The chicken-and-egg problem that kills two-sided marketplaces is a sequencing failure, not a paradox. Fix the sequence and the problem largely dissolves.
Building a two-sided marketplace startup looks impossible on paper. Supply won't arrive without demand, and demand won't arrive without supply. Founders feel stuck before they've written a line of code. But understanding how to build a marketplace startup that actually works comes down to one insight: don't try to grow both sides at once. Decide which side is the bottleneck, solve for that one first, and let the other follow. The paradox doesn't get cracked. It gets sidestepped.
Every marketplace has a scarce side and a more recruitable side. Getting that diagnosis right early, before you've spent six months running acquisition campaigns in the wrong direction, is most of the actual work.
OpenTable is the clearest working example of what this looks like in practice. When it launched in 1998, restaurants wouldn't join a reservation platform with no diners on it, and diners had no reason to use a platform with no restaurants. So OpenTable didn't pitch restaurants on a marketplace at all. It sold them reservation management software: a standalone product with genuine standalone value. The marketplace came after, once the supply base was real. By the time Priceline acquired OpenTable in 2010 for $2.6 billion, it had more than 20,000 restaurant partners, most of them signed up initially because the software was worth using before a single end-user arrived.
This approach is sometimes called single-side subsidization, but the label makes it sound more complicated than it is. Give the scarce side an offer that works whether or not the other side exists yet. It doesn't have to be free software. Guaranteed minimum payouts work. Exclusive access to a category where suppliers have no comparable alternative works. The only requirement is that the supplier is better off joining than not joining before a single customer ever shows up.
Uber's early driver subsidies in San Francisco followed the same logic. They weren't charity. They were a deliberate, time-limited investment in the scarce side of a market where rider demand could be generated through marketing but driver supply couldn't. The company tracked driver hours, wait times, and rider growth carefully because those numbers told it when the subsidy had done its job and liquidity had become self-sustaining. Most founders subsidize both sides indefinitely and call it user growth. That's not a strategy. It's a delayed reckoning with unit economics that catches up with you when the next funding round doesn't close.
Identifying the scarce side isn't always obvious. In ride-sharing it was drivers. For food delivery entering a new city, it's restaurants. For a B2B software marketplace, it's often the enterprise buyer, because landing one large procurement team willing to vet new vendors is worth a hundred smaller suppliers sitting idle. The question worth asking early: if you could get one side for free today, which would change more?
Constrain the Launch Until Liquidity Is Genuinely Real
In 2008, Airbnb's founders identified the Democratic National Convention in Denver as a structural demand moment: thousands of attendees, not enough hotel inventory. They didn't try to launch a global platform. They targeted one city, one event, a few dozen listings. Supply was personally reachable. The demand was already there and supply was the constraint. They positioned the company at an existing gap rather than trying to manufacture one.
Craigslist started even smaller. Craig Newmark launched a single email list for the San Francisco technology community in 1995 and built supply by posting listings himself in the early weeks. By the time the site went live as a proper website, it had documented users with documented behavior. It expanded city by city over years. The logic applies to categories as much as geography. A B2B marketplace that starts with one vertical and owns it before expanding has a structural advantage over one that tries to serve every category from launch. Founders who go wide immediately because they're afraid of looking small are making an ego decision dressed up as strategy.
There's a related approach worth naming: aggregating supply before the marketplace is fully live. Yelp had user-written reviews of local businesses long before those businesses were active participants on the platform. Restaurants and shops didn't need to opt in for the early flywheel to turn. By the time Yelp approached them about claiming pages or running ads, there was already measurable user intent arriving through search. The supply side was visible before it was willing. Some job boards and freelance platforms used similar mechanics, surfacing public listings before building genuine two-sided engagement. The limits are real: passive supply can't fulfill orders or respond to messages. But aggregation buys time and gives you real traffic data to show suppliers when you do approach them.
The Monetization Mistake That Compounds the Chicken-and-Egg Problem
Plenty of marketplace founders drop the commission entirely in the early stages, reasoning that volume today will justify charging later. Sometimes that works. More often it becomes a trap: you've trained both sides to expect zero fees, your unit economics never improve, and you're running a subsidized intermediary that collapses when the runway ends rather than a marketplace that can sustain itself.
Thumbtack, the home services marketplace, spent years navigating exactly this before tightening its model. It eventually shifted to charging service professionals per lead rather than per completed job. That change aligned incentives and gave the supply side predictable, plannable costs. The point isn't that you should charge from day one. It's that how you charge determines how both sides behave, which determines whether your liquidity is real or just activity on a dashboard. That decision belongs at the center of product design, not in a revenue model slide that comes after the growth projections.
Frankly, most marketplace pitch decks treat monetization as a formality. That's backward. The rake you charge, who bears it, and when you start all determine whether you're building a sustainable marketplace or buying the temporary appearance of one.
The founders who build lasting platforms stay in direct contact with both sides far longer than feels efficient. Not through automated surveys. Direct conversations, often uncomfortable ones, about what each side actually needs from the other. Airbnb's founders flew to New York in the early months to meet hosts personally and photograph listings. That wasn't purely a retention tactic. It told them what friction was invisible in the product and what would make supply stick through the months when the platform was thin. You can't get that from a dashboard.
How to build a marketplace startup reduces to a sequence, not a simultaneous launch. Identify the harder side. Build something that works for them before the other side exists. Keep your geography or category constrained until the liquidity is real, not just visible. Then expand. The chicken-and-egg problem doesn't need a clever solution. It needs a clear-eyed decision about what comes first, and the patience to stay there until that first thing is genuinely working.
Also read: AI accounting automation will cut your bookkeeping bill in half • How to Build an AI Agent for Your Business Without Writing Code • A DeFi Portfolio Strategy Built to Survive the Next Bear Market
