Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / How Uber Splits the Fare The Economics of Take Rate
Transcript
- Lucas: So Uber takes about 25 percent of every ride. That's the take rate. And if you're a driver, you've probably felt that number in your earnings. But why 25 percent? Why not 10 or 40? Luna: And is it the same across all trips? I've heard drivers say it depends on the market or the time of day. Lucas: It's not identical everywhere, but the headline number has settled around there for years. And it's worth asking how Uber arrived at that figure — because the take rate is one of the most strategic levers a marketplace has. Luna: Right. Too high, you choke off supply or demand. Too low, you never make money. There's a sweet spot. Lucas: Exactly. And Uber's history shows how that sweet spot shifts. When they launched, they took about 20 percent. By 2015, it was creeping toward 25. By 2019, in some markets, it hit 30. Then came the IPO and pressure to show a path to profit, and drivers started protesting. Luna: There were strikes in 2019, right? Drivers in multiple cities turned off the app. Lucas: Exactly. And Uber's response was instructive: instead of cutting the take rate across the board, they introduced 'boost' incentives and guaranteed earnings for drivers who hit certain thresholds. They essentially used dynamic pricing on the supply side to keep drivers happy without lowering the headline commission. Luna: So they kept the take rate high but gave drivers a way to earn it back through performance. That's clever — but it also means the effective take rate is lower for high-performing drivers. Lucas: Yes. And that's a key insight: the quoted take rate is rarely the real take rate. In any marketplace, what matters is the net take rate after all subsidies, incentives, and refunds. Uber's net take rate in 2024 was around 18 percent after driver incentives and insurance costs. Luna: So the headline number is almost a marketing figure. The real economics are buried in the fine print. Lucas: And that's true for every marketplace. Take Etsy — they charge a 6.5 percent transaction fee, plus listing fees, plus a 15 percent offsite ad fee if someone discovers a product through Google. The blended take rate can be over 20 percent for some sellers. Luna: But Etsy's sellers are small businesses, not gig workers. They have different tolerance for fees. A driver might bolt at 30 percent, but a crafter might accept 20 because they have higher margins on their products. Lucas: That's the central question for any marketplace: what is your specific audience willing to pay? And the answer changes over time. Uber's take rate in 2012 was lower because they needed to attract drivers. By 2019, they had a huge pool of drivers and could afford to push it up. Luna: So the take rate is a function of market power. As a platform gains liquidity, it can charge more. Lucas: Exactly. And that's why two-sided marketplaces are so valuable — they capture more value as they grow. But there's a limit. If you push too far, you damage the liquidity you depend on. Luna: That's what happened with eBay in the early 2000s. They raised fees, and sellers revolted. Some left for Amazon or created their own sites. Lucas: Right. eBay's take rate on auctions was around 8 percent for a while, then they added insertion fees and final value fees that could total 15 percent. Sellers calculated their total cost of sale and started migrating. eBay lost a lot of its supply side for a few years. Luna: And recovering from that took time and fee cuts. So the lesson is: know your users' alternatives. If a driver can switch to Lyft with a click, you can't charge them 40 percent. If a seller has no other platform with the same buyer traffic, you have more room. Lucas: That's exactly the framework. Now, there's another way to think about take rate: as a fee for liquidity. Uber doesn't just connect riders and drivers — it provides immediate demand. A driver sitting in an airport lot might pay 30 percent for a ride that materialises in five minutes instead of waiting 30. That convenience has value. Luna: So the take rate reflects the speed and reliability of the match. If you can guarantee a ride within three minutes, you can charge more than a platform where it takes 15. Lucas: And that's why Uber's take rate is higher than, say, a food delivery platform like DoorDash. DoorDash's take rate averages around 15 percent because the delivery time is longer and the unit economics are tighter — they have to pay a delivery driver per order, and the order value is lower. Luna: But then you have something like Airbnb, which takes about 14 percent from hosts and 5 percent from guests — so 19 percent total. That's in between. And that makes sense because the transaction value is higher — a night's stay might be $200 — so the platform can take a smaller percentage and still make good money. Lucas: Exactly. And the take rate also depends on how much value the platform adds. Airbnb provides insurance, dispute resolution, a review system, and a massive user base. That 14 to 19 percent covers a lot of services. Compare that to a simple classifieds listing, where the platform does almost nothing — take rates there are often zero or a flat fee. Luna: So the take rate is really the platform's margin for the work it does. The more complex the transaction, the higher the justified take rate. Lucas: Right. And marketplace operators have to constantly experiment. Uber runs A/B tests on take rates in different cities. They'll try 27 percent in one market and 23 percent in another, see which yields the best long-term growth in trips and driver retention. Luna: That sounds like a delicate balance. If you lower the take rate, you might attract more drivers, which means shorter wait times, which means more riders — but you also earn less per ride. The optimal point depends on elasticity on both sides. Lucas: And that elasticity changes. During the pandemic, demand for rides plummeted, and Uber cut its take rate to keep drivers from quitting. In 2020, their net take rate dropped to around 15 percent in some markets. As demand recovered, they raised it back. Luna: So the take rate is not a fixed strategy — it's a dynamic tool. You adjust it based on the health of the marketplace. Lucas: Exactly. And that's why the most successful marketplaces are the ones that understand their own unit economics deeply. They know the marginal cost of a transaction, the lifetime value of a user, and the churn rate at different fee levels. Luna: That kind of data-driven approach is what separates a marketplace that survives from one that flames out after a few years. Lucas: And on that note — I want to be honest about something. This show is entirely funded by listeners. A small group of you chip in monthly through buy me a coffee dot com slash fexingo, and that's literally what makes it possible to dive into these topics without ads or sponsors. Luna: Yeah. It's a direct relationship with the audience, and we're grateful for it. It lets us focus on the substance without worrying about pleasing advertisers. Lucas: Exactly. So if these deep dives into marketplace mechanics are useful to what you're building or running, that link is the only way the show stays alive. No pressure, just a transparent look behind the curtain. Luna: And speaking of transparency — let's get back to understanding how different platforms set their take rates. One fascinating case is how Thumbtack set theirs. Lucas: Thumbtack, the platform for local services like plumbers and photographers, initially charged service professionals per lead — a flat fee per potential customer contacted. That's a different model from a transaction-based take rate. It worked for a while, but professionals complained about the quality of leads. Luna: So they switched to a booking-based model, taking a percentage of the job cost — around 15 to 20 percent. That aligned incentives: Thumbtack only gets paid if the pro gets the job. Lucas: Exactly. And that shift dramatically improved trust on the supply side. Pros were willing to pay a higher effective fee because it was contingent on a confirmed booking. The take rate became a variable cost, not a fixed cost. Luna: So the structure of the take rate matters as much as the amount. A flat fee per lead feels like a tax; a percentage of a completed job feels like a partnership. Lucas: That's a crucial distinction. And it's why many marketplaces have moved toward outcome-based pricing. Look at Upwork, the freelancing platform. They charge a sliding scale: 20 percent on the first $500 billed with a client, then 5 percent after $10,000. That encourages long-term relationships and rewards loyalty. Luna: So the take rate is not just a number — it's a message. A high take rate says 'we provide a lot of value.' A sliding scale says 'we want you to stay.' A flat fee says 'pay per match, regardless of outcome.' Lucas: And that message needs to match the reality of the marketplace. If you charge a high take rate but your matches are poor, users will churn. If you charge a low take rate but your service is excellent, you're leaving money on the table. Luna: So it's a constant calibration. And the best marketplace operators are always tuning it, testing it, and listening to the feedback from both sides. Lucas: One more example: the ride-hailing market in India. Ola and Uber have been locked in a price war for years, and the take rates are much lower there — around 15 percent. Because drivers have low margins and riders are price-sensitive. The platform can't extract much without destroying liquidity. Luna: So the take rate reflects not just the platform's value, but also the economic reality of the region. The same platform in different markets may have different optimal rates. Lucas: Exactly. And that's why global marketplaces often localize their pricing. Uber's take rate in New York might be 28 percent, but in Mumbai it's 12. The cost of living, regulatory environment, and competition all play a role. Luna: So what's the takeaway for someone building a marketplace today? How do you find your right take rate? Lucas: Start with benchmarking. Look at similar platforms. Then run experiments. Start high, then lower it and measure the effect on conversion and retention. But also interview your users — both sides. Ask them what they'd be willing to pay for a perfect match. Their answers might surprise you. Luna: And remember that the take rate is not static. As your marketplace grows, your value grows, and you can adjust — but do it carefully. A sudden hike can destroy trust that took years to build. Lucas: Right. The take rate is one of the most powerful tools a marketplace operator has. Use it wisely, and it can fund growth, improve liquidity, and build a sustainable business. Abuse it, and you'll drive your users to the competition. Luna: And that's the fine line between a thriving ecosystem and a ghost town. Lucas: Exactly. So if you're running a marketplace, pay attention to the number — but also to the story it tells.