The Chicken-and-Egg Problem in Marketplaces

By Brian Tr · Updated Oct 6, 2026

Every marketplace founder meets the same wall on day one. Buyers won't show up to a place with nothing to buy. Sellers won't list in a place with nobody to buy from. Whichever side you recruit first looks at the empty other side, shrugs, and leaves.

That's the chicken-and-egg problem. Most explanations stop at the one-liner and a list of tactics. This article does something different: it defines the problem precisely, explains why it's brutal for some marketplaces and mild for others, introduces the named frameworks founders actually use, and sorts the solutions into families so you can pick the right one for your market. The detailed how-to lives in the spoke articles linked throughout, and this page is the hub that ties them together.

If you want the short overview first, the sections in two-sided markets and the marketplace business model cover the basics. Read on for the version you'd want before you commit a budget to it.

What the problem is, precisely

The popular phrasing makes it sound like a chicken-and-egg riddle with no answer. In economic terms it's a coordination failure. Each side's willingness to join depends on how many people from the other side have already joined. If neither side moves first, both stay out, even though both would be better off in a populated marketplace.

Economists call the dependency a cross-side (or indirect) network effect. The value a seller gets from the marketplace depends on how many buyers are there, and vice versa. It's different from a direct network effect like a phone network, where the value comes from people on the same side. We cover the general idea in network effects. The marketplace twist is that you need two groups at once, and each group is the product for the other.

Two consequences follow, and they explain why the problem feels so unforgiving:

• An empty marketplace isn't a small marketplace. It's a broken product. A note-taking app with ten users still works for each of them. A marketplace with ten sellers and no buyers delivers no value to anyone.
• The problem repeats. You don't solve it once. You solve it for a first city, then a second, then a new category. Each new pocket of the market starts cold again.

Where the term comes from

The phrase entered the economics literature through Bernard Caillaud and Bruno Jullien's paper "Chicken & Egg: Competition among Intermediation Service Providers", published in the RAND Journal of Economics in 2003 (volume 34, issue 2, pages 309 to 328). The paper models how intermediaries compete when indirect network externalities are present, and it asks which market structures emerge in equilibrium.

The same year, Jean-Charles Rochet and Jean Tirole published "Platform Competition in Two-Sided Markets", which examines markets where a platform has to attract several customer segments at once and works out how it should price across them. Together these two papers are the standard academic roots of two-sided market theory.

A practitioner-friendly follow-up is David Evans and Richard Schmalensee's 2010 paper "Failure to Launch: Critical Mass in Platform Businesses" in the Review of Network Economics. It argues that new two-sided platforms face an initial critical mass hurdle, and that, under fairly weak assumptions, a platform has to get a sufficient number of members of both sides on board to launch successfully. In plain terms: you can't win by having one side. You need enough of both.

Public companies say the same thing in their risk factors. Uber's 10-K acknowledges that if it can't attract or maintain a critical mass of drivers, consumers, merchants, shippers, and carriers, its platform will become less appealing to users. It's a risk the company still lists years after launch, which tells you the problem never fully goes away. It just changes shape.

Why some marketplaces have it worse

Not every marketplace faces the same cold start. Five variables decide how steep your climb is. Think of them as dials, and note where yours are set before you plan a launch.

Here's how those dials play out in practice.

Local versus global. A global catalog marketplace can aggregate demand and supply from everywhere, so a small base still produces some matches. A local one can't. A ride-hailing app with ten drivers spread across a metro area isn't a ride-hailing app to anyone, because each rider only sees the drivers nearby. Local markets need density in a small area, not totals. That's why launches there tend to go neighborhood by neighborhood.

Real-time versus asynchronous. If a buyer can post a request and wait a day, a thin supply side can still serve them. If the buyer wants a car in four minutes, thin supply fails instantly. The less patience your product allows, the more supply you need before the first buyer has a good experience.

Frequency. High-frequency markets give you many chances to fix a bad first match. Low-frequency ones (a house sale, a once-a-year specialist) give you one shot, and a bad first transaction may mean you never see that buyer again.

Effort on the supply side. Listing a spare item takes a minute. Becoming a driver, a host, or a service provider takes real commitment. The more effort supply must invest, the more convincing you have to be about the demand waiting for them. That is why supply is usually the side people call hard, and it's the subject of the spoke on marketplace supply acquisition.

Standalone value. Some marketplaces can give one side a reason to show up that doesn't depend on the other side. A tool that helps a seller manage inventory, or a directory that's useful as a plain list, gives you a way to start before the network exists. If your product is worthless until both sides are present, you're at the hard end.

The type of marketplace changes the profile too. Our guides to B2B vs B2C marketplaces and horizontal vs vertical marketplaces show how buyer behavior and category breadth move these dials. A vertical B2B marketplace with a small set of repeat buyers faces a very different cold start from a horizontal consumer one.

The vocabulary founders use: Andrew Chen's framework

The most widely used practitioner framework comes from Andrew Chen in his book The Cold Start Problem. The book's own site lays out its structure as a progression of stages: the cold start, the tipping point, escape velocity, the ceiling, and the moat. It also has chapters titled "The atomic network" and "The hard side." Here's what each idea means and how it applies to a marketplace.

The atomic network

Chen's term for the starting unit is the atomic network. In an excerpt published in Lenny's Newsletter, he describes it as the smallest network that can stand on its own. The point of the concept is that you don't try to launch a whole market. You build one tiny network that works, then build another.

The examples in that excerpt show how small "small" can be. Slack's atomic network was a single team of fewer than ten people inside one company. Credit cards were the opposite case: they needed an entire city's worth of consumers and merchants before the network made sense. Uber, per the excerpt, started hyperlocally, around a specific time and place (5 pm at one train station), rather than citywide.

For a marketplace the lesson is that "enough supply" is local and specific, and usually smaller than founders expect. You're not asking "how many sellers do I need?" You're asking "what's the smallest slice of buyers and sellers, in one category, in one place, where a buyer reliably finds a good match?"

The hard side

Chen's second idea is the hard side: in a network with several groups, one group is harder to attract and does more of the work. On his own blog, he argues that a minority of users usually create disproportionate value and, as a result, hold disproportionate power. Winning that group is the priority, and the way to win them is to solve a real problem for them.

On that same page he describes the typical marketplace sequence as supply, demand, then supply, supply, supply. In other words, start with the hard side, bring in the easy side, and then keep coming back to the hard side, because it's the one that keeps the whole thing standing. Which side is hard in your market is an empirical question, and getting it wrong wastes months. The spoke on supply-first vs demand-first walks through how to decide.

The stages after the cold start

The later stages matter because they show what "solved" looks like. Chen's progression runs from the cold start to a tipping point, where adding atomic networks becomes repeatable. It then moves through escape velocity, a ceiling, and finally a moat. The later stages deal with growth limits and defending the network once it exists. The practical use is sequencing: don't spend ceiling-stage money on a cold-start problem. If you're still trying to make one atomic network work, a national brand campaign is premature.

The families of solutions

There are dozens of tactics, but they cluster into a handful of families. Pick the family first, then the tactic. Each family trades speed against cost and against how much quality you control.

We don't re-explain each tactic here. The two-sided markets page covers the basic list. What matters is matching a family to your dials from earlier.

• Local, real-time, high effort for supply? Narrowing and subsidizing do most of the work, because density is everything and supply needs a reason to commit.
• Global, asynchronous, low effort? Borrowed audiences and standalone tools tend to carry you further, since thin supply still produces matches.
• High-stakes, low-frequency? Operating it yourself is common, because the first transaction has to be good. That's the territory of the managed marketplace.

The next layer down is in the spokes of this section. Supply acquisition covers how to find and sign the first sellers. Seller onboarding covers what happens in the first days after they sign, because a seller who joins and then lists nothing hasn't solved anything. Supply quality covers how to keep the early listings good enough that buyers come back. And supply-first vs demand-first covers the sequencing question that comes before all of those.

If you're building a very early product, pair this with the thinking in minimum viable product. The goal at the start isn't a full marketplace. It's the smallest version that proves one atomic network can sustain itself.

Common ways founders get this wrong

• Launching broad. Opening a whole country or every category spreads a small supply base too thin for any buyer to find a match. Narrow beats wide at the start.
• Counting sign-ups instead of matches. A thousand registered sellers who never list mean you still have an empty marketplace. Measure what a buyer experiences, not what your dashboard counts.
• Solving the easy side. Buyers are often cheaper to acquire, so founders collect them first. If they find nothing, they churn, and you've paid to teach them that your product doesn't work.
• Subsidizing with no exit. Incentives that never taper train participants to expect them. Decide up front what replaces the subsidy: organic matches, repeat behavior, or fee revenue.
• Treating it as a one-time hurdle. Every new city or category restarts it. Plan the expansion unit, not just the launch.

How you know you've solved it

There's no single switch. Instead you see a cluster of behaviors that show the market is sustaining itself. The measurement frameworks come later in this collection, so here are the signals described in words:

1. Buyers find a good match quickly, without help. A new buyer arrives, searches, and transacts without you personally steering them.
2. Sellers get enough business to stay. Supply doesn't quietly drift away after the first few weeks because nothing came of it.
3. Both sides come back. Repeat use shows up on the buyer and the seller side, not just new sign-ups.
4. Growth no longer needs a manual push for each unit. You can open the next neighborhood or category with less hand-holding than the last one.
5. Demand arrives that you didn't buy. Referrals, search, and word of mouth begin to bring in participants who cost you little.
6. The subsidy can shrink without the market collapsing. If removing incentives empties the marketplace, you haven't escaped yet.

None of these is a guarantee on its own. Together, they're what Chen's framework would call a network that has passed from cold start toward its tipping point.

Key Facts: The Marketplace Chicken-and-Egg Problem

• The phrase is the title of a 2003 paper by Caillaud and Jullien, "Chicken & Egg: Competition among Intermediation Service Providers", RAND Journal of Economics 34(2), pages 309 to 328.
• Rochet and Tirole's 2003 paper, "Platform Competition in Two-Sided Markets", studies platforms that must attract multiple customer segments at once.
• Evans and Schmalensee argue that new two-sided platforms face an initial critical mass hurdle and need a sufficient number of members of both sides to launch.
• Andrew Chen defines the atomic network as the smallest network that can stand on its own (Lenny's Newsletter excerpt from The Cold Start Problem).
• Chen's book organizes the journey into five stages: cold start, tipping point, escape velocity, the ceiling, and the moat.
• Uber's 10-K lists failure to maintain a critical mass of drivers, consumers, merchants, shippers, and carriers as a risk to the platform.

Related Reading

• Supply First vs Demand First
• Marketplace Supply Acquisition
• Marketplace Seller Onboarding
• Marketplace Supply Quality
• What Is a Marketplace Business?
• Two-Sided Markets
• Marketplace Business Model
• Network Effects