The Anatomy of a Two-Sided Marketplace Moat

In October 2007, two designers in San Francisco couldn't make rent. A design conference had sold out every hotel room in the city, so Brian Chesky and Joe Gebbia bought three air mattresses, set them up in their apartment, and charged $80 a night for what they called "Airbed and Breakfast." Three guests showed up. It worked well enough to repeat — and the repeating is the part that matters, because anyone can rent out an air mattress once.

The harder question is what keeps an idea like that, years and several pivots later, from being trivially copied once it's obviously working. A two-sided marketplace doesn't manufacture anything and doesn't own the inventory moving through it. What it sells is access to liquidity — enough buyers that sellers show up, and enough sellers that buyers bother to look. That's an unusual asset to build a moat around, because liquidity is largely invisible until you try to compete without it.

This is the structural question worth asking of any two-sided marketplace: what would it take to displace them, once they've actually solved the cold-start problem? eBay, Airbnb, and Etsy each answered that question differently. I'll get to a fourth question too — what eventually starts to crack that liquidity once it's built — because the same forces that make these moats hard to dig also tell you where they erode.

Solving the Chicken-and-Egg Problem

Every two-sided marketplace starts empty, and an empty marketplace is worthless to both sides at once. Sellers won't list on a platform with no buyers. Buyers won't visit a platform with no listings. Somebody has to move first, and the company that figures out how to manufacture that first move — without it looking manufactured — gets a head start that, if liquidity actually takes hold, becomes very hard to take away.

eBay: First and Alone

eBay launched in September 1995 as a single auction category on a personal website, with no real online competitor of any meaningful scale. There was no incumbent to dislodge, no entrenched seller base loyal to somebody else's platform. The chicken-and-egg problem mostly solved itself through timing: eBay was the only auction-format marketplace online when collectors, then casual sellers, started showing up in volume. By the time well-funded competitors arrived, eBay already had the listings, the feedback-score reputation system, and the seller habits that make multi-homing — using more than one platform at once — more trouble than it's worth for the average part-time seller.

Etsy: Liquidity Through Curation, Not Scale

Etsy, founded in 2005, took the opposite approach. Rather than compete with eBay or Amazon on breadth, it built liquidity inside a narrower category — handmade and vintage goods — where buyers specifically wanted a curated experience the general marketplaces didn't offer. The narrower scope meant a smaller absolute pool of buyers and sellers, but a denser one: a buyer searching for hand-thrown ceramics is far more likely to find a relevant seller there than on a marketplace fifty times its size but built for general merchandise. Niche-first liquidity is a slower path to scale. It's also harder for a horizontal competitor to replicate quickly, since matching it means building real category depth, not adding a checkbox filter.

Airbnb: Liquidity That Doesn't Travel

Airbnb's problem was structurally different from eBay's or Etsy's, because travel liquidity is local. A surplus of hosts in Austin does nothing for a guest searching Lisbon. Airbnb had to win individual cities — New York, San Francisco, Paris — more or less one at a time, seeding host supply through founder-led outreach before any of it compounded. That city-by-city slog is exactly why the resulting moat holds up reasonably well: a new entrant doesn't just need national scale, it needs to win the same dense, local liquidity contest in every market it wants to compete in, against a company that already has the host relationships and review history there.

The Take-Rate Test

A marketplace's take rate — the share of transaction value it keeps as revenue — is one of the more honest signals of moat strength available, because it functions as a price the platform charges both sides for access to each other. A marketplace with weak liquidity can't raise its take rate without sellers leaving for somewhere cheaper. A marketplace with genuine liquidity can, within limits, because the alternative for sellers is having no buyers at all. It's also why two-sided marketplaces sit among the more extreme cases of asset-light businesses: take-rate revenue converts to free cash flow with little capital reinvestment required, which is exactly why getting the take rate right matters so much to long-term economics.

eBay's fiscal 2024 10-K reported GMV of roughly $74.7 billion against net revenue of $10.3 billion — a take rate near 13.8%, with 134 million active buyers and 2.3 billion live listings at year-end. That take rate has drifted down for years, which is itself informative: it's the visible cost of multi-homing pressure from Amazon, Facebook Marketplace, and category-specific competitors chipping at eBay's edges without ever mounting a full frontal assault.

Etsy's economics look different. Fourth-quarter and full-year 2024 results, reported in February 2025, showed consolidated gross merchandise sales of $12.6 billion — down 4% year over year — against record revenue of $2.8 billion, implying a consolidated take rate near 22%, well above eBay's. Sellers tolerate that premium because Etsy's curated buyer intent is worth paying for. But the GMS decline, paired with active sellers falling roughly 20% to 5.6 million from 7 million the year before, is the kind of divergence — rising take rate, shrinking marketplace — worth sitting with rather than explaining away.

Airbnb sits in between. Full-year 2024 gross booking value reached $81.8 billion against revenue of $11.1 billion, a blended take rate near 13.6%, built from a roughly 3% host fee and a guest service fee that runs 14% to 16% depending on the listing. Unlike eBay's drifting-down take rate, Airbnb's has held closer to steady — modest evidence that its liquidity, so far, hasn't faced the same erosion eBay has.

Why Single-Homing Is the Real Moat Source

The mechanism underneath all of this is single-homing versus multi-homing. A marketplace moat is strongest when both sides only use one platform — when a seller lists exclusively on eBay, or a host only takes bookings through Airbnb. It weakens the moment either side starts splitting volume across several platforms simultaneously, because the marketplace stops being the only path to the buyer or seller on the other end. It becomes one path among several. So when a seller can reach the same buyer without paying a particular platform's toll, that platform's pricing power is exactly as wide as the inconvenience of going around it.

This is where network effects and switching costs diverge in a marketplace context, and the distinction matters more than most investors give it credit for. A strong network effect gets you to liquidity. It doesn't, by itself, keep either side from also building liquidity somewhere else. Real lock-in requires switching costs layered on top — accumulated reviews and reputation a seller doesn't want to rebuild from zero, a payment and logistics system a buyer doesn't want to re-enter, a search-ranking history that took years to earn. Marketplaces with thin switching costs on top of real network effects are vulnerable in a specific, identifiable way: liquidity that's easy to reach is also liquidity that's easy to leave.

Where the Moat Cracks

Two forces erode an otherwise functioning two-sided marketplace, and they tend to show up together rather than separately.

The first is disintermediation — buyers and sellers meeting on the platform, then transacting off it to avoid the fee. It's most visible in categories with low repeat-purchase friction and high per-transaction value, where the fee saved by going around the platform is worth the inconvenience. eBay has fought this for two decades through policies restricting off-platform contact information in listings; the fact that the policy still needs enforcing tells you the incentive to circumvent never goes away, it just gets managed.

The second, more structural, force is vertical integration and multi-homing by the highest-value supply. Professional Airbnb hosts — property managers running a portfolio of units rather than an individual renting a spare room — increasingly list the same inventory across Airbnb, Vrbo, Booking.com, and a direct booking site simultaneously, using channel-management software built for exactly that purpose. No single platform fee. No single platform's discoverability either, but that's a trade professional operators increasingly make once their reputation no longer depends on any one site's review history. The same dynamic shows up on Etsy, where a meaningful share of established sellers maintain a Shopify storefront alongside their Etsy shop specifically to reduce dependence on Etsy's fee structure and algorithm changes.

We covered the broader pattern in moat erosion warning signs: margin or take-rate pressure that doesn't reverse, sustained across multiple periods, is the late-arriving confirmation of a competitive position that already shifted. eBay's gradually declining take rate over the past decade reads less like a pricing choice and more like exactly that signal, observed in slow motion.

Wide, Narrow, or No Moat

Grading these three against MoatScope's classification produces three different answers, and the differences are instructive.

eBay looks narrow-moat today rather than wide. The liquidity is real and the switching costs tied to an established seller's account history are genuine, but the take-rate trend and the steady share loss to Amazon and Facebook Marketplace in casual-resale categories both point toward eroding, not durable, economics. The marketplace still works. It's no longer obviously the only place that works.

Etsy is the harder call, and I'll admit I'm less certain about it than the other two. The curated-niche argument for a real moat — buyers who specifically want handmade goods have nowhere else with comparable density — sits uneasily next to four straight quarters of declining sellers and buyers. A 20% drop in active sellers in a single year isn't noise. Call it narrow and weakening rather than wide; the niche liquidity is still genuinely valuable to the buyers who want it, but the platform isn't currently proving it can hold the supply side.

Airbnb comes closest to wide-moat among the three. Local liquidity density across thousands of cities is difficult to replicate quickly, the trust infrastructure built from years of reviews carries real switching costs for both guests and hosts, and the take rate has held steadier than eBay's. The professional-host multi-homing trend is the genuine risk to watch — it's the same erosion mechanism, just in earlier innings.

Key Takeaways

• A two-sided marketplace's moat comes from liquidity, not inventory or technology — and liquidity is hardest to value precisely because it's invisible until a competitor tries to build it from zero.
• Take rate is a genuine pricing-power signal. eBay's 13.8% (FY2024) has drifted down for years; Etsy's roughly 22% has held up even as its marketplace shrinks; Airbnb's 13.6% has stayed comparatively steady — three different trajectories worth more than any one snapshot.
• Single-homing, not the network effect alone, is what actually locks in liquidity. The moment either side can reach the other without the platform, the platform's pricing power narrows to the cost of the workaround.
• Disintermediation and multi-homing by the highest-value supply — professional hosts, established sellers building owned channels — are the clearest signals a marketplace moat is narrowing, and they show up in seller-retention and take-rate data well before they show up in headline growth numbers.