What is a Quality of Earnings (QoE) analysis?

A quality of earnings analysis is an independent, third-party examination of whether a company's reported earnings are real, repeatable, and sustainable. In M&A, it's the work that tests the earnings a buyer is actually paying for before a deal closes.

Price follows earnings. Buyers pay a multiple of EBITDA, so a QoE pressure-tests that number — stripping out one-time events, accounting distortions, and owner-specific costs to isolate what the business can be expected to earn once the current owner is gone.

What a quality of earnings analysis answers

Reported net income rarely reflects a business's true earning power. A QoE report goes past the surface accounting to assess how sustainable and accurate a company's earnings really are. The question it answers is simple: how much of this is repeatable, reliable, and transferable after the deal?

A QoE separates sustainable earnings from one-time events, accounting distortions, and operational noise, and produces a defensible view of the company's economic performance. High-quality earnings are stable, predictable, and driven by recurring operations, which signals lower risk. Low-quality earnings get inflated by one-time gains, aggressive accounting, or trends that won't hold, which signals the opposite.

The deliverable is usually an Excel databook plus a narrative report, and it forms the backbone of a broader financial due diligence (FDD) effort — for which the AICPA publishes practice guidance — that lenders, PE investment committees, and boards rely on.

Why QoE matters in a deal

For buyers, a QoE confirms that the price assumptions reflect reality, and it surfaces risks like revenue recognition problems, customer concentration, or unsustainable cost structures that could otherwise appear after close.

For sellers, a sell-side QoE done before going to market lets you control the financial narrative, fix issues before a buyer finds them, and defend your valuation. Done well, it turns diligence into confirmation rather than discovery. Sell-side QoE reports are now standard practice for all but the smallest transactions.

The stakes scale with the multiple. At a 5x EBITDA multiple, a $50,000 expense legitimately added back raises the purchase price by $250,000; a $200,000 error in normalized EBITDA moves enterprise value by $600,000 to $1.2 million at a 3x–6x multiple. That leverage is why buyers commission a QoE to independently verify each adjustment before committing to a price.

QoE vs. audit vs. valuation

This is the most common point of confusion, so it's worth being precise. An audit gives an opinion on whether financial statements are presented fairly under GAAP; a QoE assesses whether earnings are real and repeatable from an investor's perspective. A QoE will flag items that are perfectly correct under GAAP but still distort a buyer's read of the business. Both look largely at historical data, so the difference isn't timing — it's what each one is for.

A QoE is also not a valuation. The sequence runs: an audit confirms the books are clean, a QoE confirms the earnings are durable, and a valuation translates those earnings into a number.

What's inside a quality of earnings report

There's no universal format — content varies by firm, industry, and the specific concerns of the buyer. But most QoE reports cover the same core areas.

The EBITDA bridge is the centerpiece. A QoE walks reported EBITDA to a defensible base, then validates every add-back along the way:

To see the leverage: an owner paying themselves $800,000 where a replacement manager costs $250,000 is suppressing EBITDA by $550,000. A QoE typically evaluates monthly data over the most recent three fiscal years plus the trailing twelve months (TTM) to expose trends and isolate unusual items. Beyond the EBITDA bridge and add-backs, the other core workstreams:

• Revenue quality and recognition — confirming revenue is booked in the correct period and is actually collectible. A classic finding: a company bills an annual contract up front and books it all immediately, when GAAP requires recognizing it across the service period. (The three-way match — sales order, shipping document, invoice — is how this gets tested transaction by transaction.)
• Customer concentration — top-10 or top-20 customer analysis by revenue and gross margin. Will the customers stay once the founder steps down?
• Working capital normalization — setting the target net working capital the business needs to operate, which drives the closing working capital adjustment — a frequent source of post-close disputes when it isn't nailed down early.
• Net debt and debt-like items — because most deals are cash-free / debt-free, a QoE identifies what counts as debt at closing. Beyond funded debt it flags debt-like items that reduce the equity a buyer receives: unpaid taxes, deferred or contingent consideration, unfunded pension or earn-out obligations, accrued unpaid bonuses, capital leases, and factored receivables. This schedule is what bridges enterprise value to equity value, and it's a frequent late-stage negotiation because one side's "working capital" is the other side's "debt."
• Proof of cash — reconciling reported EBITDA back to actual operating cash flow, often using bank statements as the independent check. Strong EBITDA with weak operating cash flow can hide a problem the income statement doesn't show.
• Hidden liabilities — accrued PTO, deferred rent, lease obligations, contingent liabilities, and tax exposure.

Buy-side vs. sell-side QoE

The sell-side version is increasingly the default. One rule holds in both cases: one firm should never perform a QoE for both sides of the same transaction.

What a QoE costs and how long it takes

Cost scales with the size and complexity of the business. Based on 2026 deal conditions, a rough guide:

Multi-entity structures cost materially more. A draft typically takes four to six weeks from engagement to draft report, plus a week or two for revisions — heavily dependent on how fast the owner can supply data and answer questions. QoE reports are standard above roughly $2M EBITDA and increasingly common at $1M+, with limited-scope reports emerging as a faster, cheaper option for buyers who already know the business.

A worked example: reading CoreWeave's Q1 2026 earnings

What does this look like on a real company? Below is a QoE-style read of CoreWeave's Q1 2026 10-Q — the same reported → adjusted → run-rate bridge applied to a GPU-infrastructure business, where one assumption (how long the GPUs really last) swings the answer more than anything else.

The bridge starts at the GAAP operating loss and walks to a normalized EBITDA. CoreWeave's Q1 2026 10-Q reports a $144M operating loss that becomes $856M reported EBITDA once depreciation and stock comp are added back — then normalizations for concentration risk and an aggressive depreciation schedule pull it down to a defensible ~$605M.

The single most consequential input is GPU economic life. Re-pricing depreciation against a shorter useful life moves the business from a ~29% margin to a ~2% margin — without touching a single contract or customer. This is exactly the kind of assumption a buyer's QoE stress-tests first.

Seen against peers, the size of the adjustment is the story: diversified data centers keep ~92–94% of reported EBITDA after diligence, while a concentrated, GPU-heavy business keeps closer to ~70%.

See the full interactive example: Quality of Earnings — CoreWeave Q1 2026 ↗

Edge cases and common errors

• "Non-recurring" that recurs. The first buyer question for every add-back is: did this category of expense appear in prior years too? A "one-time" legal or consulting cost that shows up three years running isn't non-recurring — it's operating cost, and adding it back inflates the base.
• Unsupported add-backs erode everything. An adjustment with an invoice behind it survives; one without gets struck. Aggressive add-backs the QoE reverses don't just lose that line — they damage the seller's credibility on every other line.
• Don't double-count across workstreams. The same benefit can hide in the EBITDA add-backs, the run-rate annualization, and the working-capital normalization. Adjust each dollar once.
• Proof of cash is the reality check. Strong reported EBITDA that doesn't convert to operating cash flow is a red flag — channel stuffing, aggressive revenue timing, or capitalized costs that should be expensed.
• Run-rate is not a wish. Annualizing a mid-period contract is fair; annualizing a hoped-for pipeline is not. Run-rate adjustments need a signed, supportable change behind them.
• Match the metric period and definition. Be consistent about LTM vs. forward, and about how "adjusted" EBITDA is defined, across every period in the analysis.

Limitations

A QoE is decision-support, not a guarantee. It relies on the data the seller provides — a thin or disorganized data room limits what can be verified, and outright fraud can defeat any diligence built on the company's own records. It's largely historical: it tests whether past earnings were real, not whether they'll persist through a downturn or a lost anchor customer. And it's judgment-heavy — reasonable professionals disagree on whether an add-back is truly non-recurring — which is exactly why the signed report and its conclusions belong to a qualified professional, not a checklist.

The part everyone underestimates: the work behind the number

The cost and timeline numbers don't show the real work. The formula for normalized EBITDA is one line; the diligence behind it is weeks of someone reading a data room and tying every proposed adjustment back to a source document. General ledgers, ERP exports, bank statements, customer contracts, payroll detail, legal files — the quality of the answer depends on whether each adjustment is supported, traceable, non-duplicative, and consistent with revenue recognition, working capital, and cash flow.

This is the floor of manual effort no one enjoys: pulling the same figures across three years of monthly data plus a TTM period, reconciling EBITDA to cash, and assembling add-back tables where every line needs evidence behind it. An adjustment with an invoice behind it survives buyer scrutiny. One without it gets struck, and takes valuation with it.

This is where an agentic workspace changes the economics of the work. With MacrosLM's QoE & EBITDA Bridge Automator, you drop the full data room (contracts, statements, GLs, ERP exports, PDFs, scans) into one project and brief it in plain language, the way you'd brief an associate. It reads across the documents, builds the reported-to-adjusted-to-run-rate EBITDA bridge, drafts the add-back validation tables, and ties every figure back to its source page through an evidence layer. Click any number and see where it came from. The associate-level re-keying and tie-out that made QoE slow and expensive drops from days to a first draft you review.

The judgment is still yours. Deciding whether an add-back is really non-recurring, whether the revenue holds, whether a concentrated customer is a risk worth pricing in — that's the analyst's call, and it should be. What MacrosLM takes off your plate is the data entry that was crowding it out. A QoE is decision-support, and the professional who signs it owns the conclusion.

Bottom line

A quality of earnings analysis is the financial foundation of an M&A deal — an independent test of whether reported earnings are real and repeatable, built around a defensible normalized EBITDA, validated add-backs, revenue quality, working capital, net debt, and proof of cash. It's not an audit and not a valuation. It's the bridge between clean books and a credible price — and the side with every figure traced to its source is the side that holds its number at the table.

This article is for general information and is not accounting, tax, legal, or investment advice. A quality of earnings analysis should be performed or reviewed by a qualified professional. Costs, timelines, and market practices reflect publicly reported 2025–2026 conditions and vary by transaction.