Two-Sided Markets: The Shopping Centre With No Shops

Imagine opening a shopping centre. You have the building. You have no shops and no shoppers.

You approach a retailer. They ask a reasonable question: how many people walk through here? None yet, you admit — you have just opened. They decline. Sensibly, because a shop in an empty centre earns nothing.

So you try shoppers instead. They ask what shops you have. None, you admit. They also decline, and for the same reason.

Neither side is being difficult. Each is waiting for the other, and each is right to wait. This is the chicken-and-egg problem of a two-sided market, and it is the defining challenge of every marketplace business ever built.

A two-sided market is a platform whose value comes from connecting two distinct groups who each want the other to be there — buyers and sellers, riders and drivers, diners and restaurants, readers and advertisers. The platform itself produces nothing much. What it sells is access to the other side.

What makes them different from ordinary businesses

An ordinary business improves by making its product better. A two-sided platform improves by getting more of the other side, which is a strange thing to have as your core competence.

Three consequences follow, and they explain a lot of otherwise baffling behaviour.

Growth is self-reinforcing once it starts. More sellers make the platform more useful to buyers, which attracts more buyers, which attracts more sellers. This is a network effect — specifically a cross-side one, where each group benefits from growth in the other group.

Pricing can be deliberately lopsided, even free or negative. Because the two sides are worth different amounts, platforms routinely lose money on one side to attract the other. Nightclubs let women in free. Card networks charge merchants and reward cardholders. Newspapers historically sold below cost and made it back from advertisers. This looks irrational until you realise the subsidised side is the product being offered to the paying side.

The hard part is the beginning, not the scaling. A platform with nobody on it has essentially zero value, so the early period is unlike any other kind of business — you are asking people to join something that does not yet work.

How the chicken-and-egg problem actually gets solved

Nobody solves both sides simultaneously. The workable approaches all involve breaking the symmetry.

Pick the harder side and buy it. Usually one side is more difficult to acquire — often supply, because sellers need a real reason to invest effort. Platforms routinely pay that side, guarantee them income, or recruit them individually by hand. This is why so many marketplaces lose money for years and why that is sometimes a deliberate strategy rather than a failure.

Shrink the market until it is dense. A platform with a thousand users spread across a country is useless; the same thousand in one neighbourhood is genuinely valuable. Winning a tiny market completely beats a thin presence everywhere. Density, not size, is what makes a marketplace work early on.

Be useful to one side with no other side present. The strongest opening move available. Offer a tool that is worth using alone — inventory management, a booking calendar, a portfolio page — and let the marketplace form later on top of the users you have accumulated. This sidesteps the problem entirely rather than solving it.

Fake it, honestly. Fill one side yourself at the start: buy the inventory, employ the first providers, do the work manually behind the interface. Legitimate as a bootstrapping method, provided nobody is deceived about what they are getting.

Why some platforms are far weaker than they look

The reflex is to assume marketplaces become unassailable monopolies. Some do. Many are much more fragile, and two specific properties decide which.

Multi-homing. If participants can cheaply use several platforms at once — a driver running two apps, a seller listing on three sites, a diner with four delivery apps — then the network effect is badly diluted. Nobody has to choose, so no platform captures anyone exclusively. Multi-homing is the single most reliable solvent for marketplace power, and the strategies platforms use against it (exclusivity terms, loyalty tiers, penalties for taking business off-platform) are usually where they attract regulatory attention.

Local versus global networks. Some networks are only valuable locally. Riders in one city do not benefit from drivers in another. That means a competitor does not need to beat the incumbent everywhere — only in one city at a time. A business that looks like a global giant may really be a portfolio of separately contestable local monopolies. Contrast that with a network where every user genuinely benefits every other user regardless of geography, which is far harder to attack.

There is a third pressure worth naming. A platform that becomes dominant faces a standing temptation to increase its take rate, since neither side can easily leave. Each increase is profitable and each one raises the return to a challenger. This is economic moats being harvested rather than maintained — the classic way strong positions are eventually given away.

The shopping centre problem never fully disappears. It just changes shape: first you cannot get anyone to come, and later you cannot resist charging them too much for having come.