Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / Why Marketplaces Need Both Sides to Win
Transcript
- Lucas: You know that feeling when you open a food delivery app and there's only three restaurants within delivery range, and two of them are pizza places you don't want? Luna: Yeah, that's basically the whole app failing in that moment for you as a user. Lucas: Right. And that failure — we usually call it the chicken and egg problem. But I think that phrase actually obscures something more specific. It's not that the marketplace lacks participants. It's that it has participants on only one side. Luna: One-sided liquidity. I've heard you talk about this before — the idea that you can have plenty of supply but zero demand, or demand but no supply, and the marketplace still feels dead. Lucas: Exactly. And today I want to look at a case where that one-sided liquidity was the explicit reason a well-funded marketplace shut down. I'm talking about Google Shopping Express. Luna: Oh, that's a good one. Google's same-day delivery service, launched around 2013, I think. They had massive warehouses, partnerships with big retailers like Target and Walmart. Lucas: They had the supply side incredibly well built. You could order from dozens of major stores, thousands of SKUs. The problem was that in most neighborhoods, only a handful of people were ordering each day. The delivery density was too low. Luna: So the supply was there, but the demand was spread thin. And without enough orders per route, the cost per delivery was astronomical. Lucas: Exactly. I've seen estimates that in some suburban areas, Google was spending over $30 per delivery in subsidies, just to get the order density up. But they never reached the critical mass where the economics flipped. Luna: And they eventually shut it down in 2019. But they weren't the only ones. There's a graveyard of marketplaces that had supply but no demand. Lucas: Right. And the opposite is true too — think of all those peer to peer rental platforms for cameras or power tools that launched around 2015. They'd get thousands of listings because people love making money from idle stuff. But renters never showed up in enough density. Luna: Because I'm not going to rent a camera from someone thirty minutes away when I can just rent it from a store five minutes away. The micro-market density wasn't there. Lucas: That's the key insight. A marketplace can look healthy at the aggregate level — a million listings, a hundred thousand users — but if those users are spread across a country, each local market is empty. And marketplaces are fundamentally local. Luna: So the lesson is: don't just track total supply and total demand. Track the ratio within each geographic or categorical pocket. Lucas: Yes. And the most successful marketplaces I've studied don't even launch nationwide. They launch one city at a time, ensuring both sides are seeded simultaneously. Uber Eats did this really well. Luna: Yeah, they'd onboard a bunch of restaurants in a new city, but also run heavy consumer promotions for the first month — first delivery free, that kind of thing. They wouldn't let supply get too far ahead of demand. Lucas: Right. And they'd often use a technique called 'demand shaping' — they'd offer discounts on specific restaurants or cuisines to steer demand toward the supply that was available, rather than letting users order from places that weren't ready. Luna: That's smart. Whereas Google Shopping Express tried to launch simultaneously across dozens of metro areas, which diluted their marketing dollars and made each market under-dense. Lucas: There's research from MIT and Stanford that shows that the cost of acquiring a customer in an under-dense market is about three times higher than in a dense one, because the word of mouth effect doesn't kick in. If your friend isn't using the app, you're less likely to try it. Luna: So the network effect works on both sides. And if you only have one side, you lose the cross-side network effect entirely. Lucas: Exactly. Which is why some marketplaces have tried to solve this by owning both sides. Like, if you can't get enough independent suppliers, become the supplier yourself. Luna: You mean like how Amazon started as a retailer before opening its marketplace? Or how some home-services platforms hire their own workers in low-density areas? Lucas: Precisely. There's a term for it: 'managed marketplace'. The platform acts as the supplier of last resort to ensure liquidity. But that's capital-intensive and hard to scale. Luna: So what's the better approach if you're starting a marketplace today? How do you avoid the one-sided liquidity trap? Lucas: I think the best advice is to think in 'atoms', not 'bytes'. Each micro-market — a neighborhood, a category, a time slot — needs its own critical mass. Don't just track total users. Luna: And to actually measure that, you need to define what 'good enough liquidity' looks like. For Uber, it was 'a car available within five minutes in 80% of the city at peak times'. For a rental marketplace, maybe 'at least five listings within a two-mile radius'. Lucas: Right. And then you don't expand to the next micro-market until you hit that threshold in the current one. It's a disciplined, surgical approach. It's not sexy, but it works. Luna: Before we go further, I want to mention something that keeps this show going. We don't run ads — we keep the content focused and ad-free because listeners support us directly. Lucas: Yeah. And it makes a big difference. Even a couple of dollars a month genuinely helps keep these episodes coming. If you've gotten value out of the show, you can support us at buy me a coffee dot com slash fexingo. Luna: It's small, but it means we can keep digging into these marketplace mechanics without worrying about sponsors. So thanks to everyone who's chipped in. Lucas: Alright, back to density. So we've talked about geographical density, but there's also temporal density. Some marketplaces only work at specific times. Luna: Like a lunch delivery marketplace? Or a weekend-only service? Lucas: Exactly. For example, early versions of Uber had a hard time on Tuesday afternoons — supply would be idle because demand was low. They solved it by introducing UberPOOL and lower prices during off-peak times, which flattened the demand curve. Luna: So they used pricing to shift demand into the low-density time slots. That's a version of demand shaping too. Lucas: Right. And there's another angle: the take rate. When a marketplace has one-sided liquidity, its take rate is constrained. Because if you're a seller on a platform with few buyers, you won't tolerate a high commission. Luna: So the take rate is a function of liquidity, not just value. That's a great point. Lucas: Yeah. I've seen data that marketplaces with balanced two-sided liquidity can command take rates of 15 to 30 percent, while those with thin one side are stuck below 10 percent. And that's often the difference between a profitable business and a dying one. Luna: So the real work of building a marketplace isn't just getting users. It's getting the right users in the right place at the right time. Lucas: Exactly. And that's why I find this topic so fascinating. It's not about growth hacking. It's about designing for density. Luna: I want to ask about a specific industry: B2B marketplaces. Do they face the same one-sided liquidity issue? Lucas: Absolutely, but it's often more extreme. In B2B, the supply side might be a handful of manufacturers, and the demand side might be a handful of buyers. If one of them drops off, the whole marketplace collapses. Luna: So how do they handle it? I know some B2B platforms use a 'matchmaking' approach, where they manually curate both sides before opening the platform. Lucas: Right. They essentially pre-sell the demand to the supply and vice versa. It's like throwing a party where you call everyone beforehand to make sure enough people are coming. You don't just send out a mass invite. Luna: That's a good analogy. And I think that's the core lesson for anyone building a marketplace: don't just build the platform, build the density. Lucas: And measure it obsessively. Track the number of active buyers per seller in each postal code, each category, each hour of the day. If that ratio is below a threshold, don't scale — fix it. Luna: So if you're listening and you're working on a marketplace, what's one concrete thing you can do tomorrow to check for one-sided liquidity? Lucas: Pull the data on your top ten percent of micro-markets. See what the buyer to seller ratio is in those. Then compare it to your bottom ten percent. If there's a big gap, you've found your problem. Luna: And then you can run experiments to close that gap — like targeted subsidies, demand shaping, or even restricting supply in under-dense areas. Lucas: Exactly. Because a marketplace with one-sided liquidity isn't a marketplace at all. It's just a listing site with a high burn rate. Luna: Well said. And that's the note I think we should leave on.