Margins

Two companies report to the market. One is a software firm with an operating margin of 20%; the other is a supermarket chain with an operating margin of 3%. Put like that the contest looks over - 20% dwarfs 3% - yet the supermarket can, and often does, earn a higher return on the money its owners put in, because it turns that money over many times a year while the software firm turns it slowly. The margin, read on its own, pointed the wrong way. illustrative

The mistake is treating a margin as a score. It is not a score; it is a description of cost structure. It answers one narrow question - of every hundred rupees of sales, how many survive down to this line? - and nothing else. It does not know how much capital the business needed, how fast that capital recycles, or whether the margin is durable or about to snap.

And it does not answer that one question once. Profit is not a single subtraction; it is a staircase of them. Revenue comes in at the top. The direct cost of what was sold is taken out, leaving gross profit. The running costs - staff, selling, administration - are taken out, leaving EBITDA, a rough proxy for operating cash profit. The wearing-out of the assets, depreciation, is charged, leaving EBIT, the operating profit. Then interest, then tax, leaving PAT, the bottom line. Each margin is simply one of those rungs written as a percentage of revenue - and the gap between any two rungs is a cost family. So when a margin moves, the useful question is never "did it go up or down?" It is which gap changed, and therefore which cost family moved? Read the levels alone and you see that profit shrank; read the gaps and you see why.

Lay the ladder out and read it downward, rung by rung, watching each gap.

A single level tells you almost nothing until you know whose business it is - a 9% net margin is superb for a supermarket and dismal for a software firm. But the gaps are diagnostic in any business. A wide gross-to-EBITDA gap means heavy running costs - a big sales force, or expensive marketing to hold a brand up. A wide EBITDA-to-EBIT gap means heavy depreciation, the fingerprint of an asset-heavy business with a lot of plant to wear out. A wide EBIT-to-PAT gap means interest and tax take a large share, often the sign of a leveraged balance sheet. When you compare a company with its own past or with a rival, compare the gaps - that is where the cost story lives.

Operating leverage lives between those rungs. Some costs are variable: they rise and fall with sales and hold their share of the rupee roughly steady. Others are fixed - rent, core staff, the depreciation on a plant - and do not move with sales. When revenue rises over a fixed cost base, those fixed costs are spread thinner, so margins expand without anything improving in the product itself. That is the healthiest kind of margin expansion, and it is the whole of Firm P's story in the quiz below. The lever works just as cruelly in reverse: when revenue falls, the fixed costs do not fall with it, and margins collapse faster than sales. A business with heavy fixed costs has a margin that swings hard in both directions, so reading its margin at one point in the cycle, without asking where the volume is, tells you very little.

One trap worth naming before the sectors: a margin is not a mark-up. A margin is profit as a share of sales; a mark-up is profit as a share of cost. A 50% mark-up on a good that costs ₹100 gives a selling price of ₹150 and a gross margin of only 33% - ₹50 of ₹150, not 50%. They are different fractions of different bases, and swapping one for the other overstates profitability every time. When a company or a broker quotes "a margin," check quietly which one they mean.

Here is where the margin most often lies. The same EBIT margin means completely different things in different businesses, because each industry has a different natural cost structure - so a fat margin can be the baseline in one sector and a warning in another, and comparing the levels across them answers nothing.

Two of those cells invert the naive reading, and both are worth holding onto. The retailer's thin margin is strength, because it is paired with fast capital turnover - a fact the margin alone conceals. The cyclical's fat margin is fragile, because it is struck on peak prices that will normalise; a fat margin at a cyclical peak is the moment for most caution, not most confidence. Most of what a single year's margin tells you about a cyclical is where prices happen to sit right now, not what the business durably earns - read it as a verdict and you buy the peak. The discipline is the one carried from the previous module: ask what question the margin is asking of this business - about its cost structure and where it sits in its cycle - before you let the level mean anything.

A composite branded-goods maker reports a net margin of 8%, down from 12% two years ago. The screener flags "margin compression," and the instinct is to conclude the business is deteriorating. Walk the ladder instead, and find which rung the money fell through. illustrative

Gross margin first: it has held at 44%, barely moved. So the product is fine - raw-material costs are being passed through, and the firm has not been forced to cut prices to sell. That single fact rules out the scariest explanation and redirects the search below the gross line. Next rung: the gross-to-EBITDA gap has widened sharply, because advertising and promotion spend has jumped from 9% of sales to 15%. There is the leak. The firm is spending heavily to defend or grow share against a new competitor - a very different story from "the business is deteriorating." It may be a deliberate, temporary investment in share, or a sign the brand has lost pricing power and must now buy its volume. Which one it is, the ladder cannot say; but it has told you exactly where to ask, and ruled out the product entirely.

Keep going. EBITDA-to-EBIT has barely changed, so depreciation is stable - no big new plant distorting the picture. EBIT-to-PAT has widened a little, because the firm took on debt to fund the push, so interest now takes a slightly larger bite. Put the walk together and the four-point fall resolves into one sentence: the product economics are intact; the firm is pouring money into marketing, part-funded by new debt, to hold its share. Whether that is a good decision is a judgement about the competitive fight - but you now know what you are judging, instead of a vague "margins fell." The level told you profit shrank; the ladder told you the story.

When margins move, a good analyst refuses the single-line headline and asks management to walk the ladder: "EBIT margin fell 300 basis points. Bridge it for us - how much was gross margin, how much the higher ad spend, how much depreciation from the new line? And which of those is structural versus temporary?" The question is a demand to know which rung the money fell through.

A good answer decomposes the move and separates the temporary from the structural:

"Happy to bridge it. Gross margin was actually up 50 basis points on better sourcing. The 300-point fall is entirely below the gross line: about 200 points is the step-up in brand investment, a two-year push we expect to taper; the other 100 is depreciation on the new plant, fixed now and earning its keep as volumes ramp. So none of it is product economics - it's investment and a fixed-cost step."

An evasive one names no rung and no number:

"Margins were a bit soft this quarter on cost inflation and some investments for the future, but we remain confident in our medium-term trajectory and industry-leading profitability. We expect margins to normalise as conditions improve."

"Cost inflation and some investments" names nothing; "normalise as conditions improve" is a hope, not a bridge. The follow-up that settles it - "give us the gross-to-EBIT walk in basis points, and split the ad-spend increase into defending share versus building new share" - forces the vague "investments" into a checkable story. If the room lets the soft answer stand, either gross margin fell too and management would rather not separate it, or the marketing spend is defence dressed as offence. A margin move that management will not decompose usually has an unflattering rung in it.

Even walked rung by rung, a margin has three blind spots, and each is a way careful people get fooled.

It is silent on capital. This is the biggest limit. A margin measures profit per rupee of sales, never per rupee of capital - and per rupee of capital is what the owner actually earns. A 20% margin on a business that swallows enormous plant and working capital can be a worse investment than a 4% margin on a business that needs almost no capital and recycles it many times a year. That is why the next module is return ratios: margin is only half the picture, and the returns supply the other half. Mistaking a fat margin for a good return is the single most expensive error the level invites.

It is silent on durability. A fat margin can be the reward of a genuine moat, or the temporary gift of a supply shortage, a cycle peak, or a competitor's stumble - and it looks identical in every case. Durability is a judgement about competitive position and cycle, not something the percentage carries within it; a high margin with no moat behind it is an invitation to the competition that pulls it back down. Trusting a cycle-peak or one-off margin is the same error wearing a time dimension.

A blended margin can hide a moving mix. A company selling one fat-margin line and one thin one reports a single blended number, and that number can hold perfectly steady while the profitable line shrinks and the cheap line grows underneath it. The single margin conceals the shift; only splitting the segments - the revenue mix - reveals it. Whenever a margin looks suspiciously stable in a changing business, suspect a mix moving under the surface.

Behind all three sits the first error of all: comparing margin levels across unlike businesses. A 20% margin and a 3% margin are not a ranking; each is normal for its cost structure. Margin is comparable only within a genuinely similar business model - and even there, only once you have asked what the level assumes about capital, durability and mix.