What Is a Good EBITDA Margin? Buffer and Industry Ceiling
No published threshold says what a good EBITDA margin is. Two questions do: how much price the company can give up, and where its industry ceiling sits.
See a sample reportThis is a method, not a recommendation. Nothing here, or anywhere else on Taufolio, is investment advice. Treat every example as a starting point for your own research.
A company earns a six percent margin. Is that bad? The number alone does not settle it and never will, until you answer two others: how much price can this company give up before it stops earning, and where does the ceiling sit in its industry. EBITDA margin is a good figure for asking both, as long as you stop looking inside it for a threshold nobody ever set. A single reading is never information, and a margin is no exception: what gives it meaning is level, trend and comparison against the rest of the industry.
Start with the thing teaching material gets wrong most often, because it ruins everything downstream: a margin is a quotient, not a difference.
Every number in the examples is invented and round, except one, which carries its source and its date in the same sentence.
How to calculate EBITDA margin
EBITDA margin is EBITDA divided by revenue. On 1 billion of revenue and 260 million of EBITDA it comes to 26 percent, so the company keeps twenty-six cents of every dollar it sells, before it pays interest and tax and before it records the wear on its assets.
The error you meet in guides is to take EBITDA and subtract costs from it, then call the result a margin. That produces a figure which is no margin at all, and which looks enough like a percentage to travel into comparisons and stay there. If a margin has ever come out of your spreadsheet in a way you could not rebuild from the income statement, this is usually why.
The same arithmetic one line up or down gives different margins: gross, operating, net. Each answers a different question and none of them may be set against another. When two databases show different margins for one company, the data is usually identical and the level it was taken from is not.
Margin as a buffer against a price cut
The popular version says a high margin proves quality. Something simpler and more useful sits closer to the truth: a margin is the room a company has to move when a competitor cuts its price.
Take two companies with the same 1 billion of revenue. The first runs a 6 percent EBITDA margin, so 60 million. The second runs 26 percent, so 260 million. Now somebody enters the industry and cuts the list price by 5 percent, and both have to answer with the same move.
At the same volume, a 5 percent cut removes 50 million of revenue from each of them, and the costs stay exactly where they were. The first is left with 10 million instead of 60, so it has surrendered five-sixths of its EBITDA. The second is left with 210 million instead of 260, so it has surrendered a little under a fifth.
Which means the identical price cut is a question of survival for the first company and a weaker quarter for the second. That is the whole content of the word buffer, and it is why I read the margin before any valuation multiple.
This is worth separating from the claim the pricing power test makes: there, a margin is evidence that a company can raise prices without losing customers. Here it is something else, the reserve for a situation where prices have to come down. One belongs to good years, the other to bad ones. Only the second tells you what happens when a cheaper competitor shows up.
The ceiling belongs to the industry, not to the board
The second half of the answer is unwelcome news for anyone hunting a single number: the level of a margin largely does not belong to the company.
The test is one question. Does the customer choose by comparing prices? Fuel, a current account, motor insurance and a mobile plan are bought on price, because the product is close to identical everywhere and the comparison takes a minute. Nobody in an industry like that will run a high margin however well they are managed, because every attempt to lift it ends with the customer clicking on the next offer.
Where the customer does not compare on price, because they cannot, will not, or the product genuinely differs, the margin can be several times higher. That is the same mechanism an economic moat describes, seen from the income statement rather than from the story a company tells about itself.
The ceiling also moves with the age of the industry, and that is the second thing a single figure hides. In the growth phase margins rise with volume, because fixed costs spread across more and more sales. At the turn, once everyone who wanted in is in, the margin cracks first, before revenue has stopped growing. In maturity it settles high or low depending on how the customer chooses, and it stays there for years.
So a three-point fall in margin in an industry three new entrants just walked into describes something other than the same fall in an industry nobody has entered for a decade. The first is the price of new competition and may or may not reverse. The second is information about the company.
Hence the conclusion many readers find the most practical thing in this whole subject: the best company in a thin-margin industry has a harder job than an average company in a fat-margin one, and no spreadsheet reverses that. A margin compared across industries does not measure the quality of a board. It measures how the customer chooses.
Why I will not give you one good number
The thresholds circulating in guides fall into three classes, and only one of them is a number you may copy.
The first class is an accounting rule: the formula itself, EBITDA over revenue. It contains no threshold, it is a definition, and a definition does not age.
The second class is market background, a figure describing where the market stands right now. The aggregate net margin of the broad US index is 14.9 percent on the estimate for the third quarter of 2026, against a five-year average of 12.4 percent, as reported in this quarterly earnings review published on 4 September 2026. That is a net margin, a different line of the income statement from EBITDA, and it describes an index rather than your company. It is a scale rather than a threshold, and a year from now it has to be pulled again, because it will have aged.
The third class is an arbitrary threshold: a number somebody once picked that travelled onward without its paperwork. I went looking for a margin threshold with a publication or an institution behind it and found none. So I am not going to hand you a number below which a margin is bad, and I think anyone who hands you one without a source is handing you their own calibration dressed as a standard.
What goes in the place of a threshold is a zone, and a zone is something you compute yourself.
A percentile instead of a threshold, worked by hand
Say the peer group holds twelve companies from one industry, and their EBITDA margins, ranked from lowest, run 4, 6, 7, 9, 11, 12, 14, 15, 17, 19, 24 and 31 percent.
The median sits between the sixth and seventh entries, so 13 percent. A company on a 19 percent margin clears nine of the twelve, which puts it in the top quarter of the distribution. A company on 6 percent clears one, which puts it in the bottom.
Which means the same six percent we started with reads as "nearly the lowest in the industry" in this group, while in a group of thin-margin retailers it would read as "above the median". One figure, two conclusions, and neither of them follows from the figure alone.
Three conditions before that arithmetic means anything:
- The group holds at least ten companies. Below that, skip the percentile and write down plainly that the group is too small.
- The companies are comparable by model rather than by sector label. A manufacturer and a distributor of the same goods have different margins by construction.
- The reading carries a date. Whole industries move together through a cycle, so a percentile from two years ago describes a different industry.
What if no peer group can be assembled, because the company does something nobody else does at that scale? Then the second axis is its own history: the same margin across five or ten years, with the year of the industry's worst quarter marked. Its own history is a weaker reference than a peer group, because it says nothing about competitors. It does say the one thing a peer group cannot: whether today's number is this company's normal state or its peak.
Half an hour in a spreadsheet buys you a zone instead of a threshold from a guide.
When the raw number is enough on its own
The rule that a level means nothing without comparison has an exception, and naming it is the honest thing to do, because otherwise this piece promises more caution than it practises.
An EBITDA margin around sixty percent needs no peer group. A level like that says on its own that the company sells something whose next unit costs close to nothing to make, and that nobody is stopping it so far. The far end works the same way: a margin near zero says that whatever goes wrong will go wrong immediately.
Between those two extremes the raw number says nothing. Nine companies in ten live there.
Where exactly does the extreme begin? I do not know, and I am not going to pretend otherwise. For me it starts where the gap to the industry median stops fitting inside that industry's ordinary spread, and the spread itself varies. Where everyone sits within a few percentage points, clearing the median by five is an event. Where the range runs to tens of points, the same five points are noise.
High and falling against low and rising
The level is one third of the answer. The trend is the second third, and the comparative background above is the last.
A company on a 10 percent margin that stood at 7 three years ago is doing something that works. A company on 15 that stood at 19 is also doing something that works, except that something stronger is happening in its market. I prefer the first, even though it looks worse in a table, and that is an opinion rather than a rule.
The mechanism behind the opinion runs like this. A rising margin on rising revenue means costs are growing more slowly than sales, which usually means the company holds something the customer does not want to swap. A falling margin on rising revenue means the sales are being bought with price.
The symmetry applies in both directions and is worth applying where it is uncomfortable: a high and falling margin earns the same caution as a low and rising one earns goodwill.
What a margin will not tell you
A margin skips depreciation, interest and tax, which are the three things that genuinely cost money at a capital-intensive company. Machines wear out whether or not EBITDA notices, and replacing them is an expense rather than an opinion.
Which is why I read the margin next to the cash: when it rises and operating cash flow does not follow, either the margin is an accounting artefact or the money is still sitting with customers. Both are questions for management rather than grounds for pride.
What a thin margin costs in a recession
In a good year the difference between six and twenty-six percent is a difference in profitability. In a bad year it is a difference in the kind of problem. How large a difference?
Back to the two companies on 1 billion of revenue. Revenue falls by 10 percent, to 900 million, and in the first year the costs barely move, because buildings do not fold away and people are let go late. The first company drops below zero and goes to talk to its bank. The second is left with 160 million and goes to talk about what to buy from the first.
This is the same buffer from the opening, seen in the worse half of the cycle. A thin margin is not a fault in itself, because a whole industry can be built on thin margins. What a thin margin takes away is the right to make a mistake, and in a recession everybody makes one.
What to do with this in the next week
Take one company you hold and do three things on one sheet of paper. Compute its EBITDA margin for the last five years from the income statement itself, so that the formula is yours rather than somebody else's. Next to it, list the margins of a dozen companies from the same industry and mark which part of the distribution yours occupies. Then write one sentence about whether customers in that industry choose by comparing prices, because that sentence sets the ceiling for every other number on the page.
If the hour ends with your company in the bottom quarter of a thin-margin industry, you have the answer to the question in the title and you needed no threshold to get there.
In Taufolio the same loop is closed by the Full report: it splits profitability across the levels of the income statement, sets them against the company's own history, and leaves a reference to the source filing next to every figure. How that is put together is described in the method.
And if you only have time for one figure rather than five, take the margin from the year the industry had its worst quarter.
That is when you can see who had a buffer.
Frequently asked questions
How do you calculate EBITDA margin?
What is a good EBITDA margin?
Why do margins differ so much between industries?
Does a high margin mean a good company?
What does EBITDA margin leave out?
How many peers do you need to compare margins?
- fundamentals
- financial-statements
- moat
Posts are produced with AI tools and go through editorial review by the Taufolio team before publishing. How we write and correct posts