Start with the definition, and watch the one word doing the work. Contribution per unit = revenue per unit − all costs that vary with the unit. That doesn't mean "COGS", and it doesn't mean "the costs management labels variable". It means every one of them: product cost, shipping, packaging, payment-gateway fees, returns processing, per-ticket customer support, the voucher, the referral bonus, the rider incentive.
Which costs belong in contribution margin?
The test looks at how a cost behaves, whatever it's called. Ask the question if this unit had not happened, would this cost have happened? If the answer is no, the cost is variable. It belongs above the contribution line. You classify by behaviour and ignore the P&L caption. Performance-marketing spend is filed under "advertising" on the P&L, but it's variable per acquired customer. A streaming service's content is filed under "cost of revenues", but it's fixed with respect to the next subscriber. That's the Netflix operating-leverage lesson, restated per unit.
Don't confuse contribution margin with gross margin. Gross margin is an accounting presentation, while contribution is an economic answer. Take an e-commerce company with a 35% gross margin and ₹205 of fulfilment cost per ₹1,800 order. Its contribution margin is eleven points lower than its gross margin, and that gap is the money that vanishes if you naively multiply gross margin by GMV growth.
Advertisement
What are CM1, CM2 and CM3?
The CM ladder. Indian consumer-tech reporting has standardised a three-rung ladder in its DRHPs and investor decks. Zomato/Eternal, Nykaa, Honasa and Delhivery all use versions of it. You need to be able to build it and interrogate it:
- CM1 = net revenue − product COGS, the merchandising margin. Is the thing sold for more than it cost?
- CM2 = CM1 − variable fulfilment: shipping, warehousing pick-pack, packaging, payment gateway, per-order support, returns logistics. Is a delivered order profitable?
- CM3 = CM2 − variable marketing: performance ads, vouchers, first-order discounts (where not already netted from revenue). Is a delivered order profitable after paying for the demand?
Below CM3 come the costs that really are fixed: technology, brand marketing, corporate. Take those off CM3 at scale and you have EBITDA.
Contribution margin per order example: GlowKart
Here's one order at "GlowKart", a synthetic firm modelled on the scale and shape of a listed Indian beauty e-tailer like Nykaa. Once you've followed it, pull the listed company's actual investor deck and rebuild it with their numbers.
| Per delivered order | ₹ | % of net revenue |
|---|---|---|
| Net revenue per order (AOV net of returns, GST, and discounts-as-contra-revenue) | 1,800 | 100.0% |
| − Product COGS | 1,170 | 65.0% |
| CM1 (product margin) | 630 | 35.0% |
| − Shipping & last-mile | 110 | 6.1% |
| − Warehouse pick-pack & packaging | 38 | 2.1% |
| − Payment gateway (~1.8% of AOV) | 32 | 1.8% |
| − Customer support + returns processing (allocated per order) | 25 | 1.4% |
| CM2 (fulfilled-order margin) | 425 | 23.6% |
| − Variable marketing per order (performance spend ÷ orders) | 240 | 13.3% |
| CM3 | 185 | 10.3% |
| − Fixed costs per order at current scale (tech + brand + corporate ÷ orders) | 145 | 8.1% |
| EBITDA per order | 40 | 2.2% |
Read the ladder like an X-ray. CM1 at 35% says merchandising works. That 11.4-point drop from CM1 to CM2 is the cost of moving atoms. It shrinks as AOV rises, because a ₹3,600 order pays the same ₹110 shipping, and it shrinks with density. So AOV and order density are e-commerce's real margin levers.
The 13.3-point drop from CM2 to CM3 is the cost of renting demand, and it hides the first big trap. That ₹240 blends two different things. Vouchers on repeat orders are a true per-order variable cost. CAC on new customers is an investment in a future stream. A business with high repeat rates has a CM3 close to CM2. A business that has to re-buy every order has CM3 ≈ CM2 − CAC-per-order, which may never go positive. So before you admire a CM3, ask for the new/repeat split.
Common mistakes when reading contribution margins
You need two more habits for reading ladders. First, read the definitions note before you compare anything. The ladder is a convention, and no standard governs it. Eternal defines contribution as a % of GOV, so a "7% contribution margin" there is on the gross order value base. That's the point about bases, turning up again. Nykaa's ladder is on net revenue. A logistics company's CM2 might include line-haul that a marketplace leaves out. Comparing CM across companies without checking definitions is the unit-economics version of comparing a franchisor's revenue to an operator's.
Second, watch where the discounts go. Under Ind AS 115/ASC 606, consideration payable to customers, such as a voucher redeemable against the bill, is contra-revenue. It reduces the top line. Some companies book aggressive discounting as "marketing expense" instead. That inflates net revenue and CM1, and it parks the pain lower down, where it reads as "investment". Same cash, different story. The reconciliation test in the games section catches it.