पाठशाला Pathshala · हिसाब Hisāb, Unit economics · Lesson 02 · Start

Contribution margin: the first number a founder must know

Gross margin says what the product costs to make. Contribution margin says what one more sale actually leaves behind after the courier, the gateway, the discount and the return. Compute it before anything else.

Pathshala, The Founder Library · 11 October 2026 · 9 min read

Every founder can say what the product costs to make. Far fewer can say what one more sale leaves behind once the payment gateway, the courier, the box, the discount that closed it and the one order in twelve that comes back have all been paid for. That residue is contribution margin. It is the first number to know because every other number in the business, from the units needed to break even to whether growth is worth buying, is built on top of it.

This lesson defines the number precisely, separates it from the gross margin it is usually confused with, works one order through to the rupee, and ends with the two uses that matter: break-even and the decision to grow. The figure in the middle runs your own product.

Gross margin is not contribution margin

Gross margin is revenue less the cost of the goods or service sold. Andreessen Horowitz’s 16 Startup Metrics tells founders that investors want all costs of manufacturing, delivery and support of the product inside that line, and want the founder able to explain what was included and what was left out. Most seed decks stop there. The slide says sixty per cent gross margin and the room nods.

Contribution margin keeps going. It subtracts every cost that occurs because this particular unit was sold and would not have occurred otherwise: the payment gateway fee, the forward courier, the packaging, the discount, the reverse courier and lost revenue on a return, a per-unit support cost if support scales with orders, a sales commission or a marketplace referral fee. What is left contributes to fixed cost and then to profit, which is where the name comes from. Rent, the salaries of the core team, software subscriptions and the accountant do not move when one more unit sells. They are fixed and they belong in the break-even arithmetic later, not here.

The test for any cost is one question: if the company sold nothing next month would this cost disappear? If yes it is variable and comes off contribution. If no it is fixed. The grey cases, a warehouse, a customer-support team, a delivery fleet, are fixed until volume forces them to grow in steps; treat them as fixed and revisit the choice each quarter. The labels CM1, CM2 and CM3 that Indian decks use for successive cuts of this number are not standardised, so define yours on the slide. Swiggy’s Q4 FY25 shareholder letter is a model of how to do it: food delivery contribution is adjusted revenue less delivery charges less platform-funded discounts less other variable costs, stated as a share of gross order value. One sentence and nobody has to guess.

One order, to the rupee

A company in Jaipur sells a ceramic planter online at ₹1,000 ex-GST. Every figure here is ex-GST because the tax collected belongs to the government and is not revenue; a founder who computes margin on the GST-inclusive price has overstated it by the tax rate before beginning. A ten per cent discount closes most orders, so net price is ₹900. The planter costs ₹380 to make and pack at the pottery. The payment gateway charges two per cent plus eighteen per cent GST on its own fee, which is Razorpay’s stated standard rate and a fair proxy for the market: ₹21 on a ₹900 order. The box and padding cost ₹25. The courier charges ₹70 one way. Eight orders in a hundred come back; the planter is restocked, the courier is paid again and the ₹900 is refunded.

Per order shipped, the expected revenue kept is 92 per cent of ₹900, which is ₹828. Cost of goods is charged on the 92 per cent kept, ₹350. The gateway fee is paid on every order, ₹21. Packaging is ₹25. The courier is ₹70 forward plus eight per cent of a ₹70 return leg, ₹76. Contribution is ₹828 less ₹350 less ₹21 less ₹25 less ₹76, which is ₹357 a unit, or 43 per cent of the revenue kept. The gross margin that would have gone on the slide, ₹900 less ₹380 over ₹900, is 58 per cent. Fifteen points of margin vanished between the pottery and the doorstep, and none of it was in the deck.

Why investors read this number before any other

Bill Gurley’s 2011 essay All Revenue is Not Created Equal put the reason plainly: a revenue stream saddled with large variable costs cannot generate much cash, so low-margin revenue is worth less per rupee than high-margin revenue, and public markets price it that way. Amazon, Walmart and Best Buy all traded at a fraction of their sales because twenty to twenty-five per cent gross margins leave little behind. The a16z follow-up, 16 More Startup Metrics, draws the line for private companies: e-commerce businesses typically have relatively low gross margins while software companies should sit in the eighty to ninety per cent range.

The mechanism is leverage. At 43 per cent contribution every additional ₹1 crore of sales sends ₹43 lakh toward fixed cost and profit. At eight per cent it sends ₹8 lakh, and the company must sell more than five times as much to pay for the same office and the same ten people. Low contribution is not fatal; some of the largest companies in the world run on it. But it fixes the kind of company you are building, a volume business in which every rupee of per-order cost has to be fought for, and a founder should know that on day one rather than discover it at Series A.

The figure is the Jaipur order by default. Move the return rate from eight to twenty-six per cent, which is what Unicommerce measured across marketplace orders in 2023, and watch the green bar lose a third of its height. Move the discount to thirty per cent for a festive sale and it is nearly gone. Then move the fixed-cost slider and read how many units a month the company needs.

From one unit to break-even

Monthly fixed cost divided by contribution per unit is the number of units the company must sell in a month to exist. The Jaipur company carries ₹18 lakh a month in salaries, a studio and software. At ₹357 a unit that is 5,042 orders a month, 168 a day, about ₹42 lakh of revenue kept. If it currently ships 1,400 orders a month the gap is the business plan, and it is a plan with a number: three and a half times current volume, or a higher price, or a lower courier rate, or some combination, by a date.

Compute a second, smaller break-even as well. Paul Graham’s 2009 essay Ramen Profitable defines the state in which a startup makes just enough to pay the founders’ living expenses, and argues that its main significance is that the company is no longer at the mercy of investors. Strip the fixed-cost line down to what it would be with founders on a minimal salary and no hires that are not yet essential, and divide again. That is the number of units a month at which nobody can shut you down. Write both numbers where the founders see them.

Revenue is what the customer pays. Contribution is what the company keeps. Only the second one can pay the rent.

Where the margin leaks

Discounts. At 43 per cent contribution an extra ten per cent off a ₹900 order is ₹90, a quarter of the contribution, given away to close a sale that might have closed anyway. A discount is a cost of sale and belongs above the line where it can be seen, not in the marketing budget where it cannot. Price the product so the list price is one you can hold, and treat every coupon as a line in this calculation.

Free shipping. It is not free. ₹70 on a ₹900 order is eight per cent of revenue, and the company pays it on the return leg too. A free-shipping threshold set a little above the current average order is one of the few levers that raises contribution and order value at once; shipping absorbed on every order is a discount with a different name.

Returns. Unicommerce’s India Ecommerce Index 2023 measured returns at 10.4 per cent of all orders in FY23, with marketplace orders returned at 26.3 per cent against 6.2 per cent on brand websites, and cash-on-delivery orders returned at 20.9 per cent against 5.8 per cent for prepaid. Each return costs two courier legs, a repack and the revenue, and a company that sells mostly through marketplaces on cash on delivery is running a different business from one that sells prepaid on its own site, even with the same product.

Marketing counted below the line. Contribution as defined here stops before customer acquisition, so that the product’s own economics can be seen. But a founder who spends ₹300 of advertising to win a ₹357 order is running a ₹57 business and should know it. Compute contribution after marketing as the next cut, and read the lesson on [CAC, LTV and payback](/library/cac-ltv-and-payback-the-three-numbers) before scaling spend.

Negative contribution is a plan, not a business

Swiggy’s Q4 FY25 letter reports food delivery contribution at 7.8 per cent of gross order value after a decade of work, and Instamart, its quick-commerce arm, at minus 5.6 per cent. Every Instamart order delivered that quarter cost the company money, and the company grew the business 101 per cent year on year anyway. It could do so because it had a listed balance sheet, a stated path to positive contribution and shareholders who had agreed to fund the gap. A seed-stage company has none of those things.

If contribution is negative the only acceptable reason to add volume is a tested route to positive by a date: courier rates that fall at a known volume tier, a price rise the first cohort has accepted, a mix shift toward prepaid orders, a product change that cuts returns. Write the route down with the number it must hit and the month. Growth before that is not growth. It is buying losses in bulk, and the larger the company gets the more expensive it becomes to stop.

The monthly check, in twenty minutes

On the first working day of each month, pull last month’s orders and compute contribution per unit as above, by channel and by payment mode: own site against marketplace, prepaid against cash on delivery. Compare each with the previous month. Recompute break-even units against the current fixed-cost line. Then ask three questions. Did contribution per unit fall while revenue rose? If so the growth was bought with discounts or a worse channel mix. Did the return rate move? If so find out which product and which pin codes. Is any variable cost up for renegotiation, because courier and gateway rates are tiered by volume and the tier you qualified for last quarter is not the one you are paying for now?

Decide one thing from the answers and write it down. The discipline is not the arithmetic, which a sheet does in seconds. It is looking at the number every month before someone else does.

Nothing here is legal, tax or investment advice. It is arithmetic on one unit, which is the only place a business can be understood before it is large.

Sources

1. Jeff Jordan, Anu Hariharan, Frank Chen and Preethi Kasireddy, 16 Startup Metrics, Andreessen Horowitz, August 2015 — Gross profit should include all costs of manufacturing, delivery and support; LTV is built on contribution margin.
2. Anu Hariharan, Frank Chen and Jeff Jordan, 16 More Startup Metrics, Andreessen Horowitz, September 2015 — E-commerce gross margins are typically low; software should be 80–90 per cent.
3. Bill Gurley, All Revenue is Not Created Equal: The Keys to the 10X Revenue Club, Above the Crowd, May 2011
4. Swiggy Limited, Q4 FY2025 Shareholder Letter, May 2025 — Definition of contribution margin; food delivery at 7.8 per cent of GOV, Instamart at minus 5.6 per cent.
5. Paul Graham, Ramen Profitable, July 2009
6. Unicommerce, India Ecommerce Index 2023 — Returns at 10.4 per cent of orders in FY23; 26.3 per cent on marketplaces against 6.2 per cent on brand websites; COD 20.9 per cent against prepaid 5.8 per cent.
7. Razorpay, Payment Gateway Pricing and Fees Explained, February 2026 — Standard domestic rate of 2 per cent plus 18 per cent GST on the fee, checked October 2026.