The Add-Backs Buyers Accept (and the Ones They Reject)

The plumbing owner had a one-page schedule of SDE add-backs on his broker's CIM. Twelve line items, $238,000 of owner adjustments, neat columns, all of them defensible in his mind. His broker had told him to add back his salary, his truck, his wife's wages, the boat slip the company paid for, the consulting work his cousin did, the lawsuit they settled in 2024, two trips to Cancun that were "really sales meetings," and three smaller items. The headline EBITDA on the cover went from $310K reported to $548K adjusted — and at a 5× multiple, that's almost $1.2M of extra price.

Then the buyer's quality-of-earnings team came in. Six weeks later, the schedule looked different. Eight of the twelve add-backs got crossed off in full or in part. The "adjusted EBITDA" the buyer underwrote to wasn't $548K. It was $402K. At the same 5× multiple, the price the owner was actually arguing for had dropped by $730K — before anyone had touched the multiple itself.

(The story is illustrative; the way a buyer pulls the add-back schedule apart is not.) That gap, between what an owner proposes as add-backs and what a buyer's diligence team accepts, is where most retrades on small trade-business sales happen. And almost all of it is predictable from the categories the engine recognizes, the documentation behind each one, and a single rule of thumb: would a hired manager with no personal relationship to this business spend the same money?

What an add-back actually is (and the bar a buyer applies)

Add-backs are the lines you take out of reported profit because they aren't real costs of running the business. Done right, they bridge from accounting profit to SDE (the owner-operator's earnings — kept whole for the owner) and then to Buyer's EBITDA (what the business throws off after replacing the owner with a hired manager). For the underlying math on which earnings number applies to which buyer, see SDE vs EBITDA.

The bar buyers apply is simple, but it's tighter than most owners think. A line survives diligence if it is:

• Discretionary or owner-specific. It only exists because you own the business. A new owner wouldn't spend it.
• Documented. A line on the books, a check stub, a credit-card receipt — not a number the owner remembered while flipping through a year-end statement.
• Non-recurring (if it's claimed as one-time). Buyers look at three years; if the same "one-time" cost shows up in two of them, it isn't.
• Not already covered by another line. Replacing the owner with a market-rate manager is the biggest negative offset — and skipping it is the single most common reason a buyer kills an add-back schedule.

The categories of SDE add-backs the engine recognizes

What the offers.ai engine classifies from a GL — and what a buyer's QoE team independently checks — fits into six clean categories. Each one carries a confidence rating (high / medium / low) tied to how the engine identified it, and a tier on the supporting evidence (Tier 1 means a sourced document — a W-2, a payroll register, a QBO account, a tax return; Tier 3 means owner-stated only).

Each category surfaces with the underlying transaction trail. That trail is the part a buyer's QoE team uses to keep or kill the line — and the part most CIM schedules don't show.

The owner-comp pair: the biggest single SDE add-back swing

The largest dollar movement in almost every small-business SDE bridge is the owner-compensation pair: the owner's W-2 wages go in as an add-back, and the market-rate replacement manager comes out as an offset. The net is what actually moves EBITDA.

The engine builds this pair off two inputs. The owner's W-2 wages come from a payroll register, the QuickBooks payroll account, or the federal tax return — Tier 1 evidence a buyer accepts on sight. The replacement number comes from a Bureau of Labor Statistics Occupational Employment and Wage Statistics lookup for the right SOC role and metro at the 75th percentile — the buyer-credible market rate. Owner-attested numbers without a source document drop to Tier 3; they're still added, but a buyer will discount them harder.

Illustrative. An HVAC owner pays himself $240K. The BLS 75th-percentile General Manager comp in his metro for a $3M-revenue contractor is $145K. The bridge looks like this:

Two honest earnings numbers, $145K apart. A search-fund buyer using an SBA loan and planning to run it himself underwrites to the $550K; a PE platform installing a hired GM underwrites to the $405K. The add-back schedule is the same in both cases — it's the buyer profile that decides where the math stops.

What it means in dollars

Take a trade where the engine puts EBITDA in a 3.0×–8.0× band with a ~5.0× median — HVAC, plumbing, and electrical all sit here today. Every $10K of add-back the buyer accepts is worth roughly $30K–$80K of enterprise value, with $50K as the most common outcome.

Run that on the plumbing owner from the opening:

• Twelve proposed add-backs totaling $238K — at 5×, that's $1.19M of headline price.
• Eight rejected in part or in full, leaving $92K of accepted add-backs.
• At the same 5×, the part the buyer actually funds is $460K — a $730K reduction without anyone touching the multiple.

This is the precise mechanism by which a deal "retrades" mid-diligence. The multiple stays inside the band the comps support; the number it gets multiplied by shrinks, because the QoE team rebuilt the SDE bridge with a stricter pencil. Owners who haven't pre-graded their own schedule walk into that meeting and lose the argument because they don't have the documentation to defend each line.

Common mistakes — the patterns that get add-backs rejected

After enough trade-business diligence cycles, the same five errors drive most of the rejections:

• The recurring "one-time" cost. A $30K "consulting study" that shows up every November is the firm's bookkeeper or a relative's side business — not a one-time expense.
• A full owner salary with no manager replacement line. If the business genuinely can't run without you for two weeks, that salary is a real cost of operations. The buyer puts the BLS replacement number right back in to reach the EBITDA they'll underwrite.
• Personal expenses with no receipts. "About $15K in personal meals and travel" gets crossed off if the corresponding charges aren't tagged or traceable. The engine can flag the candidates; the documentation has to come from the owner.
• Family wages where the role is actually filled. A spouse on payroll who does run AR for ten hours a week is not an add-back. The buyer is happy to have AR run; they'll pay someone else to do it after close.
• Above-market rent to an owner-controlled property. The buyer accepts a non-arm's-length rent add-back only down to fair-market rent — not the full rent line. The rest stays as a real cost the new owner will have to pay (or renegotiate).

Each is the kind of pattern the engine surfaces with a confidence rating and the underlying GL transactions, so the owner sees the case for and against the line before a buyer's team rebuilds it. The same logic shows up in a formal Quality of Earnings report — the difference is that the QoE arrives on the buyer's clock, after the LOI.

The takeaway

Every add-back is a small argument with a buyer about what kind of cost it is. Owner-discretionary, documented, non-recurring, and not already replaced by a manager-cost line — those four tests decide which SDE add-backs survive diligence. The owners who hold their price aren't the ones with the longest schedule; they're the ones with the shortest, every line traceable, every category defensible, and the owner-comp pair already pencilled in so the buyer isn't the first person to think of it.

This is exactly the gap offers.ai surfaces from your real books — the add-back trail, the owner-comp normalization against BLS market rates, and the SDE / Buyer's-EBITDA bridge — before a buyer rebuilds it for you. New to the basis-and-multiple math? Start with SDE vs EBITDA. Want to see the rest of what a buyer prices? Read what buyers really pay for and the quality of earnings workup. Pick your trade on the industries overview or see what the diagnostic shows on a sample file.

---

Sources

1. Add-back categories (owner_compensation, personal_expenses, one_time, vehicle, family_wages, non_arms_length_rent), confidence ratings (high / medium / low), and Tier 1 vs Tier 3 evidence grading — classified by the offers.ai engine. Engine logic verified 2026-06-12.
2. SDE / Buyer's-EBITDA bridge (Adjusted EBITDA + sellerCosts = SDE; SDE − buyerCosts = Buyer's EBITDA) and the owner-comp normalization pair (owner W-2 wages added back; market-rate replacement at BLS 75th percentile subtracted) — computed by the offers.ai engine. Engine logic verified 2026-06-12.
3. Per-trade EBITDA multiple bands used in the dollar-impact arithmetic — HVAC, plumbing, and electrical all carry a 3.0×–8.0× engine band with a ~5.0× median; bands vary by trade and operator quality. getBlendedMultiple in the offers.ai engine. Engine logic verified 2026-06-12.
4. Market replacement compensation anchored to the U.S. Bureau of Labor Statistics Occupational Employment and Wage Statistics by SOC role and metro tier. https://www.bls.gov/oes/current/oes_nat.htm
5. IBBA & M&A Source — Market Pulse Report — main-street vs lower-middle-market deal-size bands; the bar above which a formal QoE is standard. https://www.ibba.org/resources/market-pulse/