Quality of Earnings: From Diligence Finding to Final Price

A quality of earnings report tests whether the EBITDA on the seller's P&L can be trusted as the base for price. It checks whether the earnings are real, whether they'll keep showing up after close, and whether they walk out the door with the owner or stay with the business. It also reconciles the target's accounting policies against what the buyer will run post close, and quantifies how much of that EBITDA the business actually converts to cash once working capital and capital expenditure are accounted for.

An audit tests whether historical statements follow GAAP. A QoE tests something different, whether a buyer can rely on the numbers to price and finance this specific deal. Audited financials can be clean and still describe a business that's entirely dependent on one customer, or run by an owner who underpays himself by $300K a year to flatter the books. The QoE exists to catch exactly what an audit isn't built to look for.

A QoE report earns its cost by giving a buyer four key things:

1. A corrected EBITDA with every adjustment traced to a dollar figure,
2. A read on whether revenue is durable or concentrated,
3. A view of how much cash the business actually consumes to run, and
4. A quantified list of debt-like items that reduce proceeds at close. Each one shapes a different part of the final agreement.

The EBITDA bridge

Every QoE report centers on one exhibit: a line by line bridge from reported EBITDA to adjusted, or pro forma run rate, EBITDA. Each adjustment falls into one of five categories: discretionary expenses, unusual or nonrecurring items, timing and accounting policy differences, revenue quality corrections, and run rate adjustments for mid year M&A activity. Every line should move a specific dollar figure. A finding that can't be quantified is an observation, not something either side can negotiate against.

Take a lower middle market services company with reported EBITDA of $4.2 million, heading into a deal priced at 6.0x.

EBITDA bridge from reported to adjusted / pro forma run rate EBITDA. Two of the five adjustments cut against the seller.

The net adjustment is $640,000, about 15% of reported EBITDA. At 6.0x, that's a $3.84 million swing in enterprise value from five line items a buyer would otherwise take on faith. Two of the five cut against the seller, which is the point: a bridge that only ever moves upward hasn't been tested hard enough.

Revenue quality: customer concentration and churn

EBITDA tells you what the business earned. It doesn't tell you whether that revenue survives the transaction. A customer and revenue rollforward does, it separates growth in the existing base from revenue lost to churn and revenue added through new logos, so a buyer isn't pricing a number that's actually three different stories netted together.

Lost customers churned at 12.8% of the prior year count but only 3.2% of prior year revenue, so the accounts that left were smaller than average, a better outcome than losing fewer, larger ones. Revenue growth from the retained base, net of churn, was 1.5%. The headline 5.9% year over year growth is mostly new logo revenue, which has no retention history yet.

That growth also has to be checked for concentration, which the rollforward above doesn't show on its own:

The largest customer alone is 12.3% of current year revenue; the top five are a third of it. Above roughly 10% for a single account, a QoE report should flag concentration by name, because the multiple being paid assumes that customer renews. Buyers and targets can also schedule a direct conversation with a highly concentrated customer before close, to confirm the relationship and gauge whether that customer would view a continued partnership under new ownership positively.

From EBITDA to cash: working capital and capital expenditure

A multiple is applied to EBITDA, but a buyer writes a check for cash flow. The gap between the two is capital expenditure and the net working capital the business needs to fund growth. Neither shows up on the EBITDA bridge, so a QoE report further quantifies both separately.

Net working capital

Net working capital swings $1.07 million between its low and high points across the year, seasonal buildup in inventory and receivables that has to be funded regardless of EBITDA. The trailing twelve month average, $3.465 million, is the natural starting point for the net working capital peg written into the purchase agreement: the target the seller has to deliver at close, with a dollar for dollar price adjustment for any shortfall or surplus, which would affect seller proceeds. (How that peg is negotiated from LOI through close, and trued up after, is its own topic, which I will cover in a separate article.)

Capital expenditure

Depreciation covered 77% of capex over three years, meaning $320,000 of cumulative spend never passed through the income statement as a deduction. EBITDA adds depreciation back on the assumption that it roughly approximates ongoing capex; here it understates it, because the gap is funding real reinvestment, equipment and systems upgrades, not just maintaining what already exists. That $320,000 doesn't show up in EBITDA and isn't optional. A buyer who prices the deal on EBITDA alone is implicitly assuming capex equals depreciation, which this company's own numbers show is wrong by roughly $107,000 a year. In the final agreement, this surfaces as a lower multiple, an explicit capex reserve built into the model, or a maintenance capex adjustment to the EBITDA bridge itself.

Net debt and debt-like items

Net debt and debt-like items both sit between enterprise value and what the seller actually walks away with, but they get quantified differently.

Net debt

Net debt is funded borrowings, term loans, revolver draws, capital leases, minus cash and cash equivalents on hand at close. It isn't simply read off the balance sheet. Differences in accounting practices between buyer and seller routinely require adjustments and reclassifications before the two sides agree on one figure, the same reconciliation work a QoE report does for EBITDA. What counts as cash is its own question: restricted cash, held as collateral or tied to a specific use, doesn't offset debt the way unrestricted cash does. Net debt also moves between signing and close, so the purchase agreement needs its own true-up mechanism, similar in spirit to the working capital peg, and a common source of post-closing disputes when the calculation isn't pinned down precisely upfront. Cash and debt get negotiated once this finding is quantified, because wherever the number lands moves seller proceeds dollar for dollar, the same mechanic as the debt-like items below.

A further $70,000 of restricted cash, held as a landlord security deposit, sits on the same balance sheet but doesn't offset debt under the purchase agreement's definition, exactly the kind of question about valuing debt and cash accurately that makes net debt its own piece of negotiation rather than a number read off the books.

Debt-like items

A QoE report also catalogs debt-like items sitting on the balance sheet that reduce proceeds dollar for dollar at close, separate from funded debt: unfunded accrued paid time off, customer deposits and unearned revenue collected but not yet delivered against, and pending sales tax or payroll tax exposure from prior periods.

Let's assume that on this deal, diligence quantified two of them:

These reduce the equity check the same way funded debt does, and get negotiated the same way the EBITDA bridge does: itemized, quantified, and subject to pushback, not accepted on the seller's word. How net debt and the $520,000 of debt-like items both roll into the final number is below.

How this shapes the final agreement

Every number above feeds one bridge, from enterprise value to the amount the seller actually walks away with at close.

Enterprise value to equity value: how net debt, debt-like items, and the working capital peg true-up turn a headline multiple into actual proceeds at close.

For reference: the unadjusted $4.2 million of reported EBITDA, at the same 6.0x multiple with none of these deductions applied, implies $25.2 million, almost the same figure the fully adjusted analysis lands on once net debt, debt-like items, and the working capital shortfall are subtracted. The QoE found $640,000 of real EBITDA the reported number missed. That finding didn't inflate the price. It got spent absorbing net debt, debt-like items, and a working capital shortfall the unadjusted number never accounted for.

The adjustments determined the price, line by line, instead of leaving it to whichever number the seller led with.

A full QoE typically runs three to four weeks from complete data to draft report, and most of that time is set by how fast the seller produces clean data, not by the analysis itself. Request the data the day exclusivity starts. A report is only worth what it finds in time to use it.