How to build a cash flow forecast a lender will actually believe

Ask any commercial lender what they read first and you get a version of the same answer: the bank statements, then the forecast, then the story that is supposed to connect the two. The bank statements are fact. The forecast is opinion. The whole assessment turns on whether the opinion looks like it was written by someone who understands their own business, or by someone who worked backwards from the number they wanted to borrow.

That second version is easy to spot. Sales rise in a straight line from the month the money lands. Costs stay flat. There is no VAT payment in the quarter there should be one. Every customer pays on the day the invoice falls due, which has never happened in the history of trade. A lender does not need to disprove any of it — they just downgrade everything else you have said.

Here is how to build the other kind.

Two forecasts, not one

You need a 13-week forecast and a 12-month forecast, and they do different jobs.

The 13-week is weekly, detailed, and denominated in actual payments and receipts. It answers the only question a lender genuinely worries about in the short run: does this business run out of money before it can pay me? It is also the single most useful management document a small business can own, borrowing or not.

The 12-month is monthly, and it answers whether the loan repayment is affordable once the immediate problem is solved. It carries the seasonality, the tax dates, the repayment schedule of the facility you are asking for. Lenders want to see the debt service sitting inside your own numbers, not bolted on afterwards.

Start from the bank balance, not the sales target

A credible forecast opens with the actual cleared balance in the account on the day you built it, and every week after that is that number plus receipts minus payments. That sounds obvious. A surprising number of forecasts start from a revenue assumption and never reconcile to a bank balance at all, which makes them profit forecasts wearing the wrong hat.

Then build three schedules underneath it and let them feed the top line.

**Receipts.** Take your actual debtor list, invoice by invoice, and date each one by when that customer actually pays — not by the terms on the invoice. If a client is reliably 45 days on 30-day terms, put it in at 45. Do the same for the work you expect to win: pipeline value, multiplied by an honest conversion rate, dated at when the cash arrives rather than when the job is done.

**Payroll and the taxes on it.** Wages, PAYE and National Insurance, pension contributions, and the dates each leaves the account. This is the most fixed cost you have and the one you can least afford to get wrong by a week.

**Everything else.** Suppliers on their real terms, rent and rates, insurance, subscriptions, loan and finance payments, and the tax payments — VAT quarters, corporation tax nine months and a day after year end, payments on account in January and July if you are a sole trader or in partnership. Put them in on their dates. A forecast with no tax payments in it tells a lender you have been surprised by tax before and will be again.

A lender is not testing whether your forecast comes true. They are testing whether you are the kind of owner who knows what next Thursday looks like.

A worked example

The figures below are illustrative, but the shape is the thing worth copying. Say a fabrication business with a £6,200 cleared balance wins a £48,000 order. Materials are £26,000, payable 30 days from delivery. Subcontract labour is £7,500, paid weekly as the work happens. The customer pays on 60-day terms and has historically taken 68.

Lay that out week by week and the picture is unambiguous. Materials leave the account in week five. Labour drains out across weeks two to seven at roughly £1,070 a week. The £48,000 arrives in week fourteen. The ordinary running costs of the business carry on underneath all of it. The forecast bottoms out at about minus £14,800 in week nine and does not recover until week fourteen.

That number — the peak deficit, and the week it lands in — is the answer to "how much do you need and what for?" You are not asking for £48,000 because that is the order value. You are asking for a £20,000 facility to cover a £14,800 hole for five weeks, with headroom, and it repays itself out of a receipt you can point to. That is a fundable request. "We'd like £50,000 for growth" is not.

If that arithmetic is unfamiliar, the longer version is in how much funding do you actually need.

The assumptions page is what makes it believable

Every forecast should have a single page listing the assumptions behind it: debtor days used and why, conversion rate on pipeline, price and wage inflation, seasonality, anything that changes in the period. Three or four lines each.

This page does more work than the spreadsheet. It shows the numbers came from somewhere. It also protects you: when reality diverges — and it will — you can point to which assumption moved rather than looking like the whole thing was fiction. Lenders reading dozens of applications a week notice immediately when this page is missing.

Show them the version where it goes wrong

Include a downside case. Receipts 15% lower, or your largest customer paying 30 days later than usual, or the order slipping a month. Show what the balance does and what you would do about it — pause a hire, delay the van, extend a supplier, draw on the facility you are asking for.

Owners resist this because it feels like arguing against yourself. It does the opposite. Presenting only the good case makes an assessor build their own downside, privately, using pessimistic assumptions you never get to see. Presenting your own means the conversation happens on your numbers.

The three mistakes that get spotted instantly

Forecasts that ignore VAT. If you are registered, your receipts include VAT you are holding on someone else's behalf, and it leaves in a lump every quarter. Model gross figures in and out, with the VAT payment as its own line.

Forecasts where the loan has no cost. The repayment, the arrangement fee and any personal guarantee implications belong in the model. If the business only works without the repayment in it, the answer was never a loan.

Forecasts that contradict the accounts already filed. Whatever you send will be read next to your filed accounts and your last six months of bank statements. If the forecast assumes a margin you have never achieved, say why — new pricing, a changed supplier, a dropped loss-making line — and show the evidence. Unexplained optimism reads as carelessness, and carelessness prices your borrowing. The rest of the file a lender builds on you is covered in what a lender actually asks for.

Build it monthly, whether you are borrowing or not

The businesses that get funded quickly are almost always the ones that already had the forecast. They are not producing a document for the bank; they are sending the one they use to run the place, which is why it holds up under questioning.

An hour a month, rolling the 13-week forward one week at a time and marking last week's forecast against what actually happened, builds something no application form can fake: a track record of predicting your own cash correctly. If the difference between cash and profit is still fuzzy, start with cash flow vs profit and build from there.

Common questions

How far ahead should a cash flow forecast go for a loan application?

Send both a 13-week weekly forecast and a 12-month monthly one. The weekly version shows you can survive the period before the funding does its work, and it is the document that answers a credit assessor's real worry about short-term liquidity. The monthly version shows the loan repayment is affordable across a full trading year including seasonality, tax payments and the facility's own cost. If a lender asks for a longer horizon, three years monthly is normal for larger or asset-backed lending, but the detail should stay in the first twelve months. Anything beyond that is a modelled trend rather than a forecast, and experienced assessors read it as such.

What is the difference between a cash flow forecast and a profit forecast?

A profit forecast records income and costs when they are earned or incurred. A cash flow forecast records money when it actually moves. The gap between them is where profitable businesses fail. A £48,000 order delivered in March is profit in March, but if the customer pays in June and the materials were paid for in April, the cash position is deeply negative for two months of a highly profitable quarter. Lenders look at profit to judge whether the business model works and cash flow to judge whether you can pay them next month. You need both, and they should reconcile to each other.

Do lenders check whether the forecast actually came true?

On any ongoing facility, yes. Many lending agreements require management accounts quarterly or monthly, and a relationship manager will compare what arrived against what you forecast when you applied. Consistently over-promising damages your position when you next need a variation, a covenant waiver or more money, because the forecast is the only forward-looking evidence you can offer and its credibility is now discounted. Businesses that forecast conservatively and slightly beat it are treated very differently from ones that forecast ambitiously and miss. This is a repeated game, and the reputation you build across it is worth real money in pricing.

Should the forecast include VAT and other taxes?

Yes, and their absence is one of the fastest ways to be marked down. If you are VAT registered, receipts arrive including VAT and payments leave including VAT, so model everything gross and add the quarterly VAT payment as its own line on the date it is due. Do the same for PAYE and National Insurance monthly, corporation tax nine months and one day after your year end, and self assessment payments on account in January and July for sole traders and partners. These are large, dated and entirely predictable. A forecast that omits them is not showing the business the lender is being asked to fund.