Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / How Uber Killed the Taxi Medallion
Transcript
- Lucas: Look at the price of a taxi medallion in Chicago in twenty fifteen. It was trading at over one point three million dollars. By twenty eighteen, it was worth less than two hundred thousand. That is not a gradual decline; that is a structural collapse. Luna: It wasn't just a price drop though. The underlying asset — the right to drive a cab in the city — didn't change quality. So what actually broke? Lucas: The liquidity did. Before Uber, the taxi market was a closed loop with a hard cap on supply. You needed a medallion to enter, which meant high barriers to entry, limited availability, and chronic shortages during peak hours. Uber opened the floodgates by removing the capital requirement for drivers. Suddenly, you had millions of potential suppliers waiting at the gate, ready to join the network the moment the incentive was right. Luna: But didn't Uber have to convince riders first? Most people still hailed cabs or used fixed routes back then. Lucas: They did, but they flipped the traditional marketplace playbook. Usually, you acquire demand first, then solve for supply. Uber solved for supply first by subsidizing drivers aggressively. They offered sign-on bonuses and guaranteed earnings floors. This created an immediate surplus of cars, which reduced wait times for riders, which increased rider volume, which justified more driver subsidies. It was a self-reinforcing liquidity loop that the taxi industry couldn't mathematically compete with. Luna: So the taxi companies were stuck with their expensive assets while Uber built a floating inventory that cost nothing to maintain? Lucas: Exactly. And here is where the take rate becomes the weapon. Uber initially charged a very low take rate, often around ten percent, compared to the fifty percent or more that taxi companies extracted through medallion leases and dispatch fees. This low take rate meant drivers kept more of every fare, making the platform instantly attractive to them even before they understood the consumer value. Luna: That feels risky for a startup. Burning cash on both sides simultaneously usually kills companies. Lucas: It does, unless you are playing for total market dominance rather than quarterly profitability. Uber accepted losses for years because they were buying network effects. Once the liquidity threshold was crossed — once there were always enough cars available within three minutes of a request — the taxi business model became obsolete overnight. The medallion holders were holding debt against an asset that no longer commanded scarcity. Luna: Did the regulators see it coming? I remember reading about medallion owners going bankrupt while Uber was expanding. Lucas: Most missed it because they were focused on the wrong metric. They looked at total ride volume and thought, 'Well, people still need rides.' They didn't realize that the value wasn't in the ride itself, but in the predictability and ease of access. Uber traded convenience for price control. The medallion system sold scarcity; Uber sold abundance. Luna: I think the real shock was how quickly the supply side turned. Once drivers realized they could make more money driving for Uber without paying lease fees, the exodus was instant. Lucas: Right. In New York, medallion revenues dropped by nearly half between twenty sixteen and twenty nineteen. But it wasn't just revenue per car; it was utilization. Taxis sat empty longer because the algorithmic matching of Uber was far more efficient at reducing deadheading. The average taxi driver spent more time cruising without a passenger than ever before. Luna: So efficiency killed the premium. If you can fill seats faster, you don't need to charge as much per mile to cover fixed costs. Lucas: Precisely. The taxi model was built on high fixed costs and low utilization. Uber’s model had near-zero fixed costs and high utilization. When you combine that with dynamic pricing, Uber could suppress prices during off-peak hours to keep drivers engaged, and raise them during peaks to balance supply. The taxis were stuck with flat meters. Luna: It reminds me of how streaming killed physical media. Not because the movies were better, but because the distribution layer changed entirely. Lucas: That is a good analogy. But in marketplaces, the distribution layer IS the product. You cannot separate the app from the service. If the app has fewer cars, the service is worse. So the competition isn't just for customers; it is for the entire ecosystem's health. Luna: Do you think other industries will face this same kind of liquidity shock? Like healthcare or legal services? Lucas: Absolutely. Any industry with fragmented supply and opaque pricing is vulnerable. Look at home services. Companies like TaskRabbit or Thumbtack are doing exactly what Uber did: lowering barriers to entry for providers, aggregating demand, and using data to optimize match rates. The key is identifying where the current incumbents are protecting rent-seeking behavior rather than providing value. Luna: Rent-seeking sounds fancy, but isn't that just what taxi unions and medallion lenders were doing? Lucas: It is. They were extracting value from the regulatory moat. Once the moat was breached by technology, the value evaporated. The lesson for operators today is that if your business model relies on artificial scarcity, you are building a sandcastle. Build models based on efficiency and liquidity instead. Luna: Yeah, because scarcity is easy to regulate, but efficiency is hard to stop. Lucas: If these conversations are useful for what you're building or running, consider supporting the show directly. It keeps the research sharp and the ads out. Honestly, if today was worth a coffee to you, that's the link — buy me a coffee dot com slash fexingo — that's F. E. X. I. N. G. O. Luna: Appreciate it. It really helps us dig into these deeper structural shifts without cutting corners. Lucas: Back to the liquidity trap. One thing people overlook is how Uber handled the transition after dominance. Once they won, they raised take rates significantly. In some markets, it went up to twenty-five percent or more. Luna: Didn't that cause driver churn? I heard complaints about drivers leaving for competitors like DoorDash or Lyft. Lucas: It did cause friction, but the network effect acted as a lock-in. Drivers stayed because the volume was there. Even with a lower cut, the absolute number of trips was higher on Uber than on any smaller competitor. This is the classic winner-take-all dynamic. Once you cross the liquidity threshold, you can extract more value because switching costs become prohibitive for users who value reliability over minor price differences. Luna: So the initial subsidy was basically a customer acquisition cost, but for the supply side? Lucas: Exactly. It was the cost of buying the network. And unlike traditional marketing, this cost compounds. Every new driver adds value to every existing rider. That compounding is what makes two-sided markets so powerful, and so dangerous for incumbent players who aren't built for scale. Luna: It makes me wonder if the taxi companies tried to adapt at all. Did any of them launch their own apps? Lucas: Some tried. Yellow Cab launched its own hailing app in various cities. But they were hamstrung by their own infrastructure. Their drivers were leased medallion holders, not independent contractors. Changing the employment model would have required massive capital restructuring. Uber had no legacy baggage. They could experiment with pricing, insurance, and routing in real time. Luna: Legacy baggage is a killer. It ties your hands when you need to move fastest. Lucas: Always. And let's talk about the data advantage. Uber wasn't just moving cars; they were mapping the city in real time. They knew exactly where demand spikes were happening, down to the block level. Taxi companies relied on historical patterns and random cruising. This data asymmetry allowed Uber to position cars proactively, further reducing wait times and increasing driver earnings. Luna: So data became part of the liquidity engine. Better data means better matching means more liquidity. Lucas: Yes. It creates a feedback loop. More usage generates more data, which improves the algorithm, which attracts more users. The taxi industry was flying blind by comparison. They couldn't compete on intelligence. Luna: It seems like regulation finally caught up, though. Cities are now capping licenses or requiring minimum wages for gig workers. Lucas: They are. And this is the next phase of the story. Platforms like Uber are now facing the same regulatory scrutiny that taxis faced decades ago. But the dynamic is different. Now, the pressure is on labor standards and classification, not just entry barriers. The medallion system was about controlling supply; the current debate is about controlling labor costs. Luna: Does that threaten the liquidity model again? If drivers earn less, will they leave? Lucas: It creates volatility. We've seen fluctuations in driver availability when wage laws change. But the core mechanism remains. As long as the platform offers better liquidity than alternatives, it will attract participants. The challenge for these companies is balancing profitability with the sustainability of their supply base. Luna: Which brings us back to take rates. If you raise them too high, you break the supplier trust. If you keep them too low, you never make money. Where is the sweet spot? Lucas: There isn't one static number. It depends on the elasticity of supply in that specific market. In dense urban areas, where many drivers are already competing, you can push take rates higher. In rural or emerging markets, you need to be generous to bootstrap supply. It's a constant calibration act. Luna: Calibration is a polite word for walking a tightrope while the ground is shaking. Lucas: That is exactly it. And for investors watching these metrics, understanding that calibration is more important than the headline take rate percentage. You need to look at retention, utilization, and net promoter scores to see if the liquidity is actually healthy. Luna: Healthy liquidity means happy drivers AND happy riders. Anything else is just temporary manipulation. Lucas: Agreed. The medallion crash taught us that assets lose value when they fail to facilitate efficient exchange. The winners are the platforms that turn friction into flow. Luna: So the question isn't who owns the medallions, but who owns the flow. Lucas: Exactly. And as new technologies emerge, whether it's autonomous vehicles or drone delivery, that ownership of flow will shift again. The only constant is the need for liquidity. Luna: Interesting. Do you think autonomous fleets will consolidate the market further, or fragment it? Lucas: Likely consolidate. Robotaxis will be expensive capital assets. Only the largest platforms will afford to deploy them at scale. This could lead to even tighter oligopolies in urban transport. Luna: A future owned by algorithms and battery packs. Scary stuff. Lucas: Efficient, but scary. We'll have to see how society balances those trade-offs. Luna: One thing for sure: the taxi medallion era is gone forever. Lucas: It is. A stark reminder that in business, relevance beats regulation every time. Luna: Relevance wins, until it doesn't. What comes next? Lucas: That is the billion-dollar question hanging over every platform today. Luna: Let's hope they figure it out before we run out of cities to map. Lucas: Here is hoping. Until next time.