Platforms and network effects

A network effect means each new user adds value for the users already there. A telephone is the classic case: one phone is useless, and a million phones connect you to almost everyone. The same logic applies to messaging apps, payment networks and online marketplaces.

This is different from ordinary scale. A big factory lowers its own cost per unit. A network gets better for the customer as it grows. That is why investors pay close attention to whether a startup has a real network effect or just a large number of users.

Direct network effects occur when users benefit from other users of the same kind. A messaging app is the clearest example: it matters who else is on it. Indirect, or cross-side, network effects occur when two groups attract each other: more shoppers draw in more sellers, and more sellers draw in more shoppers. Food delivery, ride-hailing and job sites work this way.

Data network effects arise when more use makes the product better, for example a map or a recommendation system that learns from every trip or purchase. Protocol or standard effects occur when a shared system becomes the default; India's UPI is an open payment system that many apps and banks use, and the more people and merchants accept it, the more useful it is to join.

A platform is a product that connects groups of users who create value for each other, instead of selling something it made itself. A marketplace links buyers and sellers. A payment network links payers and merchants. A social app links creators and viewers.

Platforms usually earn by taking a share of each transaction (a take rate), by charging one side a fee or by selling advertising. They are often asset-light, since the sellers own the inventory or the vehicles, which allows rapid growth but gives the platform less control over quality.

A new platform has a hard start: buyers will not come without sellers, and sellers will not come without buyers. Founders solve this in a few ways. They begin in one city or one niche, where a small crowd can be enough. They subsidise one side, for example with free listings or discounts, until the other side shows up. They bring one side with them, such as a founder's existing suppliers. And they sometimes provide the supply themselves at first.

Many Indian quick-delivery and ride-hailing companies launched city by city for this reason, and spent heavily on incentives to get both sides active. The spending only makes sense if the network eventually becomes strong enough that users stay without it.

Many network effects in the real world are local. A ride-hailing app needs drivers near you, not drivers in another country. A delivery platform needs restaurants in your neighbourhood. So a leader in one city can be beaten by a rival there even though it is bigger elsewhere. This is why many such markets in India stay competitive, with several strong players per city.

Purely digital networks, like a messaging app, can be global, which tends to produce one or two large winners. Understanding whether an idea is local or global helps judge how defensible it is.

No. Users can often join several networks at once, called multi-homing: drivers work for more than one app and shoppers compare prices across sites. When switching is easy, the network effect is weak and competition falls back to price and service. Networks can also lose value when they get crowded with low-quality users or sellers, or when the platform fails to keep out fraud.

Rivals can attack with a better experience for one side, cheaper fees, or by building an open standard. Regulation is another factor: competition authorities, including the Competition Commission of India, look closely at large platforms that favour their own products or use their position unfairly.

Ask whether the product is clearly better for a user when more people join, and whether the evidence shows it. Signs include users inviting others without being paid, falling cost of winning each new user, rising repeat use per user, and a growing share of activity on the platform that needs no incentives. Weak signs are growth that stops when discounts stop, or a large user count with little repeat activity. This guide is general information, not investment advice.