the annual figure for a cost and divides it by twelve. It is a reasonable thing to do in a budget and a destructive thing to do in a forecast, because the whole purpose of the forecast is to find the dates where the payments cluster.
The dates worth putting in your model, on the actual day rather than as a monthly twelfth:
- VAT. Due one calendar month and seven days after the quarter end. A March quarter is paid by 7 May, and so on. If you are on the annual accounting scheme or making payments on account the dates differ, so check your own return rather than assuming.
- Corporation tax. Nine months and one day after your accounting period end for companies with profits up to £1.5 million. Above that you pay by quarterly instalments, some of which fall due before the year has even finished.
- PAYE and National Insurance. The 22nd of the following month if you pay electronically, or quarterly if your average monthly liability is under £1,500. PAYE is easy to forget in a forecast because payroll software reports it separately from the net pay run, and it is a substantial sum for any business with employees.
- Self assessment payments on account, on 31 January and 31 July.
- Rent, insurance, business rates and annual software renewals, which are quarterly or annual and rarely fall in a convenient month.
- Pension contributions, monthly but usually a few weeks behind the payroll they relate to.
The VAT one deserves a paragraph on its own, because it does more damage than the rest combined. From the moment you invoice a customer to the moment you hand the VAT to HMRC, up to four months can elapse. During all of that time the money sits in your account looking exactly like your money, and it spends exactly like your money. For a business that is growing, it is worse than that. The bill you pay in January is calculated on last quarter's sales, which were higher than the quarter before, while this quarter's cash is tied up in the stock and labour needed to deliver the next round of work. Growth pulls cash out of a business and the VAT return is one of the main levers by which it does so. That is a working capital problem as much as a forecasting one, and we have covered the mechanics of it separately in our guide to working capital.
Monthly, Weekly, or Both
The thirteen-week cash flow forecast has become something close to a default recommendation, and we would push back on it gently.
Thirteen weeks is a turnaround convention. It comes from restructuring practice, where a lender or an insolvency practitioner needs to see whether a company can meet its liabilities over a defined short horizon, and thirteen weeks is a quarter expressed in a unit fine enough to show the troughs. It is an excellent tool for the situation it was designed for.
For a solvent UK business on quarterly VAT it has an awkward property: a thirteen-week window contains exactly one VAT payment, and depending on where you start it, that payment is either at the beginning, in which case the rest of the grid looks reassuring, or at the end, in which case it looks alarming. The same business can produce two very different-looking thirteen-week forecasts a fortnight apart, purely because of where the window landed.
What most owner-managed businesses actually need is two forecasts serving two purposes:
- A twelve-month monthly forecast for decisions. Can we afford the hire, the unit, the second van, the price rise? These are questions about the shape of a year, and monthly is the right resolution.
- A rolling weekly forecast, six to thirteen weeks out, for operations. Which invoices need chasing this week, which supplier payment can wait until Friday, do we need to talk to the bank. Weekly is the only resolution that answers those.
"Rolling" is doing real work in that sentence. A rolling forecast is re-cut every week or every month so the horizon stays constant, rather than a fixed annual model that quietly shortens until you are in March looking at four weeks of visibility. The rolling version takes twenty minutes a month once it is set up. The fixed version takes a day to rebuild each year and is stale for eleven months of it.
The Variance Column
Here is the habit that separates a forecast that gets better from one that stays wrong, and it is missing from almost every spreadsheet we are shown.
Next to each forecast column, put the actual. Next to that, the difference. Then, once a month, spend fifteen minutes on the four or five lines with the biggest gaps and write down why.
Do that for three months and you stop guessing about your own business. You learn that your receipts consistently land eight days later than you assume, that materials run 6 percent over, that December is always worse than you allow for. Those are not failures of the forecast. They are calibration data, and after a quarter of collecting them your forecast stops being an opinion and starts being a model with known error bars.
Nearly every forecast we inherit has been built once and never compared to outturn. The owner has no idea whether it is accurate to 5 percent or 40 percent, which means they have no idea how much of a buffer to hold against it. A forecast whose accuracy has never been measured is not a plan. It is a horoscope with a chart attached.
The comparison is also considerably easier than it sounds if your bookkeeping is up to date and your bank reconciliation is done. Cloud accounting will hand you the actuals; the fifteen minutes goes on the thinking, not the data entry.
Late Payment Is About to Change by Law
There is a reason to revisit your receipt assumptions in the next eighteen months, and it is not a commercial one.
The government published its response to the late payment consultation on 24 March 2026, and the Commercial Payments Bill was introduced to Parliament in May 2026. Late payment is estimated to cost the UK economy around £11 billion a year. The measures in the Bill include a cap on payment terms at 60 days, with the government's stated intention to reduce that to 45 days over five years, mandatory statutory interest on late payments so that it can no longer be contracted out of, a deadline for raising invoice disputes, and expanded powers for the Small Business Commissioner to investigate poor payment practice and adjudicate disputes outside court.
Two practical points follow. First, if a large customer currently pays you at 90 or 120 days, your forecast's receipt lag on that account is going to shorten materially once the cap commences, and that is a one-off cash windfall worth modelling before it happens rather than being pleasantly surprised by. Second, the rights that already exist are underused. Under the Late Payment of Commercial Debts (Interest) Act 1998, statutory interest on a late business-to-business debt runs at 8 percent above the Bank of England base rate, which at the current 3.75 percent base rate means 11.75 percent, and you can also claim a fixed sum for debt recovery costs of £40 on debts under £1,000, £70 on debts between £1,000 and £9,999.99, and £100 on debts of £10,000 or more.
Most small businesses never charge it, on the entirely rational grounds that they would like to keep the customer. Fair enough. But the right to charge it is worth mentioning in a second reminder letter, and it changes the tone of the conversation without anyone having to fall out.
Our View: Most Businesses Do Not Have a Cash Flow Problem
They have a timing information problem.
The money is there, or it will be. It arrives, it just arrives after the day it was needed, and nobody knew that in advance. The distinction matters because the two problems have completely different solutions. A genuine cash flow problem means the business does not generate enough cash and something structural has to change: prices, costs, break-even, the model itself. A timing problem is solved with a spreadsheet, a phone call to a customer in week two rather than week six, and an agreed facility arranged a month before it is drawn.
The forecast is how you tell which one you have. If the trough happens once a year in the same week and the balance recovers, that is timing. If the trough gets deeper every quarter and the recovery gets weaker, that is structural, and no amount of forecasting will fix it. Either way, the value of the exercise is not the document. It is the fifteen minutes a month spent looking at the gap between what you thought would happen and what did.
Common Cash Flow Forecasting Mistakes
- Building it monthly and stopping there. The month-end balance hides everything that happens in between, and everything that happens in between is what causes a bounced direct debit.
- Forecasting receipts on your payment terms. Measure your actual debtor days and use those. They are almost always worse.
- Mixing gross receivables with net turnover in the debtor days calculation, which inflates the figure by roughly the VAT rate.
- Leaving VAT out entirely, on the reasonable-sounding logic that it is not a cost. It is not a cost. It is a payment, and this forecast is about payments.
- Smoothing the lumps. An annual insurance premium divided by twelve is a budget entry, not a cash entry.
- Including depreciation. It is not a payment and it does not belong here.
- Forgetting the loan capital, which never appears in your profit figure and always appears on your bank statement.
- Forecasting only the good case. Run a version where your largest customer pays 30 days late and your best month comes in 15 percent under. If that version survives, you have a buffer. If it does not, you know what to fix.
- Never checking it against actuals, which means the same error repeats every month for a year.
How IAK Can Help
Cash flow forecasting is one of those jobs that is genuinely quick once the underlying records are clean, and close to impossible when they are not. Most of the work in a first forecast is not the forecasting. It is establishing what your real debtor and creditor days are, getting the aged receivables and payables listings to agree to the bank, and finding the annual and quarterly payments that have never been anywhere except a bank statement.
We build forecasts in Xero alongside monthly management reporting, so the actuals flow in automatically and the variance column fills itself. The standard shape is a twelve-month monthly model for planning and a rolling weekly view for the next quarter, with the trough and its date flagged on the front page, because that is the number that decides things. Where the forecast shows a funding requirement we will help you take it to the bank with working papers behind it, which is a very different conversation from asking for an overdraft in the week you need one.
If you are approaching a big contract, a quarter with two tax payments in it, or you simply cannot reconcile a profitable set of accounts with a bank balance that never seems to grow, get in touch. We work with agencies, retailers, professional firms and construction companies across North London, and a first forecast is usually a single afternoon's work that pays for itself the first time it catches something.
Sources
- Company Insolvency Statistics, Insolvency Service, GOV.UK, for the insolvency rate of one in 196 companies on the effective register in the twelve months to 31 May 2026, and the predominance of creditors' voluntary liquidations.
- VAT Returns: deadlines, GOV.UK, for the one calendar month and seven days submission and payment deadline.
- Pay your Corporation Tax bill, GOV.UK, for the nine months and one day deadline for companies with profits up to £1.5 million and the instalment regime above it.
- Late commercial payments: interest and debt recovery, GOV.UK, for the default 30-day and 60-day terms, statutory interest at 8 percent above base rate, and the £40, £70 and £100 fixed sums for debt recovery costs.
- Commercial Payments Bill: overview, GOV.UK, for the £11 billion annual cost of late payment, the 60-day cap on payment terms, mandatory statutory interest and the expanded Small Business Commissioner powers.
- Bank Rate, Bank of England, for the 3.75 percent base rate used in the statutory interest calculation.
- Late Payments Research, Small Business Commissioner, for average payment times and the value of overdue invoices owed to small businesses.