Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / How the Best Marketplaces Use Data to Pick Their Take Rate
Transcript
- Lucas: If these conversations are useful for what you're building or running, a couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo, if you've gotten something out of them. Luna: Yeah, it's a small thing that adds up. And it keeps us ad-free, which I think makes the conversations better. Lucas: Absolutely. So today we're talking about take rates — the percentage of every transaction a marketplace keeps. And the conventional wisdom is that the number should be whatever the market will bear. Luna: But that's not actually what the best marketplaces do, right? Lucas: Not at all. The smart ones set their take rate based on data about what maximizes long-term liquidity. And there's a really interesting example from the early days of the New York Stock Exchange that illustrates this. Luna: Wait, the NYSE? That's a marketplace — it matches buyers and sellers of stocks. Lucas: Exactly. In the 1790s, when the exchange was just a handful of brokers meeting under a buttonwood tree, the take rate was essentially zero. They were just trading stocks for each other, no commission. But as volume grew, they realized they needed to charge something to cover the cost of the physical space and the settlement. So they introduced a fixed commission rate — 0.25 percent per trade. That was the first take rate. Luna: And that rate stayed basically the same for over a hundred years, didn't it? Until the 1970s when they had to deregulate? Lucas: Right. The NYSE kept that 0.25 percent because it was enough to make the exchange profitable while keeping liquidity high. They understood that if they raised it too much, brokers would just trade on the street. That's the fundamental tension: charge too much, and the marketplace disintermediates. Charge too little, and you can't invest in trust and safety. Luna: So how do modern marketplaces figure out the right number? Let's look at Uber. They take about 25 percent on average, but that varies by city and trip type. Lucas: Uber's 25 percent is actually a fascinating data story. When they launched, they experimented with different commission rates in different cities. They found that if they charged above 30 percent, drivers would start poaching riders offline — handing out business cards, building direct relationships. Below 20 percent, they weren't making enough to cover insurance and background checks. So 25 percent was the sweet spot that maximized both driver supply and rider demand. Luna: But that's not a static number. In 2026, Uber's take rate has actually crept up slightly in some markets because they've layered on advertising and subscription services. The commission itself hasn't changed much, but the overall monetization has. Lucas: That's a key point. The headline take rate is just one part of the revenue model. Etsy, for example, charges a 6.5 percent transaction fee. That seems low compared to Uber, but Etsy also makes money from listing fees, promoted listings, and payment processing. So their effective take rate is closer to 15 percent when you add it all up. Luna: And yet sellers don't revolt because they can see the value. Etsy brings them buyers they wouldn't find on their own. The data Etsy provides on search trends and pricing is actually worth more than the fee. Lucas: Exactly. And that's where the data comes in. Marketplaces that share data with their suppliers — like average selling price, conversion rates, seasonality — can actually charge a higher take rate because the supplier perceives more value. Etsy's seller analytics dashboard is a retention tool as much as a product feature. Luna: But what about the marketplaces that are lowering their take rates right now? I'm thinking of some of the delivery apps that are facing competition from restaurants doing their own delivery. Lucas: In 2026, we're seeing a wave of disintermediation in food delivery. Restaurants are using buy now pay later services to offer discounts on direct orders, and some are even building their own delivery fleets. So platforms like DoorDash have started reducing their commission for restaurants that use their own drivers — from 30 percent down to 15 percent. They'd rather take 15 percent of a transaction than zero. Luna: That sounds like a defensive move, but it's actually a data play too. By keeping the restaurant on the platform, DoorDash still gets the transaction data, which it can use for its advertising business. The commission is almost secondary. Lucas: Right. And that's the lesson: the take rate is not just a number on a pricing page. It's a strategic lever that affects liquidity, data collection, and competitive positioning. The best marketplaces don't copy each other's take rates. They derive theirs from their own unit economics. Luna: There's a startup called Clearco that isn't a marketplace itself but funds marketplace sellers. They've analyzed thousands of sellers across different platforms and found that the optimal take rate for a marketplace is actually a function of the seller's gross margin. A seller with 50 percent gross margin can afford a 20 percent take rate. A seller with 20 percent margin can't. Lucas: That's a really useful framework. So if you're building a marketplace, you should look at your sellers' unit economics, not just your own. If your take rate is pushing sellers below their break-even, they'll leave. And that destroys liquidity. Luna: Which brings us back to the NYSE. They kept that 0.25 percent for over a century because it allowed brokers to make money while keeping the exchange liquid. When they finally deregulated in 1975, the take rate dropped to near zero for large trades, but volume exploded. The exchange made more money on a lower rate. Lucas: That's the paradox of take rates: a lower rate can sometimes lead to higher total revenue because it attracts more transactions. But you need the data to know where that tipping point is. And that's why every marketplace should be running constant experiments on pricing. Luna: So for our listeners who are building or running a marketplace, what's the one number they should focus on? Lucas: It's not your take rate. It's your liquidity ratio — the number of completed transactions per day per active supplier. If that's growing, your take rate is probably right. If it's flat or declining, you need to adjust. The take rate is a means, not an end. Luna: And don't just copy Uber or Etsy. Do the math on your own sellers' gross margins. Lucas: Exactly. Because the right take rate for your marketplace is the one that maximizes long-term liquidity, not short-term revenue. And that's a data-driven decision, not a guess.