What is SDE, and which add-backs actually survive

Almost every small business in America is priced on a number that does not appear anywhere in its accounts. It is not revenue, it is not net income, and no accountant produced it. It is seller's discretionary earnings, and it is assembled by the seller, line by line, out of the reported profit plus a list of things the seller argues should not count against them. Roughly a third of a typical SDE figure is not profit at all. It is argument.

I price private companies for a living, more than a thousand of them, and the add-back schedule is where I spend a disproportionate share of my time on any small deal. Not because it is complicated, but because it is the one page in the package where a seller can move the price by tens of thousands of dollars in either direction and usually does not realize which direction they are moving it. Here is what SDE actually is, which add-backs survive contact with a buyer, and what each line is worth.

What SDE is, and why it exists at all

Seller's discretionary earnings is the total annual financial benefit a business delivers to one full-time owner-operator. The construction is standard: start at pretax profit, then add back interest, taxes, depreciation and amortization, one working owner's full compensation and benefits, and any genuinely non-recurring or personal expenses the business paid.

It exists because a small company's reported profit is close to meaningless in isolation. The owner sets their own salary, which might be $40,000 for tax reasons or $300,000 because it was a good year. They decide whether the car, the phone and the family health plan run through the business. They choose the depreciation schedule. Reported profit at this size measures the owner's tax preferences at least as much as it measures the business. SDE strips those choices out so two companies can be compared at all, and so a buyer can see what the thing would actually put in their pocket.

The scale here is worth holding in mind, because it is smaller than most people picture. In BizBuySell's Insight Report for the second quarter of 2026, across 2,117 businesses that changed hands, median cash flow was $155,921 on median revenue of $692,087, and the median sale price was $349,250. That is the middle of this market: a business throwing off about $156,000 a year for one working owner. At that size an add-back schedule is not a rounding error. It is a material fraction of the entire thing being sold.

Where SDE stops and EBITDA starts

The difference between SDE and adjusted EBITDA is exactly one line: owner compensation. SDE adds back one working owner's full pay and benefits. Adjusted EBITDA does not, because it assumes the business pays a market-rate manager to do that job. The gap between the two numbers is the cost of replacing the owner, and that is not an accounting distinction, it is a description of who is buying.

Below roughly $1 million of earnings the likely buyer is an individual who will run the business themselves, so they think in SDE: they want to know what they will take home. Above roughly $2 million the buyer is usually a fund, a family office or a strategic acquirer who will hire a manager, so they think in adjusted EBITDA. The band between is a handoff zone where different bidders will price the same company on different metrics. This matters practically: an SDE multiple and an EBITDA multiple are never comparable, and a multiple quoted without its metric is not information. I have watched sellers conclude one offer was far better than another when the two were simply quoted on different bases.

Every add-back is worth its multiple

Here is the mechanic that makes this page worth reading carefully. An add-back does not add its own value to the price. It adds its value multiplied. Across 7,601 transactions in my Felpel Private Company Index the median business cleared 2.67 times SDE, so at that multiple every dollar of defensible add-back is about $2.67 of purchase price, and every dollar a buyer strikes is about $2.67 off.

Put an illustrative package against it. A business reports $110,000 of pretax profit and presents a $46,000 add-back schedule, arriving at $156,000 of SDE, close to that BizBuySell median. At 2.67x the asking logic produces roughly $416,500. Now the buyer works through the schedule and strikes $15,000 of it, which is one contested line and part of another. SDE becomes $141,000 and the price becomes roughly $376,500. A $40,050 move on $15,000 of disputed expenses.

Note what that arithmetic does to the incentives. It makes an aggressive schedule look extremely attractive, because a $15,000 line nobody challenges is $40,050 of free price. That is precisely why buyers assume the schedule is padded before they have read it, and why the seller who pads it pays a cost that does not show up on any single line.

The add-backs that survive

After enough of these, the pattern is simple. An add-back survives when a stranger can verify it without taking the seller's word for anything. That is the whole test. Four groups clear it reliably.

1. The structural ones. Interest, taxes, depreciation and amortization, plus one working owner's full compensation and benefits. These are definitional rather than argued: they are what the metric is. Nobody pushes back, provided you add back exactly one owner.

2. Personal expenses with paper. The owner's health insurance, personal vehicle, personal phone, the family gym membership booked to the business. These survive comfortably when the invoices show plainly what they are. They stop surviving when they are a round number in a spreadsheet with no underlying document, which is more common than it should be.

3. Genuinely one-time items with third-party proof. A lawsuit settled, with the settlement agreement attached. Storm damage, with the insurance claim. A one-off legal or consulting invoice tied to a specific event that has ended. The proof has to come from outside the business, because the seller's own characterization of an expense as non-recurring is the thing in question.

4. Related-party rent, normalized. Owners frequently own the building and charge the business above or below market. Adjusting to market is legitimate and expected, in either direction, and it survives when a comparable lease or a broker's opinion is attached. Adjusting it to whatever number makes the year look best does not.

The add-backs that do not

The rejected categories are just as consistent, and four of them account for most of what I strike.

1. Recurring expenses dressed as one-time. This is the big one. The test I apply is blunt and easy to run on yourself: if a line appears in three consecutive years, it is not non-recurring, whatever it is called. Businesses that have "one-time" legal costs every single year have legal costs.

2. The second working family member. SDE adds back one working owner. If a spouse or an adult child draws a salary and genuinely does a job, that job will still need doing on the Monday after closing, and their pay is a real operating cost. Adding back the whole family payroll is the single most common overreach in small-business packages.

3. Spending the business actually needs. Marketing described as a discretionary experiment while revenue plainly depends on it. A role left unfilled. Deferred maintenance. These are not add-backs at all, and a careful buyer will go further and treat them as negative add-backs, subtracting the cost of putting them right. An owner working sixty hours a week whose replacement manager costs $80,000 has an $80,000 deduction sitting in the model whether or not anyone writes it down.

4. Anything that touches unreported revenue. Some sellers want credit for income that never appeared on a tax return. Do not put it in the schedule, do not raise it in the room. It cannot be verified, no funded buyer can underwrite it, and introducing it converts a pricing conversation into a conversation about whether the rest of the numbers can be trusted at all.

What a struck add-back really costs

The visible cost of a rejected add-back is its multiple, and we have already priced that. The invisible cost is larger and it is the reason I argue against padding even on pure self-interest grounds.

A buyer works the schedule in order. When one line fails, they do not simply delete it and continue at the same level of trust. They reprice their confidence in every other line, including the ones that were perfectly good. A schedule with one indefensible item is treated very differently from a schedule with none, and the discount lands on the honest lines too. I have seen a seller lose more on the four legitimate add-backs that suddenly required proof than on the one fabricated one that started it.

It also moves the argument to the worst possible moment. Most of this gets tested after a letter of intent is signed, inside diligence, when the seller has lost their competitive tension and has a buyer who now has leverage and a reason to use it. Axial's Dead Deal Report examined 75 lower middle market deals that broke in 2025 and found that 21.3 percent died on a quality of earnings EBITDA discrepancy, with another 25.3 percent dying on other diligence findings. Close to half of dead deals die inside diligence, and the earnings schedule is a large part of what is being tested there. I wrote up that process in more detail in the piece on quality of earnings.

How to build a schedule that holds

The method is unglamorous and it works. Attach the proof to the line, not to a folder. Every add-back should carry its document at the moment it is presented: the invoice, the settlement agreement, the comparable lease, the insurance claim. A schedule that arrives pre-proved is read completely differently from one where each line generates a request.

Run the three-year test on yourself before the buyer does. Pull three years side by side and delete anything that recurs. You will lose some SDE. You will lose far less than you would have lost having it struck in week six of diligence.

Add back exactly one owner, and write down what the replacement costs. If you work sixty hours and the business needs a manager at $80,000, put that in yourself. Volunteering the negative add-back is the single most credible thing a seller can do, and it buys goodwill on every line above it.

State which metric you are quoting, every time. Write "2.8x SDE" or "5.5x adjusted EBITDA", never "5.5x". And when you compare offers, restate both on the same basis before deciding which is higher, because otherwise you are comparing two different questions.

The thing to hold on to is that the add-back schedule is the most leveraged page in the entire package. Every honest dollar on it is worth about $2.67 of price at the median, and every dishonest one costs you that and takes some of the honest dollars with it. A defensible schedule is not a smaller schedule. It is a schedule where every line has a document behind it, and that is available to any owner willing to spend a weekend on it twelve months before they sell.

Common questions

What is seller's discretionary earnings (SDE)?

Seller's discretionary earnings is the total annual financial benefit a business delivers to one full-time owner-operator. You start from pretax profit and add back interest, taxes, depreciation and amortization, one working owner's full compensation and benefits, and any genuinely non-recurring or personal expenses the business paid. It exists because a small company's reported profit is close to meaningless on its own: the owner decides their own salary, how much of their car and phone runs through the business, and how aggressively to depreciate. SDE strips those choices out so that two businesses can be compared, and so a buyer can see what the company would actually put in their pocket.

What is the difference between SDE and EBITDA?

One line: owner compensation. SDE adds back one working owner's full pay and benefits; adjusted EBITDA does not, because it assumes the business pays a market-rate manager to do that job. So SDE is always the larger number, and the gap between them is the cost of replacing the owner. This is why the two are never interchangeable and why an SDE multiple is always lower than an EBITDA multiple for the same company. If someone quotes you a multiple without saying which metric it applies to, the number is not usable.

Which add-backs do buyers actually accept?

The ones that are definitional or documented. One working owner's compensation and benefits, interest, taxes, depreciation and amortization are structural and rarely argued. Personal expenses run through the business, such as the owner's health insurance, personal vehicle or phone, survive when the invoices show what they are. Genuinely one-time items survive when third-party paper proves it: a settlement agreement, an insurance claim, a one-off legal invoice. Related-party rent survives when it is normalized to market with a comparable lease attached. The common thread is that a stranger can verify each one without taking the seller's word for it.

What is an add-back worth in purchase price?

Whatever the multiple is. Across 7,601 transactions in the Felpel Private Company Index the median business cleared 2.67 times SDE, so at that multiple every dollar of defensible add-back is about $2.67 of price, and every dollar a buyer strikes is about $2.67 off. On an illustrative package showing $156,000 of SDE, close to the $155,921 median cash flow BizBuySell reported for businesses sold in the second quarter of 2026, striking $15,000 of add-backs moves the price from roughly $416,500 to roughly $376,500. That is a $40,050 swing on $15,000 of disputed expenses, which is the whole reason the schedule is worth building carefully.

At what size do buyers stop pricing on SDE?

It tracks the buyer, not a rule. Below roughly $1 million of earnings the likely buyer is an individual operator who will run the business themselves, and they think in SDE because they want to know what they will take home. Above roughly $2 million the buyer is usually a fund, a family office or a strategic acquirer who will hire a manager, and they think in adjusted EBITDA. The band between is a handoff zone where different bidders will price the same company on different metrics, which is worth knowing before you compare two offers and conclude one is far better.

Related reading

• Quality of earnings: the report that moves the price.
• Is a business worth 3x profit? The honest answer is a range.
• How an independent sponsor prices a deal before the money exists.
• How I price a company in 48 hours.

Go deeper

If you are preparing a business to sell and you want the add-back schedule stress-tested before a buyer does it for you, let's go through it line by line, or read more of how I think about pricing private companies.

Tomasz Felpel is an investor, founder, and advisor in private markets and healthcare, based in New York. He is a three-time founder of Value Alpha, an AI-powered private-markets valuation platform, Sonnerie VC, an early-stage healthcare venture firm, and Pond. Previously he led corporate development and M&A at Fortune 500 scale, pricing more than 1,000 private companies. Columbia Business School EMBA. Read the full story.