Go-to-Market · 6 MIN
Marketplace launch GTM: solving the chicken-and-egg problem
How to launch a two-sided marketplace: choose a narrow starting market, seed the harder side, make early transactions work, and measure liquidity before scaling.
A marketplace needs buyers to attract sellers and sellers to attract buyers. The practical way out is to narrow the starting market until you can serve both sides by hand, seed the harder side first, make the first transactions work even if it is manual, and measure liquidity before widening. This guide gives the sequence.
- Start with a narrow slice, one category in one city or niche, where you can serve both sides personally.
- Seed the harder side first, which is often the supply that is scarce or costly to find.
- Manual work in the early days is fine, and it teaches what to automate later.
- Measure liquidity, meaning how often a listing or request leads to a completed transaction, before scaling.
- Payments, trust, and dispute handling need design before launch, not after.
- Nactore is an AI-native technology and growth company that builds marketplaces and the go-to-market assets around them, scoped to each client.
What is the chicken-and-egg problem in a marketplace?
Buyers will not come to a site with no sellers, and sellers will not list where there are no buyers. Both sides wait for the other, so neither moves. The problem is worst for horizontal marketplaces with a broad promise, since they must fill every category before any feels useful.
The fix is to change what you are asking each side to believe. Instead of "come to a large network," offer "get this one specific thing done here."
How do you pick a narrow starting market?
Narrow along the dimensions that make matching easy. A marketplace works once buyers find what they need quickly, and that happens faster in a small, dense market.
| Dimension | Narrow choice | Why it helps |
|---|---|---|
| Geography | One city or district | Local demand and supply can meet in person |
| Category | One service or product type | Easier to recruit specialists and set quality |
| Customer type | One buyer segment | Clear messaging and pricing |
| Time | One season or event | Concentrates demand |
Choose a slice where you already have an advantage, such as a community you belong to, a network of suppliers you know, or a pain you have felt yourself.
Which side should you seed first?
Seed the side that is harder to get and that the other side cares about most. In many marketplaces this is supply, for example skilled providers, inventory, or listings. In others, demand is the scarce side, such as a niche buyer group.
Ways to seed supply.
- Recruit personally. Reach out to the first 20 to 50 providers one at a time and explain what they get.
- Bring existing supply online. Help providers list what they already offer, doing the data entry yourself if needed.
- Offer a clear benefit. A tool, a lead stream, or payment convenience that works even before the buyers arrive.
- Set a quality bar. Early listings define what the site feels like.
Ways to seed demand.
- Go where buyers already gather. Communities, associations, and events in the niche.
- Use search content. Guides that answer buyer questions bring in people who are already looking. See how to get mentioned in ChatGPT for the assistant side.
- Make the first search results good. A buyer who finds a relevant match in the first visit comes back.
Can you launch with manual work behind the scenes?
Yes, and you should. Manual matching is how you learn what the product must do.
- Concierge matching. Staff read each request and introduce a fitting provider by message or call.
- Spreadsheets and simple forms. Collect listings before building a complex catalog.
- Manual payouts at first. Move to automated payments once volume justifies the build, and confirm the rules for holding funds with your payment provider and a local advisor.
- A written log of every failure. Missed matches, disputes, and confusion become your roadmap.
Build the parts that are painful and repeated. Leave the rest manual until the pattern is clear. This is the same discipline we describe in how to scope an AI pilot, applied to a marketplace.
How do you build trust between strangers?
A marketplace sells trust between strangers more than it sells listings. Plan these before launch.
- Identity and verification. Decide what you check for each side and state it.
- Clear listings. Required fields, photos, prices, and terms so buyers compare like with like.
- Reviews and ratings. Collect them after each transaction, and prevent fake ones.
- Payments and disputes. Define who holds money, when providers are paid, what happens in a dispute, and who decides.
- Policies in plain language. Cancellation, refunds, and conduct rules, reviewed by a local lawyer for each market you operate in.
Payment, tax, and consumer protection rules differ by country, so confirm them with local counsel before launch.
What should you measure before you scale?
The key measure is liquidity, which is how reliably supply and demand meet. Track a short set of numbers, defined by you and measured the same way every week.
- Search or request success rate. How often a buyer finds something relevant.
- Response time. How quickly providers answer a request.
- Fill rate. The share of requests that become completed transactions.
- Repeat rate. Buyers and providers who come back.
- Concentration. Whether a few providers account for nearly all sales.
If fill rate is weak, adding more of the easier side will not help. Fix the match first.
For more on how weekly go-to-market experiments are run in practice, see GTM Engine, a weekly newsletter on go-to-market.
Run the marketplace like a service for the first 50 transactions. Introducing each pair by hand tells you exactly what the software needs to do and which assumptions were wrong.
When do you expand?
Expand when the first slice works without heroics. Signs include steady repeat usage, providers earning enough to stay, and requests that complete without your intervention. Then widen one dimension at a time, such as a neighboring city or adjacent category, and measure again.
Frequently asked questions
Should we build the full platform before launching?
No. Build the minimum that lets a buyer and a provider complete a transaction safely, then add features as the manual work shows what is needed.
How many listings do we need before inviting buyers?
Enough that a typical buyer in your slice finds a relevant option on the first visit. That number depends on the niche, so define "relevant" with your first buyers.
Should we charge a commission right away?
Not always. Some marketplaces start free to build supply and add fees once matching works. Choose based on your costs and the alternatives providers already pay for.
What if one side leaves the platform to transact directly?
It happens in many marketplaces. Give both sides ongoing reasons to stay, such as payment protection, reviews, tools, and repeat-booking features, and check your terms are clear.
Start small, match by hand, then scale
Pick a slice, seed the hard side, make early transactions work, and track liquidity. A marketplace grows from a few dependable matches, not a big launch.
Want this built for your team? Book a free 30-minute call.
Want to apply this to your business?
Book a free 30-minute call. We will tell you what we would do first.