In NYU Stern’s January 2026 dataset of U.S. public companies, the apparel sector earned a gross margin of 56.88% and kept a net margin of 3.85%. Advertising earned 36.24% gross and finished the year at negative 0.30% net. Two sectors, two very different gross margins, and almost no relationship between the top number and the bottom one.
That gap is the whole subject. What gross margin reveals about business health is narrower and more useful than the usual advice suggests. It is not a scorecard. It is a budget. It tells you how much of every dollar of sales survives the cost of delivering the thing you sold, and therefore how much is left to pay for everything else you do. Whether that budget is enough, and whether you spend it well, are separate questions that the ratio cannot answer.
The mechanics, quickly
Gross profit is what remains after you subtract the cost of sales from net revenue. The U.S. Securities and Exchange Commission’s investor guide describes the income statement in exactly that order: cost of sales comes off net revenues to arrive at gross profit, and operating expenses are deducted after that. U.S. securities rules take it seriously enough that 17 CFR § 210.1-02 requires disclosure of gross profit, or alternatively the costs and expenses applicable to net sales. Express gross profit as a percentage of revenue and you have gross margin.
One correction before anything else, because it quietly distorts pricing in small businesses every day: margin is not markup. Add 50% to a $10 cost and you get a $15 price. The gross profit is $5, which is 33.3% of the selling price, not 50%. Owners who set prices on markup and then plan their overheads on the assumption that they are earning the markup percentage end up short by exactly that difference, every month, without ever seeing the error on a report.
Gross margin is a budget, not a score
Here is the more productive way to use the number. Your gross margin percentage is the share of each sales dollar available to fund rent, salaries that are not tied to delivery, software, marketing, interest, tax and whatever profit you intend to keep. So the question is never “is 42% good?” The question is: how much fixed cost does this business carry, and how much revenue do I need at this margin to cover it?
The Small Business Administration teaches break-even with contribution margin rather than gross margin, and the distinction matters. Contribution margin is (sale price per unit minus variable cost per unit) divided by sale price per unit, and break-even in sales dollars is fixed costs divided by that contribution margin. Gross margin is usually close enough to be useful, but it is not identical, because cost of sales often contains fixed elements: a salaried production team, depreciation on equipment, a minimum cloud commitment. Those costs do not disappear when volume falls.
Run the arithmetic anyway. A studio with $18,000 of monthly fixed costs and a 55% gross margin needs roughly $32,700 in monthly sales to break even. Drop the margin to 45% and the requirement jumps to $40,000. A ten-point margin change did not reduce profit by ten percent. It raised the volume of work the business must win and deliver every month by more than a fifth, with the same team.
That is the first real diagnostic. A gross margin is healthy when it covers your fixed base at a volume you can actually sell and deliver, with room left over. A margin that is average for your industry and still cannot do that is not average for you. It is a problem.
Rate versus quantity
Percentage margin is a rate. Cash is a quantity. Businesses fail from quantity problems, which is why the obsession with the percentage sometimes points in the wrong direction.
Consider the arithmetic of velocity. A distributor working at a 22% gross margin that turns its inventory ten times a year generates 2.2 dollars of gross profit per dollar tied up in stock. A boutique at 55% that turns stock twice generates 1.1. The lower-margin operator has the better business on that measure, and no amount of staring at the margin line will reveal it. The same logic applies to service firms, where the binding constraint is delivery hours rather than shelf space. Gross profit per billable hour is often a sharper instrument than gross margin percentage, because it exposes the engagement that looks profitable on paper and consumes three weeks of your best person. If you have not mapped this, the unit economics underneath each sale are where the work starts.
Where the COGS line gets drawn, and why your number may not be comparable
Two businesses doing identical work can report gross margins ten points apart purely through classification. Service firms have the hardest time with this. Does the project manager sit in cost of sales or in overhead? What about the account manager who is half delivery and half renewal? Customer support? The software your delivery team uses but your sales team does not?
There is no universal answer, but there is a usable rule: if the cost grows when you deliver more, it belongs in cost of sales. If it stays flat whether you deliver one project or thirty, it is overhead. Judgment calls will remain at the edges. Resolve them once, write down the decision, and never quietly change it, because the value of gross margin comes almost entirely from comparing it against your own prior periods. A reclassification that improves the margin by six points has improved nothing.
The same caution applies when you compare yourself to anyone else. The SEC’s own guidance notes that non-GAAP measures are company-defined, and in practice “adjusted gross margin” in published filings can exclude stock-based compensation, amortisation, depreciation or other charges, depending on who is reporting. Outside the public markets the inconsistency is worse. When a competitor claims a 70% margin, you are hearing a number built from unknown ingredients.
When the margin moves, find out which of four things moved
A margin change is a symptom. Four drivers explain nearly all of them, and they have completely different implications.
Price
You raised prices, or you stopped discounting. This is the one everybody hopes for, and it is the one most likely to be temporary if it came from a single large client or a quiet reduction in promotions rather than a change in what you are worth to buyers. Sustained price-led margin expansion is the strongest signal of health on this list, because it means the market accepted a value claim.
Mix
This is the driver that fools people, and the clearest illustration comes from blended business models. In its Q2 2024 results, Shopify reported total company gross margin of 51.1%. Underneath that single figure sat Subscription Solutions at 82.8% and Merchant Solutions at 39.1%. The consolidated number describes neither business. A year earlier, in Q2 2023, gross margin had fallen to 49.3% from 50.7%, and the company attributed the decline to a higher mix of lower-margin Merchant Solutions, primarily lower-margin revenue from Deliverr. Nothing had deteriorated in the software. The sales simply came from a different place.
The practical instruction is blunt: split before you celebrate. If you sell more than one type of thing, a company-wide gross margin is close to meaningless as a management number. Calculate it by product line, by service type, by channel, by customer segment. Then look at what the weighting did. Half the margin movements you have ever worried about were weighting. Working out which part of the business actually carries the profit is the direct application of this.
Input cost
Materials, freight, wages, hosting. The last one deserves attention because the stale advice about software is now visibly wrong. UiPath’s fiscal 2025 gross margin fell to 83% from 85%, and the company pointed to a lower proportion of higher-margin license revenue alongside rising costs, including a $25.1 million increase in third-party hosting and software services and a $23.1 million increase in personnel-related expenses. The old line that software is “naturally” an 80% to 90% gross margin business is a description of some companies in some periods, not a law. Infrastructure and service-delivery costs are real cost of sales, and they are rising in plain sight in current filings.
Cost to deliver
The product changed shape. NVIDIA’s fiscal 2025 GAAP gross margin was 75.0%, up from 72.7% the prior year, which is a strong year by any standard. In the fourth quarter alone, GAAP gross margin was 73.0% against 76.0% a year earlier, and the company’s explanation was the transition to more complex and higher-cost systems within Data Center. That is a deliberate, strategic margin decline. A buyer reading only the quarterly percentage would reach the wrong conclusion about what was happening in that business.
Which leads to a position worth stating plainly: a falling gross margin is not automatically bad, and a rising one is not automatically good. Deliberately absorbing lower margin to own logistics, payments, hardware or implementation can be the right call if it buys retention or share. The failure mode is doing it accidentally and discovering it two years later. Treat every margin movement as something that requires an explanation in one sentence, and treat “I’m not sure” as the finding. Often it points to a cost of serving customers that was never fully counted.
The high-margin business that is not healthy
Gross margin tells you the model can work. It does not tell you that it is working.
Snowflake’s fiscal 2026 results make the point better than any argument. GAAP product gross margin of 72%, non-GAAP product gross margin of 76%, and a GAAP operating income margin of negative 31%. On a non-GAAP basis, operating margin was 10%. Excellent unit economics, and an operating result that depends entirely on how much is being spent below the gross profit line. Anyone who treats a 70%-plus gross margin as proof of a healthy business has to explain that income statement.
There is a second, quieter version of the same lesson. DocuSign reported a fiscal 2026 GAAP gross margin of 79.4%, barely changed from 79.1%, while annual recurring revenue grew 8.0% to $3.272 billion. Strong margin, moderate growth. Margin quality and growth momentum are independent dimensions, and collapsing them into a single judgment about “a great business” throws away most of the information.
Then there is cash. Gross profit is an accounting result, not money in the account. A business can post a rising gross margin while its gross profit sits in unsold inventory and unpaid invoices, which is exactly how growing sales create cash-flow problems. The honest sequence is: gross margin, then operating expense discipline, then working capital, then cash. Skipping steps is how owners end up surprised.
Benchmarking without being misled
Sector data is worth a glance for order of magnitude. The January 2026 NYU Stern figures put apparel at 56.88% gross margin, advertising at 36.24%, air transport at 24.79%, aerospace and defense at 17.48%, and auto and truck at 10.41%. That dispersion is the only universal lesson available: there is no single healthy gross margin, and any article offering one number for all businesses is selling a shortcut that does not exist.
Read those figures next to net margin, though, and the ranking scrambles. Apparel’s 56.88% gross converts to 3.85% net. Air transport’s 24.79% gross converts to 2.51% net. Advertising’s 36.24% ends up slightly negative. Gross margin predicts final profitability far less reliably than the usual “higher is better” framing implies.
Official data does not rescue you either. The Census Bureau’s Annual Integrated Economic Survey, which replaced and integrated seven prior annual business surveys starting in 2023, reports gross margin for nonstore retailers in 2024 as an aggregate dollar figure of roughly $560 billion, not a percentage you can hold yourself against. It is good economy-wide data and a poor personal benchmark.
So take the position that follows from all of this. Your own trailing twelve months, split by segment, is a better benchmark than any industry average, because it holds your accounting choices, your customer base and your business model constant. Use sector data to notice that you are at 15% in a business where peers run near 50%, which is worth investigating. Do not use it to set a target. A target borrowed from a dataset of public companies with different scale, mix and cost classification is a target you cannot interpret when you miss it.
Improving the number without damaging the business
Three moves do most of the work, and one trap catches people.
Price is first because it drops straight through. Nothing else in the business has to change. The limiting factor is usually confidence rather than market tolerance, and the signals that a pricing model has stopped working tend to be visible long before anyone acts on them.
Mix is second and underused. You do not have to fix the margin of your worst line. You can sell less of it. Steering sales effort toward the segments that already carry good margins changes the blended number without a single operational change, and it is reversible if you get it wrong.
Cost to deliver is third and slowest. Renegotiating supply, reducing rework, cutting scope creep, automating a delivery step. Real, durable, and it takes quarters rather than weeks. Several of the same levers show up when you look at improving profit without adding customers.
The trap: optimising the percentage while shrinking the dollars. Dropping your largest, lowest-margin account will raise your gross margin and may leave you with less gross profit than before and the same fixed costs. Percentage up, business worse. Before any margin initiative, check what happens to total gross profit dollars, because that is the figure your overheads are paid from. This is also where the distinction between revenue growth and profitable growth becomes concrete rather than theoretical.
A short monthly discipline
Four questions, asked against your own numbers, will extract more value from gross margin than any benchmark table.
- What is the margin by segment, and how did the weighting between segments change?
- If the total moved, was it price, mix, input cost or cost to deliver? One sentence, with evidence.
- At this margin, what revenue do I need to cover fixed costs, and how far above it am I running?
- How much of this period’s gross profit turned into cash, and where is the rest sitting?
A margin that holds steady to the decimal point month after month in a small business is usually not stability. It is an accounting routine that nobody has examined. Real operations move. The value of this number was never in the number itself. It is in the explanation you are forced to produce when it changes.
