How to Build the Next Uber or Lyft
If I had to build a competitor to Uber and Lyft, I would focus on standardization. Can we standardize time estimates? Can we standardize the kind of car that picks you up? Can we standardize the driver? Can we standardize the fare? Version 1.0 of ride-hailing was taxis, and version 2.0 is currently the anti-taxi. But version 3.0 will be a return to taxis, just with all the lessons in version 2.0 integrated.
I came to this conclusion when studying some curious frustrations I have with Uber and Lyft. Given the fierce competition between Uber and Lyft, combined with their total dominance, you’d think these flaws would have been ironed out by now. The fact that they persist says a lot about economics, incentives, and how great businesses can get stuck in suboptimal equilibrium.
Full Disclosure: I own some Lyft shares
Lyft and Uber are in an arms race to keep their drivers from working for the other guy. One technique is to automatically queue rides when a driver is almost finished completing a trip. It’s possible, then, to lock a driver into a consecutive series of trips with just their platform for an entire workday. Before the advent of auto-queuing, a driver was allowed to finish a ride, then switch to a competing platform.
In theory, auto-queuing is great for drivers, riders, and the platforms, but in practice, it leads to ghosting. Sometimes, after being assigned a driver, Uber or Lyft will mention to me that my driver is currently finishing a trip. If that happens, a sense of dread will come over me. The driver will sometimes put their phone in Airplane Mode to pretend their phone has lost its connection or that its battery is dead. This practice is similar to how video gamers will unplug their Internet when they’re about to lose a match. Uber and Lyft could ding drivers who do this, but they’re wary of doing so for fear of losing drivers.
I never get accurate time estimates for rides. These estimates should be trivial to fix, considering that Uber/Lyft have the travel history of billions of trips. However, because they know that users will comparison shop, they’re incentivized to give rosier estimates, and hope users forget which service is worse than the other in accuracy. The equilibrium state is that both services will give time estimates just low enough not to be memorable as a negative user experience uniquely attributable to that service. In other words, time estimates with mediocre accuracy are the stable state for both Uber and Lyft.
The time estimates also get worse if you use Uber Pool or Lyft Line. When you share rides with others, the fees are a bonanza for the platforms. They sometimes get 50% more fees for roughly the same cost of the ride, which leads to maybe 2x-5x more profit for the platforms. As a result, Uber and Lyft’s error rates on estimates for pooled rides are much greater than for single rides. This policy is short-sighted but inevitable. The operators know that users comparison shop, and so the rosiest option is the one that makes the most profit.
Pooled rides were supposed to be the holy grail for ride sharing. Pooling is what makes Uber and Lyft a net positive for society. If you order a Lyft for just yourself, the driver has to burn gas to retrieve you and drop you off, which is more gas than if you just drove on your own. That’s also one extra trip that a car takes up space on the road. Pooled rides are the answer, but their stupidly inaccurate estimates are a turn-off. Unfortunately, the incentives enforce the status quo.
The beauty of the Uber and Lyft business models is that they’re the ultimate two-sided networks. There’s the network of drivers, the network of riders, and the platform in between. When networks are aligned like this, especially accelerated by frictionless mobile apps, you get the fastest growing business verticals in history. The downside, though, is that all three players — the riders, drivers, and the platform operators — become locked in an incentive war. What’s good for one party isn’t always good for the others. The result is that these models, once reaching mature execution, cannot evolve. A similar problem occurs with Facebook which has a lock on a two-sided network: advertisers on one side, users on the other. Serving one can often lead to hurting the other.
The Way Forward
A parallel to the ride-sharing game comes from another transportation space: the airline industry. Passengers complain all the time about onerous fees and delayed flights, and yet we predictably go for the cheapest option. The airline model has a saving grace compared to ride-shares, though, which is that you still have options. You can spring for the classier airline if it’s available. And even among the discount ones, there are differences. For example, even though Southwest is a discount airline, they don’t nickel-and-dime, and they have relatively excellent customer service. The experience across all airlines is still awful, but at least you have more than two identical choices.
Ride-sharing is arguably already undergoing the transition beyond duopoly. In the U.S., you might have thought it was just Uber and Lyft, but don’t forget Wingz, which focus on airport pick-up and drop-offs. Wingz is hardly a blip on the Big Two’s radar, but it’s surprising they’re still around after seven years. Outside the U.S., though, Uber and Lyft aren’t even in the running. For example, in the U.K., Gett connects you with the old school black Hackney cabs, which one reviewer finds more comfortable than the typical Prius of Uber and Lyft. In India, Ola is the leader and they offers standard services such as free Wi-Fi in rides. If Uber and Lyft were the true lock-step duopoly that they appear to be in the U.S., we should see the same pattern across the world, and Wingz should be gone by now.
If there is a theme to the Uber and Lyft competitors, it’s standardization. I often yearn for a simple ride with no weird glitches. At first, it was cute to have different cars each time you took an Uber or a Lyft, just like it was cute to try different homes on Airbnb. But after some time, diversity without purpose gets old. (Remember we all used to customize our ring-tones?). Gett standardizes the car. Ola standardizes basic perks. And Wingz standardizes the airport experience. If a competitor to Uber and Lyft came around to San Francisco that was roughly competitive in price but offered a more reliable experience, I would switch in a heartbeat. There is still room in this duopoly. Or maybe we all need to get on scooters.