What EBITDA measures, and what it deliberately ignores
EBITDA answers one question: how much cash-like operating profit does this business generate before anyone argues about how it is financed, where it is taxed, and what it paid for its assets years ago? Each of the four letters after the E removes a decision that has nothing to do with running the business day to day.
Interest comes out because it reflects the capital structure, not the operations. A company owned outright and the same company after a leveraged buyout run identical factories; only one of them has an interest bill. Taxes come out because they depend on jurisdiction, loss carryforwards and structure. Depreciation and amortisation come out because they are non-cash allocations of money spent in earlier periods, and because two identical businesses can report very different depreciation simply by choosing different useful lives or by having bought their equipment at different times. Amortisation of intangibles created in an acquisition is the sharpest case: it makes an acquired business look permanently less profitable than an organically grown twin.
That is why EBITDA became the default unit of account in leveraged finance and mid-market M&A. It approximates the operating cash generation available to all capital providers, which is exactly the pool that has to service debt. It is also the denominator of the two ratios that dominate credit agreements: EV/EBITDA for pricing and debt/EBITDA for leverage covenants.
What it ignores is just as important. EBITDA treats capital expenditure as free. A trucking company and a software company with the same EBITDA are not the same business: one must spend most of it replacing trucks. EBITDA also ignores working capital swings, so a business burning cash into inventory can post a rising EBITDA while its bank balance falls. If you want the number that survives both objections, use free cash flow instead.
The two roads to EBITDA and why they must agree
There are two ways to reach EBITDA, and a correct set of accounts gives the same answer both times. Reconciling them is the fastest check on whether you have the right D&A figure.
Bottom-up starts at net income and adds back everything below the operating line: EBITDA = net income + interest + taxes + D&A. This is the form the SEC expects, because it starts from a reported GAAP measure and shows every adjustment. Note that each item is added as reported. If the tax line is a benefit, it is negative, and adding a negative number moves the running total down. That is correct: you are removing the tax effect in whichever direction it ran, not adding a positive amount.
Top-down starts at revenue and subtracts operating costs with every depreciation and amortisation charge stripped out: EBITDA = revenue − operating costs excluding D&A. This is faster off a management P&L, and it is the version that makes the margin obvious. The trap is that D&A is usually spread across more than one expense line — some inside cost of sales, some inside SG&A — so subtracting only the D&A line you can see leaves part of it in your cost base and understates EBITDA.
Both roads pass through EBIT. Bottom-up, EBIT = net income + interest + taxes; top-down, EBIT = EBITDA − D&A. Because the two routes to EBIT use completely different inputs, agreement between them is a genuine control, not a tautology. Where they disagree, the usual culprits are D&A pulled from the income statement rather than the cash flow statement, other income and expense items sitting below the operating line, or equity-method income that never belonged in operations.
Adjusted EBITDA then adds back items management argues are not representative: restructuring costs, a legal settlement, an aborted transaction, an owner's above-market salary in a private company, or the removal of a one-off gain (entered here as a negative). Adjusted EBITDA is a negotiation, not a measurement. Every add-back you claim is one the buyer's diligence team gets to test.
Worked example: a $250 million revenue distributor
Take a business with $250M of revenue that reports net income of $18M. Its interest expense is $9M, income tax expense $6M, and the cash flow statement shows $22M of depreciation and amortisation. During the year it closed a warehouse and booked $4M of restructuring costs.
- Start at net income. $18M.
- Add back interest. 18 + 9 = $27M.
- Add back income tax. 27 + 6 = $33M. This is EBIT, the operating profit.
- Add back D&A. 33 + 22 = $55M. This is EBITDA.
- Add back the restructuring. 55 + 4 = $59M of adjusted EBITDA.
- Divide by revenue. 55 ÷ 250 = 22.0% EBITDA margin; 59 ÷ 250 = 23.6% adjusted. EBIT margin is 33 ÷ 250 = 13.2%.
Now check it top-down. If the same P&L shows $195M of operating costs once every D&A charge is stripped out, then 250 − 195 = $55M, and 55 − 22 = $33M of EBIT. Both roads agree, so the D&A figure is complete.
The gap between the $55M of EBITDA and the $18M of net income is $37M, and it is exactly the three items you added: interest $9M + tax $6M + D&A $22M = $37M. If you cannot make that difference tie to the penny, one of your inputs is from a different period or a different consolidation.
How to read the margin and the gap to EBIT
The EBITDA margin tells you what fraction of every sales dollar survives operating costs. Read it three ways.
Against the company's own history. A margin that moves more than a point or two between comparable periods is either operating leverage working, a mix shift, or an accounting change — and you should be able to say which. This is the comparison that requires no benchmark and carries the most information.
Against direct competitors on the same definition. Margins are only comparable inside an industry, because the level is set by the business model rather than by management skill. A grocer that turns its inventory weekly runs on thin margins by design; a software company with negligible cost of sales runs on fat ones. Comparing across those two tells you nothing. Comparing two grocers tells you a lot.
Against EBIT. The distance between the EBITDA margin and the EBIT margin is D&A as a share of revenue, and it is the single most useful thing on this page. That gap measures how capital-intensive the business is. Where the gap is wide, EBITDA is flattering the business relative to the profit an owner can actually keep, because the assets being depreciated will eventually have to be replaced. Where the gap is narrow, EBITDA and operating profit are telling much the same story. The reference table below holds EBITDA constant and varies only D&A to show how far apart the two measures can drift while EBITDA never moves.
Finally, look at the adjustments as a share of EBITDA. There is no threshold that makes add-backs legitimate or illegitimate, but the larger they are relative to the base, the more of your valuation rests on items a buyer can simply decline to accept. In a sale process, an add-back that cannot be evidenced by an invoice, a board minute or a signed agreement is usually the first thing to fall out of the model.
Same EBITDA, three levels of capital intensity
| D&A | EBIT | Pre-tax profit | Net income | EBITDA margin | EBIT margin | Net margin |
|---|---|---|---|---|---|---|
| $10M | $45M | $36M | $27.00M | 22.0% | 18.0% | 10.8% |
| $22M | $33M | $24M | $18.00M | 22.0% | 13.2% | 7.2% |
| $45M | $10M | $1M | $0.75M | 22.0% | 4.0% | 0.3% |
Every row is arithmetic on the same $55M of EBITDA: EBIT = 55 - D&A, pre-tax = EBIT - 9, net income = pre-tax x 0.75. The EBITDA margin is blind to all of it, which is both the measure's convenience and its central weakness.
EBITDA is a non-GAAP measure, and the SEC treats it as one
Neither US GAAP nor IFRS defines EBITDA, which is why two companies can publish figures under the same label that are not comparable. For SEC registrants, presentation is governed by Regulation G and by Item 10(e) of Regulation S-K: a non-GAAP measure must be reconciled to the most directly comparable GAAP measure, must not be presented more prominently than that measure, and must not be given a title that is confusingly similar to a GAAP one. Adjusted EBITDA is permitted, but the staff has consistently objected to add-backs that smooth away normal, recurring cash operating expenses. If you are preparing a figure that will be shown to investors or lenders, present the reconciliation ladder this calculator produces alongside it, and label every adjustment.
Mistakes that make an EBITDA number wrong
- Taking D&A from the income statement. Most of it is usually buried inside cost of sales. Use the cash flow statement figure, which is complete by construction.
- Adding back stock-based compensation without saying so. It is non-cash, so many companies add it back, but it is also a genuine recurring cost of employing people. Whatever you decide, disclose it — a multiple built on an SBC-adjusted figure is not comparable to one that is not.
- Confusing EBITDA with cash flow. It ignores capital expenditure, working capital movements, cash interest and cash taxes. A business can grow EBITDA every year and still run out of money.
- Mixing periods. Every input must cover the same window. Trailing-twelve-month EBITDA built from a nine-month interim plus a full prior year double-counts a quarter unless you subtract the prior-year stub.
- Adding back an item that recurs. Restructuring in three consecutive years is not non-recurring; it is a cost of doing business.
- Forgetting minority interests. If a consolidated subsidiary is only partly owned, its full EBITDA is in your figure but only part of its value belongs to your shareholders. The enterprise value bridge has to add the noncontrolling interest back for the multiple to be consistent.
- Ignoring leases. Under IFRS 16 most leases sit in depreciation and interest, so they are excluded from EBITDA; under US GAAP an operating lease stays in operating expenses and reduces it. The same tenant can show materially different EBITDA under the two frameworks.
Where EBITDA sits among the other profit measures
Think of the income statement as a ladder and choose the rung that matches your question. Gross profit tells you about unit economics and pricing power. EBIT tells you what the operating business earns after the cost of its assets. EBITDA tells you what it earns before that, which is what a lender sizing debt capacity cares about. Net income tells you what is left for shareholders after the financing structure and the tax authority have taken their share.
For valuation, EBITDA pairs with enterprise value because both are capital-structure neutral: EV covers debt and equity claims, EBITDA is the profit available to both. Pairing EBITDA with market capitalisation instead is a category error that flatters leveraged companies. Earnings-based comparison for equity holders belongs to the P/E ratio, which is an equity-to-equity measure.
Where capital intensity or working capital swings dominate, replace EBITDA rather than adjust it. Unlevered free cash flow — EBIT less cash tax, plus D&A, less capital expenditure and the change in working capital — answers the question EBITDA dodges, and it is the input to a discounted cash flow model. Use EBITDA to screen, to size debt and to compare peers quickly; use cash flow before you commit capital.
One practical note on private companies: the EBITDA a buyer prices is almost never the EBITDA in the statutory accounts. It is a normalised figure with owner compensation rebased to market, related-party rents restated, personal expenses removed and the cost of the functions the owner performed for free added in. That normalisation is exactly what the adjustments field on this calculator is for, and it is where most of the negotiation happens.