An online platform sits between buyers and sellers and keeps a slice of every transaction. The headline number everyone quotes, gross merchandise value, is mostly vanity, because it is easy to buy growth by subsidising customers into a loss. The honest question is whether the platform keeps more of each order than it spends to serve it, and whether customers come back on their own instead of being paid to return. Repeat purchases lower acquisition cost over time and are the real compounding engine; without them the model is just burning cash for GMV.
The six questions that explain this business
Answer these six and you have understood the industry. The rest of the page is the detail behind them.
Where does demand come from?
From people placing orders online, for food, groceries, cosmetics or almost anything, through a platform that connects them to sellers. The catch is that a lot of this demand can be manufactured: give a big enough discount and people will buy almost anything, so a chunk of the orders exists only because someone is subsidising the price. Real, durable demand is the part that comes back and buys again once the discounts fade. Until you separate the two, you cannot tell habit from a one-time bargain hunt.
Who controls the price?
Fierce competition, not the platform. With rivals one tap away and switching costs near zero, no platform can simply charge more; it competes on price, speed and selection, often by handing out discounts to win the order. The nearest thing to pricing power is take rate, the slice of each transaction the platform keeps, which only rises once it has enough scale and trust to charge sellers and buyers more without losing them. Early on, though, price is set by whoever is willing to burn the most cash.
What's the hardest thing to get?
A dense, low-cost logistics network and customers who come back on their own. Anyone can build a website, but getting orders to doorsteps quickly and cheaply, at scale, takes years and enormous investment, and it only gets cheaper once enough orders flow through the same routes. Just as scarce is genuine repeat custom, because a platform that has to pay for every order through discounts never really owns its customers. Delivery density and real loyalty, not code, are the true bottlenecks.
Where does the money disappear?
Into customer acquisition and delivery subsidies. The cash pours out on discounts to pull people in, on the cost of acquiring each new customer through marketing, and on delivery that often costs more than the platform charges for it. If the company loses money on every order before even counting salaries and warehouses, then more orders simply mean bigger losses. The money disappears order by order, which is why contribution margin, what is left from an order after its own variable costs, is the number that cannot lie.
What usually breaks first?
Cash burn and churn. Many of these companies subsidise prices below what the business can sustain, funded by investor money, which works only as long as fresh funding keeps arriving. The moment it slows, discounts have to shrink, the subsidy-dependent demand walks away, and the company that looked like it was scaling turns out to have been renting customers. Running out of runway before the unit economics turn positive, with customers churning as soon as the deals stop, is the dominant way these businesses die.
Why can't rivals just copy it?
Logistics density and network effects. Once a platform carries enough orders, each delivery route and warehouse gets cheaper per order, so it can serve customers at a cost a smaller rival cannot match. At the same time more buyers attract more sellers, and more sellers bring more choice that attracts more buyers, a loop that feeds itself. A newcomer can copy the app in weeks but cannot conjure the order density or the two-sided network that took years and billions to build.
The question beginners always ask
If an online store keeps offering discounts and free delivery, how will it ever make money?
The heavy spending early on is a deliberate bet, not the plan forever. The company pours money into discounts and building a fast, dense delivery network to get people to try it and turn shopping there into a habit, accepting a loss on those first orders. The wager is that once ordering becomes routine, it can quietly pull back the discounts and each repeat order starts to pay for itself, because a loyal customer costs almost nothing to serve again. The danger is that the customers only ever came for the deals, so the moment the discounts shrink they vanish, and all that spending bought rented growth rather than a real habit.
First, what is an e-commerce business really?
An e-commerce platform is a middleman that connects buyers and sellers online and takes a cut of every transaction that happens through it. The tricky part is that it is very easy to fake growth in this model by simply spending more money than you make on each order.
01
GMV is the total pie, not what the company actually earns
Gross merchandise value (GMV) is the total value of everything sold through the platform, including the part that goes to sellers, delivery partners, and payment processors. It is the headline number every company loves to quote because it always looks big and growing. But GMV is not revenue, and it is definitely not profit, it is closer to the total volume of traffic passing through a toll booth, not the toll collected.
For exampleA platform can report 10,000 crore rupees of GMV in a year while the company itself only earns 500 crore rupees in actual revenue, the rest belongs to the sellers and logistics partners on the platform.
02
You can buy growth, but you cannot buy loyalty
Give people a big enough discount, a 50% off coupon or free delivery, and they will buy almost anything, which is why GMV growth funded by subsidies is not proof of a good business. The real test is whether customers come back and buy again once the discounts fade, because that is what separates an actual habit from a one-time bargain hunt.
For exampleA food delivery app offering 60% off to first-time users will always show strong new-customer growth; the question is what percentage of those users order again a month later without a discount.
How to read an e-commerce business
01
Take rate shows how the platform actually gets paid
Take rate is the company's revenue divided by GMV, the slice of every transaction the platform actually keeps for itself, through commissions, delivery fees, and advertising sold to sellers. A rising take rate usually means the platform has built enough scale or trust that it can charge sellers and buyers more without losing them, a sign that monetisation is maturing.
For exampleIf a platform's take rate rises from 8% to 11% while GMV keeps growing, it is squeezing more revenue out of the same transaction volume, a much healthier sign than GMV growth alone.
02
Contribution margin is the one number that cannot lie
Contribution margin is revenue from an order minus the variable costs of fulfilling it: delivery, packaging, payment processing, discounts. If this is negative, the company loses money on every single order before even counting fixed costs like salaries and warehouses, meaning more sales literally means more losses. Positive contribution margin is the first real proof that the business model can work at all.
For exampleA quick-commerce company losing 20 rupees on every 500-rupee order, after delivery and discount costs, is not close to profit no matter how many orders it adds, it needs the unit economics to flip positive first.
Where e-commerce breaks, and how to value it
01
The subsidy trap: growth that disappears when the discounts stop
Many e-commerce companies chase GMV and user counts to raise the next funding round, using investor money to subsidise prices below what the business can sustain. This works as long as new funding keeps arriving, but the moment it slows, discounts have to shrink, and a chunk of the demand that was only there for the subsidy disappears with it. The company that looked like it was scaling was really just renting customers.
For exampleWhen a company cuts discounts to improve its path to profitability and its order volume drops 15-20% almost immediately, that gap is the real size of its subsidy-dependent, non-organic demand.
02
Value it on the path from GMV to profit, not GMV itself
Because these companies often lose money for years, valuing them on earnings does not work early on. Instead, track the trend: is contribution margin improving order by order, is CAC (the cost to acquire each new customer) falling as repeat purchases rise, and is there a believable, funded runway to turn overall profitable, cash burn, not GMV, before the money runs out. A shrinking loss with rising repeat rate is worth more than a bigger loss with bigger GMV.
For exampleA company burning 200 crore rupees a year with contribution margin improving each quarter and repeat rate climbing is on a real path to profit; one burning the same amount with flat repeat rate is just spending to stand still.
The five numbers that decide the story
Every sector reduces to a demand metric, a pricing metric, an efficiency metric, a capital metric, and a risk metric. For e-commerce, these are the ones that matter.
Demand
GMV / order volume
Pricing
Take rate
Efficiency
Logistics cost
Capital
Contribution margin
Risk
CAC / churn
The full metric set
What each number tells you, and how to read it.
Metric
Why it matters
GMV (Gross Merchandise Value)
Total transaction value on the platform. A scale metric, but inflated by high-return categories. Watch take rate.
Take Rate
Revenue over GMV, how much of each transaction the platform keeps. Rising take rate means monetisation is working.
CAC (Customer Acquisition Cost)
The cost to win each new user. Rising CAC with flat repeat rates means the unit economics are breaking.
Repeat Purchase Rate
Loyalty. High repeat means organic growth and lower CAC over time, the compounding engine.
Contribution Margin
Revenue minus variable cost per order. Positive is a path to profit; negative means still subsidising growth.
Logistics Cost
A major expense. Owning versus outsourcing logistics changes the margin structure; higher order density lowers per-unit cost.
One sentence to remember
E-commerce spends heavily to build a habit and a delivery network, and only pays off if customers stay once the discounts stop.