Nobody joins an empty marketplace. So do not launch one.
The cold start problem is the reason most marketplaces die, and the usual answers are recruit harder or spend more. There is a third option in any category where the operators are already on a public register, which is more categories than people expect.
The usual answers
Three standard approaches and where each one runs out
- Recruit supply first
- Months of manual sales to get operators onto a platform with no buyers on it. The pitch is a promise, so the conversion rate is what you would expect.
- Subsidise both sides
- Works, and it is how the famous ones did it. It also needs capital on a scale that rules it out for almost everyone, and it does not survive the subsidy ending.
- Launch narrow, one city
- Sound advice that quietly caps you. One city produces one city of search demand, which is rarely enough traffic to prove anything either way.
All three treat empty as a starting condition to be escaped. Seeding treats it as a condition to be avoided.
The method
Five steps, in the order the work happens
- 01
Seed supply from the public record
Every operator in the category already exists in a licence register, a federal facility list or a county registration file. Those records are public. A marketplace can therefore launch with the supply side already populated, which removes the cold start problem rather than solving it. This is the entire trick and everything below depends on it.
- 02
Make an unclaimed listing genuinely useful
A seeded listing has to be worth landing on before anyone claims it. Real address, verified status, coordinates, category, whatever the source publishes. If a cold listing is a stub, buyers bounce and operators have no reason to care that it exists.
- 03
Let buyers arrive through the listings
Thousands of populated pages produce search demand on their own, which is the substitute for a marketing budget. Buyers find the listing, not the homepage. This is why a marketplace and a programmatic directory are the same build for the first several months.
- 04
Convert listings into claimed accounts
The claim flow is where a directory becomes a marketplace. Verification against the domain on the operator's own site, single-use tokens with an expiry, and an admin queue gating what may overwrite a public record. An operator claims because there is already traffic on the page, which is the inversion that makes it work.
- 05
Monetise the claimed side only
Placement tiers, featured slots, lead capture and routing. Charging before there is demonstrable traffic on a listing is how trust goes. Once an operator can see views, calls and leads on their own dashboard, the upgrade conversation is arithmetic rather than persuasion.
The inversion
Operators normally join a marketplace hoping for traffic. Here they claim a listing that already has it.
That reverses the hardest conversation in building a two-sided platform. You are not asking anyone to take a bet on a new site, you are showing them a page about their own business that people are already landing on.
Built this way
37,411 locations, live on day one
A car rental platform covering 37,411 locations with live geo search. A network of 22,800 interstate exits and 19,000 towing companies with proximity search. A national directory of 9,585 treatment facilities with a claim portal and operator dashboards.
None of the three recruited a single operator before launch.
FAQ
Marketplace build questions
How do you build a marketplace website without the cold start problem?
By not starting cold. If every operator in the category is already listed in a public register, the supply side can be seeded from that record set before launch. You are then running a claim and conversion problem, which is tractable, instead of a two-sided recruitment problem, which mostly is not.
Is seeded supply legitimate?
The records are public and the listings are accurate, so yes, provided the site is honest about what a listing is. An unclaimed listing should say so plainly and give the operator an obvious route to claim it. Presenting a seeded listing as though the operator opted in is where this goes wrong.
What if my category has no public register?
Then this method does not apply and you are back to recruiting supply by hand. It is worth checking properly before assuming either way. Plenty of categories people think are unregulated turn out to have county-level registrations or an accreditation body behind them.
How is this different from building a directory?
For the first several months it is not, and that is the point. A marketplace is a directory plus a claim flow plus a transaction or lead layer. Building it in that order means the traffic exists before the marketplace features have to justify themselves.
What does it cost to build a marketplace website?
Core at $12,000 covers up to 5,000 pages with geo search and a claim flow, which is a working marketplace. Full at $22,000 adds multi-source data, operator dashboards, lead capture and monetization tiers. Free marketplace website templates exist and will get you a demo, not a supply side.
How long before it is worth anything?
The build is four to ten weeks. Search traffic on generated pages takes months regardless of who builds it, and anyone telling you otherwise is selling something. What you have on day one is a complete, accurate, useful supply side, which is the asset that would otherwise have taken a year of sales calls.
Next step
Tell me the category and I will check the register
Free either way. You get the record count, how clean the source is, and a straight answer on whether the supply side can be seeded at all. If it cannot, this method does not apply and I will tell you that rather than sell you a build.