Marketplaces, platforms and moats
Some products get better as more people use them: more shops bring more customers, and more customers bring more shops. This lesson explains how network effects really work, how to start a marketplace from nothing and measure whether it works, how much a platform can keep from each sale, how a three-sided marketplace of customers, shops and riders stays in balance, and where lasting advantages come from.
By the end you can
- Tell direct, cross-side and data network effects apart, and say whether a network effect is local or global
- Plan a cold start around the smallest network that works, and decide which side to seed first
- Measure marketplace liquidity with fill rate, time to match and utilisation, per neighbourhood
- Set and defend a take rate, and spot the signs that it is too high
- Name a product's moat with the seven powers, and test whether it would survive a well-funded copy
The everyday picture
A town has a busy Saturday farmers' market. Shoppers come because there are forty stalls to choose from; farmers come because there are thousands of shoppers. Someone opens a second market on an empty lot across town. On its first Saturday it has six stalls and a few curious visitors. The shoppers find little to buy and don't come back, so the farmers sell nothing and don't come back either.
A smart organiser starts small: one row of stalls in a corner, not a whole field. She gives the first farmers their stalls free for a month, picks the morning when the neighbourhood is out anyway, and fills the gaps herself by selling bread from her own bakery. Once shoppers find what they want, more farmers ask for stalls, and she can start charging a stall fee. If the fee grows too high, farmers start selling from their trucks by the road. That is a marketplace: two groups who need each other, a hard start, a fee that has to stay fair, and an advantage that grows with size.
Network effects: when each user makes it better
A product has a network effect when each new user makes it more valuable to other users. There are three kinds, and they differ a lot in strength:
- Direct. Users benefit from other users of the same kind. A messaging app such as WhatsApp is worth more when your family and friends are on it.
- Cross-side. Two or more groups benefit from each other. More shops on a delivery app means more choice for customers; more customers means more sales for shops. This is what makes a marketplace.
- Data. More use produces data that makes the product better: more trips give better delivery time estimates. These are real but weaker, because the improvement usually levels off.
Two more questions decide how strong a network effect is. Is it local or global? A customer in one city gains nothing from riders in another, so ride-hailing and grocery delivery are won neighbourhood by neighbourhood. A travel platform is more global: guests from any country help hosts everywhere. Can it turn negative? Too many sellers competing for too few buyers, too many riders waiting for orders, or spam in a messaging app all make the product worse as it grows.
Not every product with many users has a network effect. A calculator app with a million users is no better for the millionth. Most internal tools have none either: the Store app doesn't improve for Ama because another store uses it. Some internal platforms do, though. A request desk where every team sends work to every other team becomes the one place to look as more teams join.
Quick check
Which of these has the most local network effect?
Delivery value depends on supply close to the customer, so each neighbourhood is its own network and has to be won separately. Videos, international family chats and code repositories all benefit users wherever the other side is.