Which EBITDA Add-Backs Survive Diligence (and How to Build a Schedule a Buyer Won't Cut)

Every add-back you propose is a claim that a cost won't happen again. The buyer's accountants will spend three to six weeks testing that claim, and when they reject one, the price moves by the full multiple.

Papermark, a data room software vendor, estimates that buy-side providers commonly disallow 10% to 30% of the add-backs sellers put forward (source). It's one vendor's estimate, not an audited dataset, but it matches what most deal teams will tell you over coffee. At a 6x multiple, every $100,000 of rejected add-backs costs $600,000 of headline price.

So the question isn't whether to take adjustments. It's which ones hold up.

What is an EBITDA add-back?

An add-back is an adjustment that increases reported EBITDA by removing a cost (or, less often, a loss) that you argue isn't part of the normal run-rate of the business. Adjusted EBITDA is reported EBITDA plus the add-backs that survive scrutiny. The word "survive" is doing most of the work in that sentence.

The test every add-back has to pass

A buyer's team asks three questions of each line, in roughly this order.

Is it real? Can you show the invoice, the contract, the payroll record? A number from the owner's memory fails here, no matter how true it is.

Is it non-recurring? One legal settlement is a one-off. Legal fees that show up in 11 of the last 12 months are a cost of doing business with a new label on it.

Does it actually go away? This is the one sellers miss. If the owner's brother-in-law is on payroll at $140,000 and the buyer needs someone to do his job anyway, you haven't removed a cost. You've moved it.

Anything that fails one of the three gets cut or haircut. Anything that passes all three, with paper behind it, usually stays.

Add-backs that usually hold

Documented personal expenses. The truck, the club membership, the family cell plans. Boring, auditable, and clearly not coming with the business. Keep the receipts and show them by month.

True one-time events. A lawsuit settled and closed, a facility move, a failed software implementation that was written off. The key word is "closed." If the matter is still open, expect the buyer to ask what the next bill looks like.

Owner compensation normalization. Replace what the owner actually takes with a market-rate salary for the role. People forget this one cuts both ways. If the founder pays himself $90,000 to run a $30M company, the normalization is a deduction, and buyers will find it.

Related-party rent at market. If you rent a building from an entity you own at twice the going rate, restating it to market is defensible. Bring a broker opinion or comps. Don't just assert it.

Add-backs that usually get cut

Costs that repeat every year with a different label. "Professional fees, non-recurring" for the third year running. "Recruiting costs, one-time" at a company that's growing 20% and hires constantly. If it appears on the P&L in every period, it's operating cost.

Run-rate and pro forma revenue adjustments. Annualizing a customer you signed in September. Adding back the full-year effect of a price increase that hasn't fully rolled through. These are the most contested items on any schedule, because they ask the buyer to pay for revenue the business hasn't earned yet. Some survive, with contracts attached. Most get discounted.

Stuff the buyer will have to replace. The bookkeeper you let go, the CFO role the owner covered by working weekends. If someone has to do the work after close, the cost belongs in the number.

Anything with no documentation. It doesn't matter how reasonable it sounds.

A worked example (hypothetical)

Say a $22M distribution company reports $3.0M of EBITDA and proposes $720K of add-backs, for a $3.72M adjusted figure.

Proposed add-back	Amount	Likely outcome
Owner personal expenses (documented)	$85K	Holds
Settled lawsuit (closed, with settlement letter)	$160K	Holds
Above-market related-party rent (with broker comps)	$120K	Holds
"One-time" consulting, 3 straight years	$140K	Cut
Annualized pricing change, half-implemented	$110K	Cut by half, maybe more
Uncovered finance role, owner-performed	$105K	Reversed, cost added in

Sum it up and you land near $3.3M, not $3.72M. At 6x, that roughly $400K gap is about $2.4M of headline price. The schedule wasn't dishonest. It was just built without asking the buyer's three questions first.

How to build the schedule before anyone asks

Start 6 to 12 months before a sale or refinancing, not during exclusivity. The work is mostly bookkeeping discipline.

First, tag unusual items as they happen. A "non-recurring" flag in the general ledger, with a note on what it was, costs nothing in the moment and saves weeks later. Trying to reconstruct why a $60,000 charge hit in March 2025 is miserable.

Second, keep a running add-back ledger: item, amount, month, support, and a one-line argument for why it won't recur. Review it every quarter. If something you flagged last year shows up again, remove it yourself before a buyer does.

Third, run your own skeptical pass. Rank every line by how contested it is, and pressure-test the top five. Ask what a buyer would say, then ask whether you'd accept the answer.

Fourth, be conservative on purpose. A schedule with $500K of clean, documented add-backs is worth more than one with $800K that has $300K of arguments in it, because the first one makes the rest of your numbers more believable. Lenders read the same schedule for covenant calculations, so credibility compounds in both directions.

The real point

Adjusted EBITDA isn't a number you negotiate into existence. It's a record of how well you know your own cost structure. Companies that tag, document, and challenge their adjustments every quarter go into diligence with a schedule that's mostly already agreed. Companies that wait scramble, and the buyer's team writes the schedule for them.

Plametrix runs this kind of ongoing add-back tracking as part of an outsourced FP&A service for PE-backed and growing companies between $5M and $50M in revenue, so the schedule is ready long before the process starts.