Quality of Earnings Explained: Adjusted EBITDA & Add-Backs

Key takeaways

The "E" in EBITDA is the most contested number in a deal because the multiple is applied to it. At a 10x entry multiple, every £1m of disputed EBITDA is £10m of enterprise value — so a Quality of Earnings (QoE) report is not an accounting exercise, it is a negotiation over price
A QoE rebuilds management's adjusted EBITDA from the bottom up: it starts at statutory EBITDA, accepts the add-backs that are genuinely one-off, haircuts the "run-rate" adjustments that assume the future, and rejects the recurring costs dressed up as exceptional. The output — sustainable or maintainable EBITDA — is the number the buyer actually underwrites
The fight does not end at EBITDA. Items the buyer concedes on earnings — deferred revenue, provisions, earn-outs, deferred maintenance capex, factoring — reappear as debt-like items in the net-debt bridge and shrink the equity cheque pound-for-pound, which is where a deal quietly leaks value after the multiple is agreed
In interviews, "what add-backs would you challenge?" is a real question with real answers: recurring "restructuring", unsubstantiated run-rate synergies, owner add-backs without evidence, and any cost that is capitalised to flatter the margin

The Multiple Is Applied to a Number Management Built — So Diligence Rebuilds It

A buyout price is a multiple times EBITDA, and the multiple gets all the attention. The number it multiplies gets far less, which is exactly why it is where the money is made and lost. Management has every incentive to present the highest defensible EBITDA, and the figure in a CIM is almost never the statutory one — it is "adjusted", and the adjustments are the argument.

Quality of Earnings is the buyer's answer to that argument. A QoE report — usually run by the transaction-services or financial-due-diligence team of an accounting firm, commissioned by the buyer — takes management's adjusted EBITDA apart line by line and rebuilds it into a number the buyer is willing to pay a multiple on. It is the most consequential piece of diligence on most deals, because it does not just describe the business; it resets the price.

The reason it matters so much is leverage in the literal sense. The same QoE work that strips £3m off an agreed EBITDA strips £30m off the enterprise value at a 10x multiple, before a single line of the SPA is negotiated. Understanding how that number is built is the difference between quoting a multiple and understanding a deal.

The Bridge: Reported £18m, Management's £25m, Sustainable £22m

The cleanest way to see a QoE is as a bridge with three rungs. Take a mid-market business and walk it from the audited number to the one the buyer underwrites.

1. Statutory EBITDA: £18.0m. The number that falls out of the audited accounts — operating profit with depreciation and amortisation added back. It is the only figure with an auditor's signature near it, and it is the floor the whole negotiation builds up from.

2. Management's adjusted EBITDA: £25.0m. The CIM number. Management has layered £7.0m of add-backs on top — one-off costs, owner adjustments, and run-rate items — to present the "true" earnings power. This is the number the seller wants the multiple applied to.

3. QoE-adjusted (sustainable) EBITDA: £22.0m. Diligence accepts roughly £4m of the £7m, haircuts or rejects about £3m, and lands at the maintainable figure. This is the number the buyer underwrites — and the gap to management's £25m is the heart of the price negotiation.

£30m Enterprise-value gap between management's £25m and the QoE's £22m at a 10x multiple. The £3m of rejected add-backs is not an accounting footnote — it is the size of the argument about price

The bridge reframes the whole exercise. The buyer is not trying to prove management dishonest; it is trying to establish how much of the £7m of adjustments represents earnings the business will actually repeat. That question — what recurs — is the one every add-back has to answer.

Which Add-Backs Survive, and Which Die in the Databook

Add-backs are not equal, and a QoE sorts them into three buckets by how much faith the future has to carry. The first bucket survives, the second is haircut, the third is rejected.

Bucket	Typical examples	Diligence treatment
Genuine one-offs	ERP/system implementation, one-time legal settlement, M&A transaction costs, a single restructuring, owner's above-market salary normalised to market	Survive if substantiated with invoices and shown not to recur. Owner normalisation is legitimate in a founder business — but works both ways if the owner was underpaid
Run-rate / pro-forma	Full-year effect of a mid-year price rise, annualised savings from a cost programme, "synergies" from a recent bolt-on not yet delivered	Haircut. These assume the future. Buyers accept the portion already evidenced in actuals and discount the rest — an unrealised synergy is a plan, not earnings
Recurring dressed as exceptional	A "restructuring charge" that appears every year, costs capitalised that should be expensed, deferred maintenance capex, pulled-forward revenue, stretched supplier terms	Rejected. A cost that recurs is operating cost. This bucket is where management's number is most aggressive and where diligence claws the most back

Insider tip The single most common rejected add-back is a recurring "exceptional" item. If a £1m restructuring charge shows up in three consecutive years, it is not restructuring — it is the cost of running the business, and it belongs below the EBITDA line in name only. The QoE test is not "did management call it one-off?" but "does it recur?" — and the databook of monthly actuals is where that gets settled.

Quality of Revenue Decides Whether the EBITDA Is Worth a High Multiple

A QoE does not stop at the level of EBITDA; it interrogates its composition, because two businesses with identical earnings can deserve very different multiples. The question underneath the number is how durable it is.

Three tests do most of the work. Recurring versus one-off revenue — a software business with 90% renewing subscriptions earns a higher multiple than a project business that rebuilds its order book every year, even at the same EBITDA. Customer concentration — if the top three customers are 40% of revenue, the earnings carry a fragility the headline figure hides, and diligence will model the loss of the largest. Margin sustainability — a gross margin that jumped two points last year on a one-off pricing move is not a new baseline, and the QoE normalises it back.

This is why a QoE and the comparable multiple are joined at the hip: the adjusted EBITDA sets what the multiple is applied to, and the quality of that EBITDA helps justify what the multiple should be. Get the composition wrong and you can pay a recurring-revenue multiple for project earnings.

The Quiet Leak: Debt-Like Items Move From EBITDA Into the Net-Debt Bridge

The most expensive part of a QoE is the part students never see coming. Even once EBITDA is agreed and the multiple fixes the enterprise value, the equity cheque is enterprise value minus net debt — and a QoE keeps finding things to put in the net-debt column. These are debt-like items, and they reduce the equity proceeds pound-for-pound.

What a QoE reclassifies as debt-like Deferred revenue — cash collected for services not yet delivered, a real obligation the buyer inherits. Earn-outs and deferred consideration from the target's own past acquisitions. Provisions for litigation, warranties, dilapidations or restructuring. Deferred maintenance capex — investment the seller skipped to flatter cash, which the buyer must now spend. Factoring and reverse factoring, unpaid tax, capex creditors, and an underfunded pension. Each one shifts value out of the equity cheque even though EBITDA never moved.

This is why a deal can feel agreed and then lose several million at signing. A seller who has conceded nothing on EBITDA can still surrender real value through the EV-to-equity bridge, because the QoE has moved a deferred-revenue balance or a pension deficit into net debt. The earnings argument and the balance-sheet argument are fought separately, and the second is where the late surprises live.

The Working-Capital Peg: Where the Last Few Million Move at Completion

One more QoE output decides the final cheque: the normalised working-capital target, or "peg". Because working capital swings month to month, a sale sets a normal level the seller must deliver at completion; if actual working capital comes in below the peg, the buyer adjusts the price down, and if above, up.

A QoE builds the peg by averaging working capital across a cycle and stripping out the window-dressing — the supplier payments delayed and the receivables chased hard in the run-up to a sale. Set the peg too low and the buyer overpays for a balance sheet the seller drained on the way out; set it right and the seller cannot quietly extract cash by squeezing working capital before close. It is unglamorous and it is where the last few million change hands.

Common mistake Treating the QoE as confirmation rather than negotiation. Candidates assume diligence verifies a number that is already set. It does the opposite — it produces the number, and that number resets the price, the SPA mechanics, and the equity cheque. A buyer who waits for the QoE to "check the figure" has misread who decides what the figure is.

How This Is Tested: "Walk Me Through the Add-Backs You'd Challenge"

QoE is interview-grade precisely because it sits where technical knowledge meets judgement. The weak answer recites that adjusted EBITDA strips out one-off items. The strong answer takes a side on which adjustments are real.

Interview framing Asked what you would challenge in a management EBITDA, do not hedge — name the buckets and rank them. Lead with recurring costs labelled exceptional: a restructuring charge in every year is operating cost, full stop. Then the run-rate adjustments: accept the price rise already in the actuals, discount unrealised bolt-on synergies to near zero because they are a plan. Then the owner add-backs: legitimate only with evidence of the market rate. Close on the balance sheet — point out that even with EBITDA agreed, you would reclassify deferred revenue and any deferred maintenance capex as debt-like, because that is where value leaks after the multiple is fixed. That last move is what tells the interviewer you have seen a real bridge, not just a textbook one.

The Verdict: Adjusted EBITDA Is a Claim, and the QoE Is the Cross-Examination

Management's adjusted EBITDA is not a fact — it is an argument about which costs are real and which earnings will repeat, presented by the party with every reason to round up. A Quality of Earnings report is the cross-examination: it concedes the genuine one-offs, discounts the adjustments that assume a future not yet in the numbers, throws out the recurring costs wearing an exceptional label, and then keeps going into the balance sheet to reclassify the obligations that shrink the equity cheque.

For a student, the lesson is to stop treating the EBITDA in a CIM as the starting point and start treating it as the seller's opening bid. The number that decides the price is the one the buyer rebuilds, and the skill is knowing which of management's seven million pounds of add-backs would survive a databook and which would not. That judgement — not the multiple — is where the deal is priced.

Careers: The QoE Databook Is Where a PE Associate Lives on a Live Deal

On a buyout, reconciling management's EBITDA to the QoE bridge is not a side task — it is the deal. An associate spends diligence inside the provider's databook, pressure-testing each add-back against monthly actuals, building the net-debt and working-capital bridges, and feeding the sustainable EBITDA into the LBO model that sets the offer. The investment case is, in large part, an argument about whether the QoE number can grow — so getting it right is the foundation everything else stands on.

Anyone can add up management's add-backs and accept the total — that is arithmetic. The judgement the report cannot make for you is the one that prices the deal: which of those adjustments describe earnings the business will actually repeat, and which describe a year the seller chose to present. Sorting the durable from the dressed-up is the entire skill — and it is the part a spreadsheet, and a CIM, will never hand you.

Take Your Preparation Further

Quality of Earnings sits at the junction of accounting, valuation and deal mechanics, so read it alongside the pieces it connects. Start with How to Read a CIM, since the adjusted EBITDA a QoE dismantles is the one a CIM presents, then work through EBITDA to Free Cash Flow to see why the deferred maintenance capex a QoE reclassifies matters so much to the cash the business actually throws off. For where the debt-like items land, the EV-to-Equity Bridge shows how net-debt adjustments shrink the equity cheque, and Trading Comps and Precedent Transactions covers why an aggressive add-back understates every multiple in the set. For where the diligence sits in the deal, The M&A Process.

For the diligence and process map a QoE plugs into, download our free M&A Process Cheat Sheet, and for the full set of PE interview questions and model answers — including how to challenge a management EBITDA under pressure — see the PE Interview Masterclass.

Ready for personalised feedback? Book a 1-on-1 mentoring session with an experienced IB/PE professional.

Quality of Earnings: How Buyers Pull Apart Adjusted EBITDA — and Which Add-Backs Survive Diligence