How to get your first users for a marketplace

Pick one side and one small market, then go and get that side by hand. Not both sides. Not a country. One side, one city or one narrow category, recruited personally until the room looks full.

Every other page in this cluster is about finding the people who want what you built. A marketplace differs in kind: what you built does not exist until both sides show up, and neither will show up for an empty room. That is not a copy problem - it is a structural condition, and the only reliable way through it is to manufacture one side manually, in a space small enough that manual is enough.

Growth is not linear and neither is failure

A SaaS product with twenty users is a small SaaS product. A marketplace with twenty users a side, spread across a country, is nothing at all - every search returns empty, both sides leave, and that churn is permanent, because a bad first impression of an empty marketplace is very hard to reverse. So early growth is worthless unless it is concentrated: ten thousand users at low density is a worse position than four hundred at high density, and it costs far more to reach.

You must choose the constrained side, deliberately

One side is always harder to acquire than the other, and the whole plan hangs on which. Usually it is supply, because supply requires effort and demand only requires a need - but not always. Where the good providers are already busy and the buyers are desperate, demand is abundant and supply is the bottleneck; where anyone can provide but nobody is looking, the bottleneck is demand. Work it out before you spend anything: the abundant side can generally be recruited later, at low cost, once the scarce side exists.

Manual is the strategy, not a shortcut

Recruiting each provider personally, making matches yourself, fulfilling orders behind the scenes - doing the work the software will eventually do - is how most functioning marketplaces began. It is not scaffolding to be ashamed of. It is the only way to learn what a good match looks like before you write code that makes matches badly at scale.

Single-player mode makes the impossible pitch possible

Asking a busy professional to list on an empty platform is asking them to work for a promise. Giving them something useful with zero buyers - a scheduling tool, a profile page they can send to existing clients, inventory tracking to replace a spreadsheet - changes the pitch from a bet to a benefit. They stay for the tool, and they are there when demand arrives.

Trust is a feature, not a page

Both sides are being asked to transact with a stranger through an intermediary nobody has heard of. Whatever reduces that risk - verified identity, escrowed payment, a guarantee you personally underwrite while you are small - does more for conversion than any channel, because the objection is not "I had not heard of you", it is "I do not know if this is safe".

Both sides of a functioning marketplace are already transacting somewhere, badly. Find the bad version and you have found both populations at once.

• In group chats, community boards and classified sections where the transaction currently happens with no structure. A busy local group where people ask "does anyone know a…" every week is a marketplace with no software, and everyone in it has already demonstrated the behaviour you need.
• On the platform they resent. Providers on a big incumbent complain publicly about fees, rules and lead quality. Those complaints are a recruiting list, and their grievance is your positioning.
• In trade bodies, professional groups and industry meetups. Slow, analogue, and by far the most effective way to recruit a supply side that takes itself seriously.
• Physically, where the work happens. Markets, workshops, depots, campuses. If your marketplace is geographic, the fastest supply recruitment available is usually a day of walking around with a phone.

Ranked for the first month of a marketplace with no liquidity anywhere. Note that most of the paid and scalable options sit at the bottom - not because they never work, but because they are actively harmful before density exists.

• Do not launch both sides at once. The single most common marketplace failure. You get two thin populations that never meet, and both conclude the product is dead.
• Do not fake inventory, reviews or activity. It creates demand you cannot fulfil, burns the goodwill of your first real buyers, and in many categories it is fraud rather than growth hacking.
• Do not go wide geographically because the software allows it. The software allowing it is exactly why founders do it, and it is why so many marketplaces are permanently empty everywhere.
• Do not charge a take rate before there is liquidity. You are taxing a service that does not work yet. Get transactions happening, then introduce the fee and be direct about why.
• Do not build the matching algorithm first. Match by hand until you can predict which pairings work. Then encode what you learned instead of guessing.
• Do not spend on the abundant side. Acquiring more of what you already have too much of feels like progress and moves nothing.

Marketplaces produce more flattering vanity metrics than any other product type, because signups and listings both rise while the thing quietly fails to work.

• Liquidity, measured within one pocket. What share of listings transact within a sensible window, and what share of searches end in a match. Measure it in your one city or one category - the platform-wide average is designed to mislead you.
• Searches that return nothing. The most useful early diagnostic you have. Every empty result is a specific, named gap in supply, and the list of them is your recruitment plan.
• Time from listing to first response. This decides whether the supply side comes back. Measure it in hours, not days, and watch it rather than the average.
• Repeat rate on each side separately. A buyer who returns and a provider who relists are the only real evidence the exchange was worth making. Aggregate retention hides which side is leaking.
• Ignore total signups, listing counts and gross volume. All three go up while a marketplace is dying.

Marketplaces are the slowest type here to show signs of life and the hardest to fake progress on - a mercy in disguise, because the numbers refuse to flatter you. Expect the first working pocket to take months of unglamorous, largely offline work, and expect expansion to feel like starting again rather than scaling, because it is. The compensation is that a marketplace which does reach density is far harder to displace than anything else in this cluster: a competitor has to solve the same physics problem against a market you have already filled.

The pillar covers showing up in the communities where both sides gather, and the channel picker helps if you are still unsure which side is constrained.

The channel picker asks about your product, your buyer and how the transaction happens today, then names the two or three channels worth your first month. For a marketplace it mostly resolves one question - which side is genuinely scarce - and that answer reorders everything else.

Or get the 30-day post-build playbook Send it