India now has a substantial set of listed consumer-internet companies that lose money on enormous revenue. The standard toolkit produces no answer — there is no P/E, book value is mostly cash from the IPO, and a DCF requires forecasting profits that do not yet exist.
Start with one customer
Strip the company down to a single transaction and a single customer. If the unit does not work, scale makes the problem larger — a company losing ₹40 per order loses more as it grows, and no amount of revenue growth fixes that.
The metrics that matter
| Metric | What it is | What good looks like |
|---|---|---|
| Take rate | Net revenue as a share of gross order value | Rising over time. A falling take rate means the company is buying volume with discounts. |
| Contribution margin | What remains after all costs that vary with each order — delivery, discounts, payment fees | Positive, and improving. Negative contribution means every order destroys value. |
| CAC | Marketing spend divided by new customers acquired | Falling, or at least stable. Rising CAC means growth is getting more expensive. |
| LTV / CAC | Lifetime contribution per customer against acquisition cost | Above 3 is generally considered healthy. Below 1 means each customer is a net loss. |
| Cohort retention | Whether customers acquired in 2022 still order in 2025 | The single most informative disclosure. Improving cohorts mean the product genuinely works. |
The two questions, in order
- 1Does one customer make money?
Positive contribution margin and LTV comfortably above CAC. If not, stop — scale is the enemy, not the solution.
- 2Will total contribution ever cover fixed costs?
Tech, salaries and corporate overhead do not scale with orders. Work out how many customers at current economics would be needed to cover them, then ask whether that number is plausible in the addressable market.
Adjusted metrics and what they hide
How to value one at all
- EV/Sales against comparable businesses — crude, and about the only relative measure available. Be far more sceptical than you would be with an earnings multiple.
- A DCF from an assumed steady state — model the business at maturity, at plausible margins, then discount back heavily for the risk of never getting there.
- A reverse DCF — work out what growth and terminal margin the current price implies, then judge whether that is plausible. Usually the most honest approach.
- Cash runway — how long the existing cash lasts at the current burn. A company that must raise equity in eighteen months has a dilution problem regardless of its economics.
App ₹100 ka saman ₹80 mein de raha hai aur delivery free — grahak toh badhenge hi. Sawaal yeh hai ki discount band karne pe grahak rukega ya nahi. Nayi companies mein revenue growth se zyada yeh dekho ki ek order pe paisa ban raha hai ya jal raha hai.
- No earnings means no P/E; judge one customer first, then the whole.
- Contribution margin must be positive or scale makes losses larger.
- Cohort retention is the most informative disclosure — decaying cohorts mean churn is being masked by acquisition.
- Passing unit economics does not guarantee covering the fixed cost base.
- Always ask what "adjusted EBITDA" adjusted out, and add it back.
Mark it done to track your progress through the curriculum.
Common questions
Short, direct answers to what people ask about this topic.
- contribution margin meaning in unit economics
- Contribution margin is what remains from a single order after every cost that varies with that order — delivery, discounts, payment gateway charges, packaging. It is the first test of whether a loss-making consumer-internet business works at all, because a negative contribution margin means each extra order destroys value and scale simply makes the loss larger. Fixed costs such as technology and head-office salaries sit below this line and are covered later, if at all.
- a healthy ratio of customer lifetime value to customer acquisition cost is generally taken to be
- Above 3 — a customer should generate at least three times in lifetime contribution what it cost to acquire them. Below 1 the customer is a net loss and faster growth only accelerates the burn. The ratio is only as reliable as the retention assumption inside LTV, which is why cohort disclosure matters more than the headline number.
- how do you value a company that has no profits
- Three approaches are available: EV/Sales against genuinely comparable businesses, a DCF built from an assumed mature steady state and discounted back heavily for the risk of never reaching it, or a reverse DCF that works out what growth and terminal margin the current price already implies. The reverse DCF is usually the most honest, because it converts a vague story into a checkable claim. Cash runway sits alongside all three — a company that must raise equity within a year or two carries dilution risk whatever its unit economics look like.
- what does adjusted ebitda exclude
- Whatever the company decided to exclude — adjusted EBITDA is not defined by any accounting standard, so its contents vary from filing to filing. Indian new-age companies commonly strip out employee stock compensation, which is a genuine cost paid in your ownership, and occasionally marketing spend reclassified as investment. Read the reconciliation back to reported numbers, add the adjustments in, and treat the gap between adjusted and reported as the figure worth studying.
- what is take rate for an online marketplace
- Take rate is net revenue as a percentage of gross order value — the share of every rupee transacted that the platform actually keeps. A rising take rate points to pricing power; a falling one usually means volume is being bought with discounts and incentives. Most listed Indian marketplaces disclose it, though the definition of gross order value differs between them, so compare a company against its own history before comparing it to a rival.