The chicken-and-egg playbook every marketplace founder reuses

Every marketplace founder eventually faces the question: who comes first, supply or demand? The honest answer is "neither, and both — you build them together in a specific shape." There are five shapes that have worked across most successful marketplaces; the right one depends on what you have to start with.

Shape 1: Single-player tool that becomes two-sided

How it works: build a tool that one side of the marketplace already gets value from, even without the other side. Once enough single-player users exist, introduce the other side. Examples: OpenTable started as a restaurant reservation tool (restaurants used it for table management even without consumers). Airbnb's host tool was usable as a "manage my one rental" before there was meaningful guest demand.

Best fit when: you can build something genuinely useful to one side as standalone software. The single-player version has to earn its keep — not be a marketplace pretending to be a tool.

Shape 2: Concentrated geography or niche

How it works: pick one tiny geography (one city, one neighbourhood, one campus) or one tiny niche, and saturate it on both sides before expanding. Examples: most local marketplaces (Uber started in one San Francisco district; DoorDash started on one Stanford campus). Substack first focused on independent newsletters.

Best fit when: the marketplace works locally / contextually, and density matters. A national launch starves both sides; a 5-mile radius launch gets enough liquidity to feel real.

How to execute: define the geography or niche narrowly enough that you can hand-recruit both sides. 30-50 supply, 100-200 demand within the boundary. Don't expand until you've proven the unit economics there.

Shape 3: Subsidise the harder side

How it works: figure out which side is harder to acquire (usually supply, occasionally demand), and pay for it. Could be cash, could be in-kind, could be done-for-you onboarding. Examples: rideshare companies subsidising drivers; food delivery subsidising restaurant onboarding fees; Patreon paying creators upfront in the early days.

Best fit when: you have capital to deploy and the marketplace's unit economics will eventually justify the subsidy. Disastrous when neither is true — you bleed cash forever.

How to know which side is harder: count the natural inbound on each side at zero marketing. Whichever has fewer hands raised is the harder side.

Shape 4: Aggregate existing supply

How it works: scrape, integrate with, or otherwise represent existing supply that already exists elsewhere. Build the demand side; the supply side already has product-market fit somewhere else and you're just channelling it. Examples: Kayak aggregating flight inventory from airline systems; Indeed aggregating job postings; Airbnb's early "we'll list your Craigslist apartment for free" hack.

Best fit when: the supply already exists in some form publicly or via integrations. Be very careful about whether what you're doing is welcomed by the supply (or about to be cease-and-desisted).

Shape 5: Founder-led marketplace

How it works: in the first 90 days, the founder personally handles both sides. They take the supply order, fulfil it manually, deliver to the demand side. The "marketplace" is a Stripe page and a phone number. Examples: Airbnb's early days (founders flew to every host's apartment to take photos); Reddit's early days (founders posted fake content from sock-puppet accounts to seed activity); most enterprise marketplaces start with the founder personally brokering each deal.

Best fit when: you have no liquidity and no capital but you have time. You're using your own time as the subsidy. Doesn't scale past 100 transactions but it tells you whether the marketplace shape works at all.

Picking your shape

Answer these three questions:

1. Is there a single-player version of the product that's useful on its own? Yes → Shape 1.
2. Do you have $50k+ to deploy on subsidies and you've identified the harder side? Yes → Shape 3.
3. Does meaningful supply already exist somewhere else (Craigslist, public listings, APIs)? Yes → Shape 4.

If no to all three:

• Have a tight geography / niche? Yes → Shape 2.
• No? Shape 5. Be the marketplace yourself for 90 days and use the learning to pick a different shape after.

What never works

• "Launch a two-sided marketplace and run paid ads to both sides simultaneously." Bleeds cash on both sides, neither side ever sees enough of the other to retain.
• "Build the world's best marketplace UX and the liquidity will come." Design doesn't beat liquidity. The ugliest marketplace with enough supply beats the prettiest empty one.
• "Wait for organic growth to hit critical mass." Marketplaces don't grow organically from zero; the founder has to make the first 100-1,000 transactions happen by hand.

What to do today

If you're at zero: pick Shape 5. Start brokering. The point is to learn whether the trade is real, not to scale.

If you've done 50-100 transactions and the unit economics survive: pick Shape 1, 2, 3, or 4 based on what you've learned. Don't pick two shapes simultaneously — you'll dilute focus on both.