Net income is the bottom line under the applicable accounting framework: revenue less all expenses including interest, taxes, depreciation, and amortization. It is defined, it is auditable, and it is what determines what shareholders actually earned.
EBITDA is earnings before interest, taxes, depreciation, and amortization. It is not defined by any accounting framework. It is a computed measure with no authoritative definition, which means two people can compute it differently from the same financial statements and both be defensible. That single fact explains most of the confusion and all of the abuse.
It is not a marketing invention, and dismissing it is as unhelpful as relying on it.
Comparing operating performance across different capital structures. Two businesses with identical operations, one financed with debt and one with equity, report very different net income. Removing interest lets you compare the operations rather than the financing.
Comparing across tax situations. An S corporation, a C corporation, and a partnership with identical operations report different net income because of entity-level tax treatment. Removing taxes normalizes that.
Removing accounting policy differences in depreciation. Two companies with identical assets can report different depreciation because of useful life and method choices, and one may have acquired assets in a transaction that stepped up basis while the other built the same assets. Removing depreciation and amortization strips that out.
It is the market's convention. Business valuation multiples, purchase price negotiations, and lending covenants are commonly expressed in terms of EBITDA. A practitioner advising a client in a sale or a financing works in EBITDA because the counterparty does, and refusing to engage with it on principle serves no one.
The critical framing, and the one to give a client who has fallen in love with the number: EBITDA is earnings before the cost of the assets that generate the earnings.
Capital intensity disappears. A manufacturer requiring continuous reinvestment to maintain capacity and a software business requiring almost none can show identical EBITDA and have entirely different economics. Depreciation is an imperfect proxy for that reinvestment requirement, and removing it removes the only signal in the income statement.
Interest disappears, which is exactly what you want for comparing operations and exactly what you do not want when assessing whether a leveraged business can meet its obligations.
Taxes disappear, and taxes are real cash leaving the business.
Working capital changes were never in it. A business can grow EBITDA while consuming cash, because receivables and inventory expansion appear nowhere in the measure. This is the mechanism behind profitable companies running out of money, and it is visible only in the statement of cash flows — see our post on reading the cash flow statement.
Maintenance capital expenditure is invisible, which is the practical version of the capital intensity point.
Revenue quality is not addressed. EBITDA counts revenue recognized, not cash collected.
Every transaction involves an "adjusted" figure, and the adjustments range from entirely legitimate to fictional. Knowing the difference is where a practitioner earns their fee.
Adjustments that are usually legitimate, particularly in a closely held business:
Adjustments that require real skepticism:
Three practical tests for any adjusted figure:
Demand the reconciliation, itemized from the audited or reviewed net income. A number presented without a reconciliation should not be relied upon at all.
Count the adjustments. A figure requiring twenty adjustments to reach is not a measure of the business; it is a construction. The number of adjustments is itself information.
Test each adjustment for recurrence against the prior two or three years.
|
The question |
The measure |
|
What did shareholders earn? |
Net income |
|
How does this operation compare to a peer with different financing? |
EBITDA, then check capital intensity |
|
What is the business likely worth? |
EBITDA, because that is the market convention |
|
Can it service its debt? |
Cash flow after capex, interest, and taxes |
|
Can the owner take distributions? |
Free cash flow and net income together |
|
Is it converting profit to cash? |
Operating cash flow against net income |
|
Is it a good business? |
Return on invested capital, which none of these measures |
The last row matters. Neither metric addresses how much capital the business requires to produce its earnings, which is arguably the most important question about any business. A company earning modest EBITDA on a small asset base can be far better than one earning large EBITDA on an enormous one.
EBITDA less capital expenditures, sometimes the single most useful number in an asset-heavy business, because it restores what EBITDA removed.
Free cash flow — operating cash flow less capex — which captures working capital and taxes.
Unlevered free cash flow for comparing businesses independent of financing, which is what EBITDA gestures at and does more crudely.
EBITDA margin trend, more informative than the level, because it shows operating leverage and pricing power over time.
Net debt to EBITDA, the standard leverage measure, with the caveat that it inherits every weakness of the denominator.
Structured coverage is available through Understanding EBITDA: Definition, Formula and Calculation, the Certificate in Financial Reporting and Analysis, Analyzing Financial Statements, the financial statements training catalog, and the Business Credit Analysis Bootcamp.
Practitioners should be clear about their own exposure here, because it is easy to acquire without deciding to.
For public companies, presenting a non-GAAP measure carries specific requirements: the comparable measure prepared under the accounting framework must be presented with equal or greater prominence, a reconciliation is required, and adjustments that render a measure misleading are prohibited. Adjusting to exclude normal, recurring operating expenses has drawn specific criticism.
For private company work, the exposure is subtler and more common. When a client's "adjusted EBITDA" appears in a lender package, a sale memorandum, or an offering document, the question is whether the practitioner prepared it, was associated with it, or permitted their name to be used near it. Professional standards govern association with financial information, and a figure assembled by the practitioner on a client's instruction carries responsibilities the practitioner may not have thought about.
The protective practice: state plainly what you did and did not do, keep the workpaper showing each adjustment's basis, decline to include adjustments you cannot support, and be explicit in writing when a number is management's rather than yours. A practitioner who lets an unsupportable add-back go out in a package with their firm's name on it has taken a risk for someone else's benefit.
The most valuable version of this work, and where an accountant genuinely earns a fee in a sale.
A closely held business's reported net income is frequently a tax-minimization artifact rather than an economic result, because the owner has every incentive to reduce taxable income. The purpose of normalization is to show a buyer what the business earns for a buyer — an owner paid a market salary, no personal expenses, rent at market, and no discretionary spending that a new owner would not continue.
Done well, this legitimately increases the value a seller receives, because the buyer's multiple applies to a defensible number rather than a suppressed one.
Done poorly — adjustments with no support, recurring items called one-time, personal expenses that cannot be substantiated — it destroys credibility in diligence, and a buyer who catches one unsupportable adjustment discounts every other one. The discipline is to adjust only what you can document, and to have the documentation ready before it is requested.
The clearest way to hold both ideas at once: EBITDA is a useful comparative tool and a terrible measure of financial health. Use it where the market uses it, and never answer a question about whether a business can pay its obligations with a number that excludes interest, taxes, and the cost of its own equipment.
No. EBITDA is not defined by any accounting framework, which means two analysts can compute it differently from the same statements and both be defensible. Net income, by contrast, is defined and auditable. That absence of a standard definition is the source of both the metric's flexibility and its abuse.
The cost of the assets that generate the earnings. Capital intensity vanishes, so a manufacturer requiring constant reinvestment looks like a software business requiring none. Interest and taxes — both real cash — are excluded, working capital changes were never included, and maintenance capital expenditure is invisible.
Demand an itemized reconciliation from audited or reviewed net income; count the adjustments, because a figure requiring twenty of them is a construction rather than a measure; and test each adjustment for recurrence against prior years. "One-time" items appearing every year are operating expenses with a better name, and that check is the highest-yield test available.
Cash flow after capital expenditures, interest, and taxes — not EBITDA, which excludes all three. Net debt to EBITDA is the market's standard leverage convention and inherits every weakness of its denominator, so it should be supplemented rather than relied on alone.
Yes, and it is where an accountant adds real value in a sale. A closely held business's reported income is often a tax-minimization artifact, and showing a buyer what the business earns for a buyer — market owner salary, no personal expenses, market rent — legitimately raises the price. The discipline is adjusting only what can be documented, since one unsupportable add-back causes a buyer to discount all the others.
It depends on whether the practitioner prepared the figure, was associated with it, or permitted their name to be used near it — and professional standards govern association with financial information. The protective practice is to state plainly what was and was not done, retain workpapers supporting each adjustment, decline unsupportable adjustments, and confirm in writing when a number is management's rather than the firm's.


