What do lenders look for when they read your accounts?
A credit officer reads your accounts to answer one question: will this debt be repaid, in cash, even if next year is worse than the plan. Everything else is a way of getting to that answer. It starts by rebuilding your EBITDA from the earnings it can evidence, not the ones you report. It checks how much of that EBITDA turns into cash, because interest and repayments are paid in cash and nothing else. It sizes the debt against your earnings, then tests whether the cash flow still services it in a soft year. It reads the three-year trend before the latest figure, the balance sheet for what secures the loan and what already ranks ahead of it, and the distance between your filed accounts and your management ones. This guide sets out each of those reads the way a lender runs it, at August 2026 rates, and ends with a worked path from a set of accounts to an approved £5m facility. It answers the path to a yes; how much you can raise is the separate question our guide to borrowing capacity works through.
Written for the borrower’s side of the table. This is general guidance, not advice on your specific facility. It deepens the shorter answers in our working guide.
A lender lends against the EBITDA that survives diligence, not the one you report.
They assess it by scrubbing the EBITDA line down to the earnings they can evidence, then lending against that number rather than the one in your accounts. Every add-back is tested on its own. Costs that will not recur, a relocation, a settled dispute, an aborted deal, survive the test and come back into earnings. Items labelled exceptional that reappear every year do not, and nor do owner adjustments a diligence accountant cannot stand behind: a director’s salary normalised to market only if the evidence supports the market figure, personal costs stripped out only where they are documented. Where the raise is large enough to warrant it, a quality-of-earnings review does this formally, testing how the profit was built, how repeatable it is, and how much of it becomes cash.
The number the scrub lands on is what the leverage multiple is applied to, so the arithmetic is unforgiving. At three times leverage, £200,000 of disallowed add-backs is £600,000 of borrowing capacity gone, and a borrower who presents a £2m EBITDA and leaves diligence with £1.75m has not lost £250,000, they have lost three-quarters of a million pounds of debt. Quality also means more than the level. A credit officer reads margin stability, the share of revenue that is contracted or repeatable, and any earnings that depend on a single customer or a single year, because a lender underwrites the durability of the profit, not the size of it in a good period. How that scrubbed figure sets the ceiling on what you can raise is the subject of our guide to how much your business can borrow.
The discipline that follows is to scrub your own EBITDA before a lender does. Build the bridge from statutory profit to adjusted EBITDA yourself, evidence every add-back, and drop the ones you cannot defend. A defensible number presented plainly borrows more than an optimistic one taken apart in diligence, because a lender prices what it cannot verify as risk.
A lender is repaid in cash, so it reads how much of your EBITDA becomes cash.
Because a lender is repaid in cash and not in profit, the ratio it cares about is how much of your EBITDA reaches the bank account. Cash conversion is free cash flow measured against EBITDA, and the gap between the two is where a facility that looked comfortable on the profit line comes under strain. Three things open that gap. Working capital absorbs cash as the business grows, because a bigger debtor book and more stock have to be funded before the profit on them is collected. Maintenance capital expenditure, the spend needed just to keep the business running, comes out before any of the earnings are free. And tax is paid on profit whether or not the cash has arrived. A business that converts around 90% of EBITDA into cash can carry materially more debt than one that leaks it into working capital and capex, whatever the leverage multiple would allow, because the serviceability test runs on cash rather than on an accounting line.
A credit officer does not take conversion on trust. It is reconciled from the accounts: the cash flow statement, the year-on-year movements in debtors, creditors and stock, and the trend in those movements across three years. A widening working capital absorption while revenue is flat is read as a warning, because it means growth or slower collection is consuming cash the debt was meant to be serviced from. Seasonality is read here too: a business whose cash swings hard across the year needs a committed line to absorb the trough rather than a fixed loan carrying idle cash, which is the structuring question our guide to the RCF and the term loan sets out. Strong, stable conversion is one of the quietest advantages a borrower can bring to a lender, and one of the easiest to evidence in advance.
Serviceability binds before leverage, and it is tested against a downside.
They test it against a downside rather than the plan: the debt has to be paid out of the cash the business would still generate in a soft year. The starting point is leverage, sized as net debt to EBITDA. A clearing or challenger bank lending on cash-flow terms will usually sit around two-and-a-half to three-and-a-half times a decent lower-mid-market EBITDA, a private-credit fund will stretch to four or four-and-a-half and occasionally to five for a strong credit, and the multiple is read as a ceiling rather than a target. But the multiple is only the first pass. Mid-market debt is floating-rate, priced over SONIA, which sat at 3.73% on 2 July 2026 against a Bank Rate of 3.75%, held at the Bank of England’s June meeting. Bank of England, Bank Rate. On a secured, sensibly geared bank deal the all-in cost lands near 6.75%, a margin of about three points over that reference, and at those rates the cash flow, not the multiple, is what caps the number.
Serviceability is measured through two covers. Interest cover is EBITDA over the interest bill, the simpler test and the one that matters most on a fund structure repaid as a bullet at maturity. Debt-service cover, or DSCR, is the harder one on bank debt: it sets the cash the business generates after tax and maintenance capex against the whole annual obligation, interest and scheduled repayments together, and it is the test amortising term loans live or die on. A credit officer runs both under a downside, not the base case. It flexes EBITDA down by a sensible margin, holds the interest rate at or above today’s level, and checks the debt still services with the cover above one. That is why a lender wants headroom of 25 to 30% between the base-case forecast and every covenant in the package: enough room that a bad quarter is a bad quarter rather than a breach.
The distinction from capacity is worth holding onto. How much a lender will extend is one question, answered by the lowest of the leverage, cash-flow and asset tests. Whether it will approve the facility is a different one, answered by whether that debt still services when the plan disappoints. A business can be well inside its capacity and still fail the read if the cover has no room in it, and the covenants that police that room, and what happens if they are tripped, are set out in our guides to loan covenants and to a covenant breach.
Direction of travel is read before the level, and an explained dip beats an unexplained spike.
They read the trend first, three years before the latest one, because a direction of travel underwrites forward cash better than a single strong period does. A lender is lending against next year and the years after it, so the shape of the last three matters more than the height of the last one. A rising, explicable trend borrows more than a higher flat line, and a single record year sitting above a modest history is discounted until it is understood, because a lender that leverages the peak of a good year has sized the debt off a number the business may not see again. Consistency is prized precisely because it is evidence: earnings that hold their shape through a cycle are earnings a credit committee can underwrite.
The story behind the numbers is read as closely as the numbers. A dip explained by a known one-off, a lost contract since replaced, a deliberate investment year, reads very differently from a dip with no account of it, and a borrower who narrates their own three years plainly is trusted more than one who lets the accounts speak for themselves. Forecast credibility is part of the same read: a management team that has hit its budget for three years is believed on the fourth, and a forecast that hockey-sticks away from a flat history without a mechanism behind it is marked back to trend. None of this is about presentation. It is about whether the forward cash the debt relies on is one a lender can stand behind, and the trend is the best evidence a borrower has that it is.
The balance sheet is read for what backs the loan and what already ranks ahead of it.
They look for two things: what is available to secure the loan, and what already ranks in front of it. Net debt is the starting point, not the headline borrowing, so a credit officer adds back the finance leases, any deferred consideration, pension deficits and tax arrears, then checks the charges register at Companies House to see what security is already granted and to whom. Existing fixed and floating charges tell a lender where it would stand in an enforcement, and an unexpected charge, or a pattern of late filings and county court judgments, changes the read before the earnings are even reached. The security a new facility can take is the mirror of that: a debenture over the company, fixed charges on the assets that can carry them and a floating charge over the rest, as our guide to debentures and charges explains.
Balance-sheet strength then does two jobs. It measures the cushion beneath the debt, the tangible net worth and the equity the owners have left in, because a lender is more comfortable behind a business the shareholders are still invested in than one they have already taken their money out of. And for an asset-heavy balance sheet it opens a second route to the money entirely: an asset-based lender sizes a facility from a borrowing base rather than a multiple of earnings, advancing around 80 to 90% of the eligible receivables plus smaller tranches against stock, plant and property, as our guide to asset-based lending sets out. At the conservative, owner-managed end a lender may also ask for personal guarantees where the balance-sheet cushion is thin, which is a read on the security position as much as on the borrower, and one our guide to personal guarantees covers from the borrower’s side.
Lenders start from your filed accounts and lend on the management ones.
Both, in sequence: the filed accounts anchor your credibility, the management accounts size the facility, and the distance between them is read as risk. Statutory accounts at Companies House, audited or not, are where a lender starts, because they are the independent, on-the-record version of the business and they set the base every later number is checked against. But a set of accounts filed nine months after the year end is a picture of a business that no longer exists, so the facility itself is sized on management accounts: the current-year monthly figures, the latest board pack, the aged debtor and creditor listings and the integrated forecast. The management numbers carry the weight, and the filed ones tell the lender whether to believe them.
The gap between the two is where a credit officer’s attention goes. Management accounts running well ahead of the last filed set are welcome only if the bridge between them holds together: a clear reconciliation, consistent policies, no line that appears in the management pack and vanishes in the statutory one. Thin, late or hand-built management reporting reads as a governance signal in its own right, because a lender is also underwriting the finance function that will produce the covenant certificates every quarter for the next five years. A business that can produce timely, reconciled monthly accounts on demand is telling a lender something the numbers alone cannot, and it is one of the things worth having in order before a process starts, because assembling it under diligence pressure is where a timetable slips, as our guide to how long a debt raise takes sets out.
How a £5m facility gets to a yes.
It gets approved when every lens a credit officer applies still holds under a downside, which is easier to show than any single heroic number. Take a business with £2m of scrubbed EBITDA refinancing its existing debt and funding a small bolt-on, asking for a £5m facility. These figures are illustrative arithmetic at August 2026 rates, not a quote; the point is the sequence of the read, not the decimals.
The read on the earnings
The EBITDA is £2m after diligence, with the add-backs evidenced and the exceptional items that recur left out, so the number survives the scrub intact. Around 90% of it converts to cash, the three-year trend rises steadily with no single record year doing the work, and revenue is spread across a book of repeat customers with no dangerous concentration. Before a structure is even drawn, the earnings read as durable and the cash as real.
The structure and the cost
At £5m the leverage is 2.5 times net debt to EBITDA, the bottom of the bank cash-flow range, so this is a clearing or challenger bank deal, not a stretch to a fund. It is structured as a £4m amortising term loan for the sum certain and a £1m committed revolver for the working-capital swing. At an all-in near 6.75%, interest on the drawn facility runs about £337,500 a year, leaving interest cover close to six times, comfortable on its own.
The downside test
The term loan amortises over six years, a longer profile than the usual three to five, about £667,000 of principal a year, so total debt service is near £1m. Against cash of roughly £1.4m after tax and maintenance capex, base-case debt-service cover is about 1.4 times. Now flex EBITDA down 20% to £1.6m: cash to service falls to about £1.1m, cover holds just above 1.1 times, and the debt still pays. Set the covenants to leave the 25 to 30% headroom the market expects between that base case and the test, take a debenture over the company, and the facility services through the downside rather than only in the plan.
That is the yes, and nothing in it is heroic. The earnings survive the scrub, the cash converts, the leverage sits at the bottom of the range, the cover holds under a stress and the security is clean. A credit officer approves the facility not because one number is impressive but because every lens holds when the plan does not, which is the whole of what bankable means.
We will read your accounts the way a credit officer will.
If you want to know how your accounts will read to a lender before you send them, a first conversation is confidential and costs nothing. We scrub the EBITDA, test the cash conversion and the cover, and tell you plainly where the read is strong and where a credit officer will push, so you fix it on your own terms rather than in someone else’s diligence. See how a mandate runs in how we work, or the full range of what we advise on in our services.