What is Phantom EBITDA?

Phantom EBITDA is a term coined by Precision Dental Analytics for profit that appears real on a dental practice's P&L but will not survive institutional due diligence. It has three faces. The accounting face is normalization: add-backs that fail, owner compensation below market, related-party rent, 1099 contractors doing W-2 work. The operational face is durability: EBITDA that is real today but depends on the seller, one provider's exam chair, or a marketing funnel that leaks. The clinical face is defensibility: revenue that was collected, cleanly, from coding patterns and non-billable production that a payer's payment-integrity review will later unwind. Every dollar removed is multiplied against the multiple.

What are the three faces of Phantom EBITDA?

Accounting, operational, and clinical, and each is found by a different person on the buyer's side. The accounting face is found by the financial Quality of Earnings team in about week two; it comes off the P&L and cuts the price. The operational face is found by the buyer's operations lead or the platform's integration team; it never touches the P&L and rewrites the terms instead, through a longer employment agreement, a larger earnout, and a bigger holdback. The clinical face is found by clinical diligence if the buyer runs it, and by the payer if the buyer doesn't; it cuts the price where it is caught and survives close as a liability where it isn't.

How does a buyer normalize a dental practice's EBITDA?

The buyer's analyst rebuilds the seller's number line by line. In the practice in this article, two of the seller's adjustments held: $80,000 of discretionary expense and $40,000 of one-time items came back in. Three went the other way: the owner was drawing $50,000 below what a replacement dentist would cost, 1099 contractors doing W-2 work were reclassified with payroll burden attached (about $155,000), and rent paid to the owner's own building entity was roughly $40,000 under market. Reported EBITDA of about $720,000 became about $595,000. At a 7x multiple, that $125,000 adjustment removed roughly $875,000 of price before anyone opened a chart.

Can dental practice EBITDA be collected and still not be defensible?

Yes, and it is the most expensive version. The practice in this article had insurance AR days of 21, a 93% clean-claim rate, under $500 of rejected claims all time, and insurance AR over 90 days under 16%, and it passed every revenue-cycle screen a sponsor runs. Underneath were $680,000 of coding-pattern exposure across six code families with every claim paid, $1.5 million of production under practice-created non-billable codes that never became claims, an insurance collection rate of 107% of net insurance production, and refunds of $390,000 in the practice software against $150,000 on the P&L. Payers pay clean claims first and audit the pattern later. In a transaction, later means after close, through the escrow and the indemnity.

Why do 85% of dental practice deals get re-priced between LOI and close?

Because the three faces of Phantom EBITDA land in sequence and the seller usually prepared for only the first. The accounting face lands in the QoE and the price drops. The operational face lands in the structure and the terms tilt toward more earnout, longer employment, and a bigger holdback. The clinical face lands last, sometimes after the seller has spent the money, and the escrow gets claimed. SRS Acquiom puts post-LOI price adjustments at 85% of deals. The buyer has three people looking at three faces; the seller has a CPA looking at one.

How far before a sale should a dentist run a Quality of Earnings on their own practice?

Twelve to twenty-four months before going to market, because the remedy for all three faces is early rather than complicated. The accounting face is a documentation exercise: every add-back with an invoice behind it and owner compensation reset to market for a full year. The operational face is a systems exercise: associates in the exam chair alongside the owner, presentation and reappointment processes documented and measured, and the assist matrix built from the practice software. The clinical face is a clean-up exercise: non-billable codes mapped to CDT and claimed, ratios benchmarked and outliers documented or corrected, credit balances refunded, and the ledger reconciled to the claims. Done early, all three become things the buyer pays for. Done in the ninety days after an LOI, they are the reasons the deal changed.

Financial Analytics

The Three Faces of Phantom EBITDA

James DeLuca September 21, 2026 8 min read

Every dental practice that goes to market carries some Phantom EBITDA. That isn’t an accusation. It’s arithmetic. A practice is run to pay its owner and its taxes, not to survive an institutional audit, and the two jobs produce different numbers.

The mistake sellers make is thinking Phantom EBITDA is one thing. It’s three. Each face is found by a different person on the buyer’s side, at a different point in the deal, and each one prices differently. Sellers who’ve heard of the first one walk into the second and third with no idea they exist.

I’ll use one practice to show all three. Ten million dollars of income, multi-provider, clean books, a CPA who did his job. Reported EBITDA of about $720,000.

The First Face: The Accounting Face

This is the one everybody knows about, and the one the buyer’s financial Quality of Earnings team finds in week two.

The seller presents EBITDA with add-backs. The buyer’s analyst rebuilds it. In this practice, two of the seller’s adjustments held: $80,000 of discretionary expense and $40,000 of one-time items came back in. Three went the other way. The owner was drawing $50,000 below what a replacement dentist would cost. The 1099 contractors doing W-2 work got reclassified, with the payroll burden attached, about $155,000. The rent paid to the owner’s own building entity was roughly $40,000 under market.

Reported: about $720,000. Normalized: about $595,000. The largest related-party subsidy in an owner-operated practice is almost always the owner, and it was here too.

That $125,000 isn’t the loss. The loss is $125,000 multiplied by the multiple. At 7×, the accounting face took roughly $875,000 off the price before anyone looked at a chart.

This face comes off the P&L. It cuts the price. It’s the face a good CPA can partially defend, and the only one most sellers prepare for. It is also the face The EBITDA Illusion was written about, back when I thought it was the only one.

The Second Face: The Operational Face

The second face never touches the P&L. The EBITDA is real. The question is whether it’s durable, and the buyer answers that question with structure, not price.

In the same practice, the owner performed 4.4% of the exams and originated 24.6% of everything the practice diagnosed. His treatment plans averaged $10,632; the associates’ averaged $875 to $1,393. The revenue by provider said low key-person risk. The diagnosis by provider said the practice’s comprehensive pipeline runs through one exam chair. I wrote about him last week as The Point Guard.

The broken-appointment rate was 26%. The practice added 7,544 new patients in a year and retained 1,412 of them into a treatment trajectory, on $1.88 million of marketing, the funnel from The Acquisition Addiction. None of that is an add-back. All of it is a question about whether next year’s EBITDA looks like this year’s without the seller in the building.

A financial QoE doesn’t find this face. The buyer’s operations lead does, or the platform’s integration team does, and they don’t reduce the price for it. They rewrite the terms. The employment agreement gets longer. The earnout gets larger and its metric gets tied to the production they now know depends on one person. The holdback grows. The seller reads the headline number, sees it didn’t move, and doesn’t notice that the part of it he’ll actually receive did.

This face rewrites the terms. The seller’s CPA can’t see it because it isn’t in the ledger. The broker won’t show it because it isn’t in the prospectus.

The Third Face: The Clinical Face

The third face is the one I’ve spent the last month arguing about in public, because it’s the one the industry’s diligence process was never built to open.

The practice’s revenue cycle was clean. Insurance AR days of 21. A 93% clean-claim rate. Under $500 of rejected claims, all time. Insurance AR over 90 days under 16%. By every screen a sponsor runs on day one of a data room, this practice passes.

Underneath: $680,000 of coding-pattern exposure across six code families, every claim clean, every claim paid. $1.5 million of production booked under practice-created, non-billable codes that never entered the claims process, so it could never be denied and never aged. An insurance collection rate of 107%, $4.18 million collected against $3.90 million of net insurance production, part of it AR burn-down and the rest payer overpayment parked as credits and refunds. Refunds of $390,000 in the practice software against $150,000 on the P&L.

That is EBITDA that was collected, and clean, and still not defensible. Payers pay clean claims first and audit the pattern later. The recoupment letter shows up twelve to eighteen months after the pattern established itself, which in a transaction means after close.

This face does two things. Where the buyer’s clinical diligence catches it, it cuts the price, because exposure gets reserved against. Where it doesn’t, it survives close as a liability, and it comes back through the escrow and the indemnity the seller signed. It’s the only face that can cost the seller money after the wire hits.

The buyer’s financial QoE never opens the clinical ledger. A platform’s compliance team might. A payer’s payment-integrity team eventually will. The test that reads it first, for the seller, is The Clinical Scrub.

Why the Seller Loses Three Times

The three faces are sequenced, and the sequence is the problem.

The accounting face lands first, in the QoE, and the price drops. The operational face lands second, in the structure, and the terms tilt: more earnout, longer employment, bigger holdback. The clinical face lands last, sometimes after the seller has spent the money, and the escrow gets claimed.

Each finding on its own is survivable. Together, they explain why 85% of deals get re-priced between the letter of intent and close, and why sellers who “got their number” so often end up describing a deal they wouldn’t have signed.

Now look at who on the seller’s side sees each face. The CPA sees the first one, partially, and argues the add-backs; The CPA Trap is about why the second and third are structurally out of reach. The broker sees none of them; the broker’s job is the buyer, not the audit. Nobody on the seller’s payroll is looking at diagnosis by provider, the assist matrix, the code ratios, or the credit balances, because nobody was hired to. The buyer has three people looking. The seller has one, looking at one face.

Run the Audit First

The remedy is the same for all three faces, and it’s not complicated. It’s early.

The accounting face is a documentation exercise. Every add-back with an invoice behind it. Owner compensation reset to market and the P&L run that way for a year, so the normalized number is the reported number by the time a buyer sees it.

The operational face is a systems exercise. Associates in the exam chair alongside the owner on a defined runway. The presentation and reappointment processes documented and measured. The assist matrix, treatment planned by one provider and completed by another, built from the practice software so the pipeline can be shown to be distributed rather than claimed to be.

The clinical face is a clean-up exercise. The non-billable codes mapped to CDT and claimed. The ratios benchmarked and the outliers documented or corrected. The credit balances refunded before they become a working-capital deduction. The ledger reconciled to the claims.

Done twelve to twenty-four months before market, all three become things the buyer pays for: a normalized number that holds, a practice that runs without its seller, a clinical ledger that survives the audit. Done in the ninety days after an LOI, they’re just the reasons the deal changed. That audit, run for the seller first, is the Clinical Quality of Earnings.

The Bottom Line

Your EBITDA is a claim. Diligence is the audit. The buyer runs it with three people looking at three different faces of the same number, and the seller usually shows up having prepared for one.

I’ll walk all three, with this practice’s numbers, in Session 2 of The Practice Owner’s Playbook. Registration is free and covers all five sessions and the live roundtable.

About the author — James DeLuca is the founder of Precision Dental Analytics and works in clinical data forensics and M&A defense: the independent Clinical Quality of Earnings for sellers, buyers, and lenders, run before the buyer’s team runs theirs. He is the author of Phantom EBITDA, The Dental Data Playbook, Hidden Levers, and Spartan Leadership. Meet the team →

Frequently Asked

Questions

What is Phantom EBITDA?: Phantom EBITDA is a term coined by Precision Dental Analytics for profit that appears real on a dental practice's P&L but will not survive institutional due diligence. It has three faces. The accounting face is normalization: add-backs that fail, owner compensation below market, related-party rent, 1099 contractors doing W-2 work. The operational face is durability: EBITDA that is real today but depends on the seller, one provider's exam chair, or a marketing funnel that leaks. The clinical face is defensibility: revenue that was collected, cleanly, from coding patterns and non-billable production that a payer's payment-integrity review will later unwind. Every dollar removed is multiplied against the multiple.
What are the three faces of Phantom EBITDA?: Accounting, operational, and clinical, and each is found by a different person on the buyer's side. The accounting face is found by the financial Quality of Earnings team in about week two; it comes off the P&L and cuts the price. The operational face is found by the buyer's operations lead or the platform's integration team; it never touches the P&L and rewrites the terms instead, through a longer employment agreement, a larger earnout, and a bigger holdback. The clinical face is found by clinical diligence if the buyer runs it, and by the payer if the buyer doesn't; it cuts the price where it is caught and survives close as a liability where it isn't.
How does a buyer normalize a dental practice's EBITDA?: The buyer's analyst rebuilds the seller's number line by line. In the practice in this article, two of the seller's adjustments held: $80,000 of discretionary expense and $40,000 of one-time items came back in. Three went the other way: the owner was drawing $50,000 below what a replacement dentist would cost, 1099 contractors doing W-2 work were reclassified with payroll burden attached (about $155,000), and rent paid to the owner's own building entity was roughly $40,000 under market. Reported EBITDA of about $720,000 became about $595,000. At a 7x multiple, that $125,000 adjustment removed roughly $875,000 of price before anyone opened a chart.
Can dental practice EBITDA be collected and still not be defensible?: Yes, and it is the most expensive version. The practice in this article had insurance AR days of 21, a 93% clean-claim rate, under $500 of rejected claims all time, and insurance AR over 90 days under 16%, and it passed every revenue-cycle screen a sponsor runs. Underneath were $680,000 of coding-pattern exposure across six code families with every claim paid, $1.5 million of production under practice-created non-billable codes that never became claims, an insurance collection rate of 107% of net insurance production, and refunds of $390,000 in the practice software against $150,000 on the P&L. Payers pay clean claims first and audit the pattern later. In a transaction, later means after close, through the escrow and the indemnity.
Why do 85% of dental practice deals get re-priced between LOI and close?: Because the three faces of Phantom EBITDA land in sequence and the seller usually prepared for only the first. The accounting face lands in the QoE and the price drops. The operational face lands in the structure and the terms tilt toward more earnout, longer employment, and a bigger holdback. The clinical face lands last, sometimes after the seller has spent the money, and the escrow gets claimed. SRS Acquiom puts post-LOI price adjustments at 85% of deals. The buyer has three people looking at three faces; the seller has a CPA looking at one.
How far before a sale should a dentist run a Quality of Earnings on their own practice?: Twelve to twenty-four months before going to market, because the remedy for all three faces is early rather than complicated. The accounting face is a documentation exercise: every add-back with an invoice behind it and owner compensation reset to market for a full year. The operational face is a systems exercise: associates in the exam chair alongside the owner, presentation and reappointment processes documented and measured, and the assist matrix built from the practice software. The clinical face is a clean-up exercise: non-billable codes mapped to CDT and claimed, ratios benchmarked and outliers documented or corrected, credit balances refunded, and the ledger reconciled to the claims. Done early, all three become things the buyer pays for. Done in the ninety days after an LOI, they are the reasons the deal changed.