Marketplace

Buying and selling has always carried a search cost. Someone with a used bicycle to sell had no idea where the buyers were, and the buyer had to check five neighborhood boards. Finding each other didn't end the problem either. Send the money first and the item might never arrive; ship the item first and the money might never come.

The same friction sat on every category where the two sides were scattered. Hiring a freelancer worked the same way. A small company that needed a logo knew no designers and had to ask around for an introduction, while designers dug through their own contacts looking for work. Both sides knew the other existed somewhere, and neither had a path to reach it.

A marketplace takes on both problems at once. It puts the two sides in one place so they can find each other, and it holds the money in the middle until the work is done, which absorbs the trust problem too. The product doesn't sell anything itself — it makes other people's transactions happen on top of it and takes a cut. That makes its revenue structure different from every other product form.

It takes on the time and risk one side spends looking for the other. Four things, specifically.

1. Search. One screen with filters replaces the hunt. Contacting thirty designers individually to ask their rates becomes a few clicks.
2. Comparison. Prices and terms sit side by side in the same format, instead of arriving as quotes in thirty different layouts that someone has to retype into a table.
3. Trust. Reviews and transaction history accumulate, giving each side something to judge a stranger by. Without reviews the only rational move is to pick the cheapest option; with them, paying a bit more for a safer option becomes a reasonable choice.
4. Payment. The money is held in the middle and released when the work is finished, so neither side has to be the one who goes first. That escrow step is also what gives the product a record of every deal, which is what makes charging a commission possible at all.

All four require both sides to be present at once, which is the condition of a two-sided market — a market that only produces value when suppliers and buyers are assembled together. With ten sellers, buyers find nothing worth choosing and leave; with no buyers, sellers have no reason to keep their listings up.

This is why early marketplaces run straight into the cold start problem. The usual answer is to eat a loss on one side first: waive commission for suppliers for a year, pay early sellers to post inventory, or have the company act as a supplier itself until demand exists. Narrowing the scope to one city or one category does the same work. Used textbooks around a single campus reach critical mass long before used everything nationwide does. Once both sides pass a threshold, the network effect takes over: more sellers bring more buyers, and more buyers bring more sellers again.

The same problem can be solved in other shapes, and because assembling both sides is expensive, the shapes that deal with only one side are usually considered first.

• Buying and reselling. The company purchases inventory and sells it under its own name. Quality and delivery stay under direct control, so the experience is consistent, but the company carries the stock, and how fast the catalog can grow is limited by capital.
• Listing only. Show a directory of suppliers and contact details, and let the transaction happen off the product. It's the easiest to build, but there's no record that a deal occurred, so there's nothing to charge a commission on — and once two parties connect, they skip the product from then on.
• Selling tools to suppliers. Instead of brokering the transaction, sell suppliers the software as a service (SaaS) they use to serve their own customers. Only one side has to be assembled, which makes the start far lighter. The cost is that the product never holds a relationship with the buyer.

Any of the three can turn into a marketplace later. A common sequence is to sell tools to suppliers first, accumulate the supply side, and then attach demand — open a consumer-facing booking screen once 500 shops already run their scheduling on the tool.

"This quarter's goal is supplier count, not transaction volume. Demand is already backed up."

The shortage is on the supply side, so that's where resources should go.

"Raising commission from 8 to 12 percent will push sellers to transact off-platform."

A warning that a higher cut makes both sides route around the product.

"We're not a marketplace — we sell the inventory ourselves."

A different structure of revenue and liability, because the company owns the goods.

Pagedeck isn't a marketplace, but someone once proposed letting users buy and sell the document templates they build. Mia's first question wasn't whether it could be built but whether both sides would fill. She counted how many users currently make templates good enough to sell and whether anyone would pay for them. If one side stays empty, what ships is a screen with no transactions on it.

The most common assumption is that plenty of users means the thing runs on its own. In most products a rising user count pulls the other numbers with it, so the expectation is reasonable. But a marketplace's users split into two groups, and the total means nothing when the ratio is off. Of 50,000 accounts, if 48,000 are sellers, listings pile up unsold and those sellers are gone within two months. So the number to watch here isn't total signups but the share of listings that actually transact — whether 3 or 40 of every 100 items posted find a buyer says more than the 50,000 does.

The arithmetic on commission is the other place teams go wrong. Raising the rate rarely raises revenue by the same proportion, because both sides can always transact outside the product. Once a seller and a buyer know each other, moving the next deal off-platform takes the commission to zero. That's how a quarter ends with the same number of deals happening but less volume recorded inside the product. When setting the rate, the question is whether the commission stays cheaper than the hassle of going around it.