Chapter 5 left Meridian Industrial Services with $12 million of revenue and $2.4 million of reported EBITDA. Here is what happens to that second number in a sale process: nothing. It anchors no offer, appears in no letter of intent, and survives no diligence. The number the company will actually be priced on is adjusted EBITDA, and by the end of this chapter it will be $2.85 million.
The $450,000 of distance between those figures is not sleight of hand. It is the answer to the only earnings question an acquirer ultimately cares about: what would this business generate for a financial owner, run at arm’s length, in a normal year? Reported EBITDA answers a different question: what did this business generate for this particular owner, at the salary she chose, in the building she owns, with the relatives she employs, and with whatever unusual events the trailing 12 months happened to contain. Normalization closes the gap between those two questions, one adjustment at a time. Multiply the accepted adjustments by the earnings multiple a competitive process produces, and the add-back schedule becomes, per page, the most valuable document in the deal.
It is also the easiest document to get wrong, because the habits that build a bigger number and the habits that build a believable one point in opposite directions.
Why Reported EBITDA Is Never the Number
A private company’s income statement is a record of its owner’s decisions, and those decisions were made for reasons that have nothing to do with a sale. Compensation was set with the company’s CPA to manage taxes, not to match what a hired president would cost. Rent was set by the owner writing a check to her own building LLC. The trucks, the travel, the club dues, and the nephew in the warehouse all reflect the reality that in a private company, the line between the business and the household is drawn in pencil.
None of that carries over to the next owner. A financial owner, whether a private equity firm, a strategic acquirer folding the company into its P&L, or an individual borrowing to buy it, will pay market compensation for management, market rent for the facility, and nothing at all for the club dues. So the acquirer restates the trailing period as if it had already owned the business: strip out the owner-specific choices, remove the non-recurring noise, and see what earning power remains. That restated figure, adjusted EBITDA (you will also hear normalized EBITDA), is the base every multiple gets applied to.
Two ground rules govern the exercise, and owners forget both at their peril.
First, normalization is symmetric. It removes costs a financial owner will not bear, and it also inserts costs a financial owner must bear and strips out income a financial owner will not receive. The founder paying herself $90,000 to run a $12 million company gets adjusted downward: a real president costs more, so EBITDA falls. One-time gains (an insurance recovery, a gain on an equipment sale) come out just as one-time losses do. A schedule with no negative adjustments on it reads less like analysis and more like advocacy, and acquirers price it accordingly.
Second, every line is a claim, and claims require evidence. Adjusted EBITDA is not a fact about the company. It is an argument about the company, advanced line by line, that will be tested line by line by people paid to disbelieve it. The rest of this chapter is about which arguments win.
The Add-Back Taxonomy: What Survives and What Gets Struck
Nearly every adjustment belongs to one of six categories. They appear here in roughly descending order of credibility, and the order matters: the further down the list an owner reaches, the more expensive the reaching becomes.
Owner Compensation Normalization
What it is: restating owner pay to what a hired executive would cost in the same seat. Our industrial services founder takes $425,000 in salary and bonus. A market-rate president for a company this size costs roughly $275,000 fully loaded, so $150,000 comes back to EBITDA. The adjustment runs the other way just as often: the owner who pays herself $90,000, or nothing, must see EBITDA reduced by the cost of her replacement.
Credibility: high, when the benchmark is real. Compensation surveys and a recruiter’s read on the actual replacement role settle the number. What gets struck is the fantasy benchmark: claiming that a founder who runs sales, quotes every job, and holds the top customer relationships can be replaced by a $150,000 operations manager. Acquirers benchmark the job as it is actually done, not the job title.
One-Time and Non-Recurring Items
What they are: costs the trailing period contains that a normal year will not. Our company defended and settled a lawsuit brought by a former employee: $85,000 in fees and settlement, documented and released. It also relocated its shop across town: $90,000 in moving, rigging, and downtime costs with invoices to match. Settled litigation, facility moves, casualty events, a one-time regulatory remediation, a failed expansion cleanly abandoned: this is the legitimate heart of the category.
Three tests decide credibility: documented, resolved, and non-recurring by nature rather than by hope. Credibility: high when all three pass. Two failure modes get lines struck. The first is aging: adjustments tied to 2020-2021 disruptions meant something once, but the trailing period no longer contains them, and a schedule still carrying pandemic-era add-backs in 2026 tells the acquirer it was assembled from a template. The second is the serial one-timer: a company with a “non-recurring” legal expense in four consecutive years has a recurring legal expense, and a diligence team will line the years up side by side to prove it.
Personal Expenses Run Through the Business
What they are: household costs on the company ledger. Two trucks that never visit a job site, travel that was mostly vacation, club dues, family phone plans, and the owner’s spouse drawing $65,000 in a role that exists on paper. In our worked example these total $110,000: $45,000 of vehicles, travel, and dues, plus the $65,000 of payroll.
Credibility: high, but only line by line. Each item must trace to the general ledger with a name, a date, and an amount. A round-number estimate (“call it $75,000 of personal stuff”) invites the acquirer to do the tracing for you, with less charity. Two cautions. The family-member add-back holds only if the person genuinely does not work in the business; a spouse who actually runs the back office is not an add-back, she is an unfilled position, and the honest adjustment is the difference between her pay and her replacement’s. And the deeper the personal spending runs, the more it says about the books it ran through: acquirers read heavy commingling as a records-quality signal, not just an adjustment opportunity. Cleaning it up well before a sale is cheaper than arguing about it during one.
Related-Party Items
What they are: transactions between the company and its owner’s other interests, restated to market. The classic is rent. Our founder owns the shop building in a separate LLC and charges the company $180,000 a year; brokers put market rent at $240,000. That is a negative adjustment: EBITDA falls by $60,000, because the next owner will pay market. Above-market rent runs the opposite direction and adds back. The same logic covers purchases from a family-owned supplier, management fees paid to related entities, and shareholder loans priced off-market.
Credibility: high, precisely because it is verifiable, and the direction frequently runs against the owner. That is what makes the category valuable. Volunteering the $60,000 reduction before anyone asks is one of the cheapest credibility purchases available in a sale process. Hiding it and being caught is one of the most expensive.
Pro-Forma Adjustments
What they are: adjustments that annualize something new rather than remove something old. A price increase implemented in month nine, credited as if it had applied all year. A signed contract that has not started. A completed cost reduction with only one quarter of proof. These differ in kind from everything above: the first four categories normalize the past, while pro-forma items borrow from the future.
Credibility: medium at best, and structurally discounted. Acquirers have been burned by signed contracts that cancelled before kickoff and by price increases that quietly triggered churn, so even fair pro-forma items rarely receive full credit. The strongest of the family is a completed price increase already visible in several months of invoices, annualized with the customer list to prove it. The weakest is the run-rate argument built on the best recent quarter. In practice, acquirers respond to pro-forma EBITDA in one of three ways: partial credit, a request to see another quarter of performance, or an offer to pay for it contingently rather than at close. An owner should treat pro-forma lines as negotiating positions, not as money in the bank.
The Aggressive Category: Add-Backs That Cost More Than They Claim
What they are: adjustments that fail the financial-owner test outright. Adding back marketing spend “we could cut.” Normalizing “underinvestment” by adding back deferred maintenance or training. Treating routine repairs as if they were growth investment. Adding back bonuses that are, in truth, the price of keeping the plant manager. Claiming the acquirer’s own synergies (consolidated insurance, shared back office) as the owner’s add-back: whatever those savings are worth, they belong to the acquirer’s model, and competitive tension in the process, not a line on the schedule, is how an owner captures any of that value.
Each of these fails the same test: a financial owner would have to keep spending the money to keep the earnings. Credibility: near zero, and the expected value is negative, for reasons the credibility principle below makes plain.
The Worked Schedule: $2.4 Million to $2.85 Million
The schedule below carries our Chapter 5 company from reported to adjusted EBITDA, and it includes one line a thoughtful acquirer will contest.
| Line item | Adjustment | Category | Survives scrutiny? |
|---|---|---|---|
| Reported EBITDA, trailing 12 months | $2.40M | Starting point | n/a |
| Owner compensation above market replacement ($425K vs. $275K) | +$150K | Owner comp | Yes, benchmark documented |
| Spouse on payroll, no operating role | +$65K | Personal | Yes, role verified in diligence |
| Personal vehicles, travel, and club dues | +$45K | Personal | Yes, traced to the ledger |
| Employee litigation settled and released | +$85K | One-time | Yes, documented and resolved |
| Facility relocation costs | +$90K | One-time | Yes, invoiced and verifiably one-time |
| Rent below market on owner-owned building | -$60K | Related party | Yes, and it favors the acquirer |
| Signed maintenance contract commencing next quarter | +$75K | Pro-forma | Contested: expect a discount or structure |
| Adjusted EBITDA | $2.85M |
Eight lines, seven of which will hold. The first six are the model of a defensible schedule: benchmarked, documented, and traceable, with one adjustment (the rent) that moves against the owner and is disclosed anyway. That negative line works harder than any positive line on the schedule. It tells every reader the schedule was built to be accurate rather than to be large.
The contested line deserves honest treatment, because most schedules have one and pretending otherwise helps no one. The company signed a three-year maintenance agreement with a regional customer; work begins next quarter; the first-year EBITDA contribution pencils to $75,000. The owner’s argument is that the contract is signed and committed. The acquirer’s argument is that the trailing period contains none of it, no work has been performed, and signed contracts have cancelled before. Both arguments are reasonable, which is what contested means. The realistic outcomes are the three the pro-forma section listed: partial credit against the price, full credit in exchange for another quarter of proof, or payment contingent on the contract performing. The owner should hold two numbers in her head: $2.85 million if the line survives, $2.775 million if it does not, and she should negotiate from the first without building her family’s plans on it.
The full schedule lifts earnings by nearly 19%, which is meaningful but not exotic for a founder-run company of this size. Schedules that push far beyond that territory invite a harsher read of every line on them, including the strong ones.
The Credibility Principle
Owners write add-back schedules as lists of independent arguments: if two lines fail, six still stand, and the number lands somewhere in the middle. Acquirers do not read them that way. To the reader across the table, the schedule is a single document with a single author, and it is the first sustained sample of how that author represents facts. It gets graded like handwriting, not arithmetic: the worst line sets the impression for all of it.
This is the credibility principle, and it is the most practical idea in this chapter: every weak add-back taxes the strong ones. Present seven defensible lines and one indefensible one, and you do not collect on seven of eight. The struck line changes the posture of everything downstream. The acquirer’s diligence team shifts from verifying your schedule to rebuilding it. Contested-but-fair items that would have earned partial credit get discounted harder, because you have shown the reader what you are willing to claim. And the acquirer starts applying the same discount to things that are not on the schedule at all: the pipeline you described, the customer relationships you characterized, the reason you gave for selling.
Key point. An add-back schedule is a trust document. Acquirers grade it on its weakest line, and every add-back they strike raises the price of believing the rest. A smaller number that holds through diligence is worth more than a bigger one that does not.
The arithmetic of restraint follows directly. An aggressive line might claim $50,000 of EBITDA. If asserting it costs full credit on a contested $75,000 line and hardens the acquirer’s treatment of everything else, the claim was not free money; it was a bad trade. The owners who do best here are the ones willing to publish a schedule slightly smaller than the one they could argue for, because they understand what the document is actually for. It is not there to maximize a spreadsheet. It is there to establish that when this company says a thing, the thing is true.
Common pitfall. Consider an owner who takes $2 million of reported EBITDA to market with $600,000 of adjustments, including $130,000 of “marketing we could cut” and $80,000 of deferred maintenance framed as savings. The acquirer’s accountants strike both in their first week, then re-underwrite every remaining line, trim the fair-but-contested items they might otherwise have accepted, and the acquirer reprices the deal off a lower base with a wider margin of doubt. The aggressive lines did not merely fail. They took defensible add-backs down with them, and the repricing exceeded everything those lines were supposed to add.
The advisor’s view. Before we take a company to market, we attack its add-back schedule the way an acquirer’s diligence team will: documentation pulled for every line, compensation and rent benchmarked against market, and one blunt test for anything soft: would we defend this line across the table with a straight face? We routinely strike add-backs owners are attached to, and the adjusted number we publish is sometimes smaller than the one the owner first calculated. But it holds, and in our experience a number that survives diligence intact does more for the final price than a bigger number that gets carved up in front of the acquirer.
Quality of Earnings: Not an Audit, Something More Pointed
Within days of a letter of intent being signed, an accounting firm hired by the acquirer will begin taking your adjusted EBITDA apart. The engagement is called a quality of earnings review, QoE for short, and owners who assume a clean audit history makes it a formality get surprised twice.
An audit and a QoE answer different questions. An audit delivers an opinion that historical financial statements fairly present the company’s position under accounting standards. It is compliance-oriented, works off sampling and materiality thresholds, and is largely indifferent to whether the earnings will ever happen again. A QoE delivers no opinion at all. It is an analysis, built for a specific reader with money at stake, of three questions an audit never asks. Is the EBITDA real, meaning does it convert to cash? Is it sustainable, meaning will it recur under a new owner? And is it measured honestly, meaning do revenue recognition, cutoffs, and accruals reflect economics rather than convenience? A company can hold years of clean audit opinions and still produce a QoE report that guts its valuation, because technically compliant earnings can be concentrated, decelerating, cash-poor, or propped up by estimates that were defensible on paper and generous in fact.
Four procedures do most of the work.
Proof of cash. The QoE team reconciles reported revenue and expenses to actual bank activity, month by month, typically across two to three years. Revenue that exists on the income statement but never landed in a bank account has nowhere to hide from this procedure, and neither do expenses routed around the P&L. For owners with honest books, proof of cash is a friend: it converts claims into verified facts.
Revenue recognition review. For a project-driven business like our industrial services company, this is where the sharpest questions live. Percentage-of-completion estimates get retested against actual job costs. December shipments get examined for pull-forward. Deferred revenue gets checked to confirm that cash collected for future work is not sitting in this year’s earnings. A business with contracts, milestones, or long jobs should expect its largest projects to be rebuilt from the job files.
Customer-level margin analysis. The team decomposes profitability by customer and by job, which is how acquirers discover that the growth of the last two years came from work priced to win rather than to earn, or that the largest customer generates a great deal of revenue and very little margin.
Add-back verification. Every line of the schedule, with support. This is where the taxonomy above meets its examiners, and where the credibility principle collects its tax or pays its dividend.
The most consequential shift in lower-middle-market diligence over the past decade is that owners stopped taking this exam cold. A sell-side QoE, commissioned by the owner before going to market, has become routine in the deals we see above roughly $2 million of EBITDA. The logic is simple: whatever an acquirer’s team will find, find first. Problems discovered by your own accountants are items you fix or disclose on your own terms. The same problems discovered during exclusivity are repricing events, because by then there are no competing bidders left to discipline the response. A sell-side report also arms the schedule itself, since an add-back that arrives pre-verified is far harder to strike, and it compresses the acquirer’s diligence timeline, which matters because slow diligence kills deals in ways Chapter 11 describes. The exception sits at the smaller end of the market, where SDE is the measure and the fee is large relative to the deal; there, clean books and an advisor-built schedule usually have to carry the load instead.
Even with a sell-side report in hand, expect the acquirer’s team to re-perform rather than accept. They will rebuild EBITDA from the general ledger up, interview the controller without the owner in the room, test the months elapsed since the sell-side report (the stub period, where late-stage softness likes to hide), and push hardest on exactly the pro-forma and contested lines this chapter has warned about.
The Companion Question: Working Capital
One normalization question remains, and it rides alongside earnings all the way to close. Adjusted EBITDA describes what the machine earns; working capital, the receivables, inventory, and payables that fund the operating cycle, is the fuel that must be in the tank when the machine changes hands. Acquirers expect the business to be delivered with a normal level of working capital, defined by a peg typically set from a trailing average and enforced through a post-close true-up, and a carelessly set peg can move six figures as surely as any struck add-back. Chapter 12 owns the mechanics. The point to carry forward from this chapter is that the same normalization logic, and the same credibility rules, apply there too.
The Bottom Line
- Acquirers price adjusted EBITDA: reported earnings restated to what the business would generate for a financial owner in a normal year, with every accepted adjustment multiplied by the deal multiple.
- Credibility is categorical: benchmarked owner compensation, traceable personal expenses, documented one-time items, and related-party corrections survive scrutiny; pro-forma items get discounted; “underinvestment” and synergy add-backs get struck.
- Normalization is symmetric, so include the adjustments that cut against you; they are the cheapest credibility you will ever buy.
- Every weak add-back taxes the strong ones: the schedule is graded on its worst line, and a smaller number that holds beats a bigger number that gets carved up.
- A QoE is not an audit; it tests whether earnings are real, sustainable, and honestly measured, and commissioning your own before market turns repricing events into managed disclosures.
- Working capital is the companion normalization, settled through the peg that Chapter 12 explains.