Marketplace model

What is a Marketplace model?

In a marketplace you do not own the inventory and you do not deliver the service. You own the meeting point. Sellers list what they offer, buyers find it, and the platform earns money for making that match happen safely and reliably. The business you are actually running is trust plus discovery, not supply.

That has an appealing consequence and a brutal one. The appealing part is that growth does not require you to buy stock or hire service providers, so the model scales in a way that a traditional retailer cannot. The brutal part is that you have to build demand and supply at the same time, and each one is nearly worthless without the other. Buyers will not come to an empty catalog, and sellers will not list where no buyers look.

Marketplaces also carry work that pure software products avoid: payments, disputes, fraud, quality control, and the constant temptation of both sides to transact off platform once you have introduced them. Those are not edge cases. They are the product.

How marketplaces make money

Model	How it works	Best when
Commission (take rate)	A percentage of each transaction	You process payment and add real value to the match
Listing fee	Sellers pay to post	Supply is abundant and buyers are scarce
Subscription	One side pays a flat monthly fee	Transactions are hard to track or happen offline
Promoted placement	Sellers pay for visibility	The catalog is large enough that ranking matters

Commission is the default because it aligns your revenue with value delivered, but it only works if the transaction actually runs through your platform. If you merely make introductions, subscriptions or listing fees usually survive longer.

Why the marketplace model matters for startups

The decision a marketplace forces on a small team is which side to build first, and how narrowly to start. The standard answer is to solve the harder side first, which is almost always supply, and to constrain the market geographically or by category until liquidity is achievable at your size.

Liquidity is the real metric, not signups. A marketplace is working when a buyer who arrives with intent finds something acceptable, and a seller who lists gets a response within a reasonable window. Ten thousand listings across an entire country is a ghost town. Two hundred listings in one city can be a functioning market.

The other founder-level consequence is that you carry two acquisition costs, not one, and both feed the same transaction. Your unit economics have to work after paying to acquire both a buyer and a seller.

Marketplace model in practice

Imagine you are building a marketplace for studio time: musicians booking rooms by the hour. Launching nationally gives you 400 studios spread so thin that most cities show two options, and buyers leave.

Instead you pick one city and personally onboard 40 studios, photographing rooms and setting up calendars yourself. Coverage in that city is good enough that a musician searching a Tuesday evening finds five real options. Bookings start, and each completed booking earns a 12 percent commission. Once weekly bookings hold steady for two months, you copy the playbook to the next city.

The unglamorous manual work is the strategy, not a shortcut. Marketplaces that scale are almost always assembled by hand at the start.

Benchmarks and rules of thumb

Take rates vary enormously by category, but many software-enabled marketplaces land somewhere between 10 and 20 percent, with lower rates where transaction values are high and the platform adds little beyond discovery, and higher rates where the platform handles payments, guarantees, insurance, or dispute resolution. Categories with expensive goods usually cannot support double digit percentages at all. A useful sanity check: if sellers routinely try to move the transaction off your platform after the first match, your take rate is ahead of the value you provide, or you are not providing enough ongoing reason to stay.

Common mistakes

• Launching both sides everywhere at once. Thin coverage looks like an empty store to everyone. Concentrate on one city, category, or niche until it feels full.
• Measuring listings instead of liquidity. The number that matters is what share of searches end in a transaction and what share of listings get a response.
• Ignoring leakage. If both sides can complete the deal without you, give them a reason to stay: payment protection, scheduling, records, guarantees.
• Setting the take rate before proving value. Start lower than you think and raise it once the platform is clearly doing work neither side wants to do themselves.
• Skipping quality control. One bad seller experience poisons a buyer for the whole platform, not just that listing.

Related concepts

Marketplaces are the classic home of the network effect, where each additional participant makes the platform more useful to everyone on the other side. Choosing between commission, subscription, and fees is a monetization strategy decision that is hard to reverse later, and because you pay to acquire both sides, your customer acquisition cost deserves closer attention here than in most models. Growth also depends on scalability of the manual operations you start with.