EBITDA Add-Backs: How to Tell a Legitimate Adjustment From an Aggressive One

EBITDA add-backs show up in interviews as two different questions wearing the same costume. One is definitional: what's the difference between reported EBITDA and adjusted EBITDA, and why does the adjustment exist at all. The other is judgment: given a specific expense, would you add it back or not, and why. Candidates who've only memorized the first question fall apart on the second, because the second is really asking whether you understand what adjusted EBITDA is for. This guide covers both, plus the buyer-versus-seller dynamic that turns a simple accounting concept into one of the most negotiated numbers in any deal.

What Adjusted EBITDA Is Actually Solving For

Reported EBITDA comes straight off the income statement: revenue minus operating expenses, with depreciation and amortization added back, before interest and taxes. It's accurate, but it's also a snapshot of one specific year's accounting, and that year can include things that have nothing to do with how the business will actually perform going forward under new ownership.

Adjusted EBITDA (also called normalized EBITDA) takes that reported figure and strips out the noise: expenses or gains that are one-time, non-operating, or specific to the current owner, so what's left is a cleaner estimate of the company's repeatable, run-rate earning power. That's the number buyers actually want to underwrite a valuation multiple against, because paying 8x for a temporarily depressed earnings figure, or a temporarily inflated one, means paying the wrong price for the business.

That's also why this question gets asked so often. Adjusted EBITDA sits directly underneath nearly every multiple-based valuation and every leverage calculation in a deal, so an interviewer testing whether you understand it is really testing whether you understand what a valuation multiple is actually being applied to.

The Add-Back Categories That Actually Come Up

Most legitimate add-backs fall into three buckets, and knowing which bucket an item belongs to is usually the fastest way to defend or challenge it.

Non-recurring items. One-time expenses or gains that genuinely won't repeat: a lawsuit settlement, restructuring or severance costs from a layoff, a one-time consulting engagement, moving expenses, a gain or loss on the sale of an asset. The test isn't whether management calls it "one-time," it's whether that category of cost actually shows up only once. Legal fees that appear every single year aren't a one-time expense just because each individual lawsuit is different.

Normalization items. These are recurring, but priced wrong relative to what a new owner would actually pay. The classic case is owner compensation: a founder paying themselves $650,000 when the market rate for someone running that business is $350,000. Only the excess, not the full salary, gets added back, since someone still has to do the job after the deal closes. The same logic applies to above-market rent paid to a landlord entity the owner also controls, or family members drawing a salary for a role they don't meaningfully perform.

Non-operating items. Income or expense that doesn't relate to the core business at all: interest income on a side investment, a gain on a legal settlement unrelated to operations, income from a property the company owns but doesn't use in the business. These get backed out because they're not part of what a buyer is actually paying to acquire.

A Worked Example

A company reports $8,000,000 of EBITDA for the trailing twelve months. A quality of earnings review turns up four items:

A $450,000 litigation settlement, a genuine one-time legal expense with no history of recurring. Add it back.

The owner draws $650,000 in salary; the market rate for a general manager running a business this size is $370,000. The excess, $280,000, gets added back. The remaining $370,000 stays in as a real operating cost.

The company pays $500,000 in rent to an LLC the owner also owns, against a market rate of $380,000 for comparable space. The $120,000 excess gets added back.

The company also booked a $200,000 gain on the sale of unused equipment. That's a one-time gain, not ongoing earning power, so it gets subtracted rather than added.

Adjusted EBITDA: $8,000,000 + $450,000 + $280,000 + $120,000 − $200,000 = $8,650,000.

That $650,000 swing looks small in isolation, but run it through a valuation multiple and it isn't. At a 7x multiple, the gap between reported and adjusted EBITDA is $4,550,000 of implied enterprise value, the same distinction that matters when you're moving between enterprise value and equity value in the first place. That's the entire reason this figure gets fought over instead of just accepted at face value.

The Buyer-Versus-Seller Fight

Every add-back favors the seller, since a higher adjusted EBITDA supports a higher purchase price at the same multiple. That's exactly why buyers don't take a seller's adjusted EBITDA schedule at face value and instead run their own quality of earnings review, testing every proposed adjustment against three questions: is there documentation behind it, did this category of cost actually appear in prior years, and will it genuinely not continue once the business changes hands.

Stock-based compensation is where this fight gets sharpest. Sell-side bankers frequently add it back on the logic that it's a non-cash expense. Buyers, especially strategics and credit-focused buyers, often push back hard, arguing that SBC is a real, ongoing cost of retaining employees that the buyer will have to replace with either continued equity grants or higher cash comp after close. There's no universally correct answer here, which is exactly why interviewers like asking about it: they want to see whether you can argue both sides rather than just recite a rule.

This is also where adjusted EBITDA diverges from a related concept candidates sometimes confuse it with: seller's discretionary earnings, or SDE, used mostly in small-business and lower-middle-market deals. SDE adds back the entire owner salary, not just the excess over market, because the assumption is a single owner-operator buyer will step directly into that role themselves. Adjusted EBITDA assumes a professional buyer who still has to pay someone market rate to run the business, which is why only the excess counts.

Follow-Up Questions Interviewers Ask

How would you push back on an add-back you don't trust? Ask for documentation, then check whether that expense category shows up in the two or three prior years. If a "one-time" consulting fee appears every year, it's an operating expense wearing a disguise, not a real add-back.

Is stock-based compensation a legitimate add-back? There's a real argument on both sides. It's genuinely non-cash, but it's also a genuine cost of keeping the team the buyer is paying to acquire. Strong candidates name both positions instead of picking one and stopping there.

What's the difference between adjusted EBITDA and SDE? SDE adds back the full owner salary and assumes an owner-operator buyer; adjusted EBITDA only adds back the excess over market rate and assumes a buyer who still has to staff the role.

Would you trust an EBITDA figure that comes entirely from management's own adjustments, with no independent review? No. That's precisely what a quality of earnings engagement exists to test, and it's a fair thing to say directly in an interview.

Where Most Candidates Lose Points

The most common mistake is treating every item labeled "non-recurring" as automatically legitimate without checking whether it actually recurs. The second is forgetting that normalization items like owner compensation only get partially added back, not the whole expense. The third is not being able to argue the other side of a gray-area item like stock-based compensation when an interviewer pushes back, since the question is usually designed to see if you'll defend a position under pressure rather than fold immediately.

Where to Go From Here

The EBITDA add-backs question rewards the same preparation as any judgment-based technical: work through real add-back scenarios until sorting non-recurring from normalization from non-operating becomes automatic, not something you have to reason through live in the room. It also pairs directly with how EBITDA gets used downstream, whether that's setting the entry multiple in a leveraged buyout or sizing the debt a target can support against its earnings base. Offcycle has this question, along with the surrounding valuation and quality-of-earnings concepts it tends to come up alongside, built into structured flashcards you can work through by topic and difficulty.

Beyond flashcards, you get custom practice sets built around whatever you're weakest on, quizzes that test the same material a different way, and a readiness score that tracks your progress across accounting, valuation, and M&A topics so you know exactly where you stand before you walk into the room.

The trial is free for 7 days and doesn't require a card to start.