Uber and Lyft, as they tend to do, made parallel announcements this week: The two companies, widely (and not quite accurately) synonymous with "ridesharing," are testing big new services that look, well, a little more like actual ridesharing.
Uber unveiled UberPool, in beta in San Francisco. Lyft introduced Lyft Line, a take on "personal transit" launching there as well. Both services are throwbacks to the traditional carpool as Silicon Valley would re-imagine it. Thus far, the two companies have wielded data to pair passengers and drivers-for-hire in real time. Now they're aiming to link passengers with drivers with other passengers, leveraging within their growing networks efficiencies where two or more riders heading in the same direction might travel there together.
The idea is both the logical next layer to everything both companies have built thus far and a sign of their ambitions to supplant much more than Friday night cab rides.
Trips through the new services would function a lot like they currently do through both apps. You call a car through the shared service. A nearby driver (a non-professional in his personal car) accepts the ride and picks you up. But then, on your way to your destination, the app continues to search for other potential passengers headed the same way.
Lyft claims that 90 percent of its rides replicate similar trips that someone else is taking within five minutes. That claim is vague, but it's plausible in a big, dense city: Researchers at MIT who've looked at taxi data in New York have concluded that 80 percent of all trips there could be shared with minimal inconvenience to each rider. Those MIT researchers envisioned a cab network centrally optimized through real-time data that looks strikingly like what Uber and Lyft are proposing.
With UberPool and Lyft Line, the rides will not be free (this is where the concept diverges again from the traditional carpool). But both companies are promising still bigger discounts off their cheapest services to date. Uber says a shared ride will cost 40 percent less than a solo UberX ride; Lyft is promising "up to 60 percent" off the original, with the price depending on the odds that another rider is traveling your way, too. With the even lower rates, both companies are angling for the parts of the transportation market that are seldom served by taxis: daily commutes to and from work, errand runs for which you wouldn't pay $10 a pop — but for which you might pay $5.
They're aiming, in effect, at the kinds of trips you'd more likely take in your own car—which also means they're aiming at your car itself. "At these price points," Uber declares in its blog post announcing the beta, "Uber really is cost-competitive with owning a car, which is a game-changer for consumers."
Here's how Lyft co-founder Logan Green put it to Farhad Manjoo at the New York Times: "We don’t want to be the option that people use only when they’re heading out for a nice evening. We want to be something that people use twice a day, every day.”
By invoking the idea of "personal transit," Lyft even seems to be aiming at rides you might otherwise take on the bus or train. A particularly intriguing passage from the company's announcement:
We’ve spent our entire lives moving around transit. Paying to rent near it. Sprinting to catch it. Timing our days to its timetables. But from today on, transit comes to us.
Piling more people into fewer cars makes economic sense for both companies. As they sprint to the bottom of the market, trying to collect three and four-dollar rides — opposite, for Uber, from the high-end towncar trips that launched the company — the strategy could come at a cost to their other customers: drivers. It wouldn't make sense for a driver to ferry you across San Francisco for $7. But it might make sense if he has three passengers in the back each paying $7.
Likewise, a $50 surge-pricing ride home from the bar on Friday night might not make sense for you as a rider. But the same trip could suddenly becomes more feasible if you combine it with several strangers.
For now, both companies say riders will be guaranteed the cheaper shared-ride rate even if the network doesn't turn up additional passengers to join you (yes, their models are strikingly similar in many ways). That means that Uber and Lyft will be on the hook to pay drivers the difference as they test whether the ride-matching math behind all of this will actually work in the real world.
What both companies are proposing is logistically and technically more complex than what they've done to date. Such a system optimizing shared rides across an entire city requires more data from each passenger (each now must give origin and destination details up front). It requires coordinating data among more parties (several of whom will be moving targets). It also requires nailing the details at the margins of each transaction: Are people willing to go two minutes out of their way — three minutes? four minutes? — to collect a fellow passenger? Will you wait on a street corner a little longer for a cheaper shared ride than a solo one? How much longer? How dense a network of users and drivers do you need to make those numbers work?
For Lyft, the idea of returning closer to ridesharing's roots — more people, using fewer cars efficiently — is at the core of its business. For Uber, the strategy is one piece of a bigger picture that involves providing transportation for every context and every kind of customer, from the businessman on a corporate Amex to the guy who normally takes the bus. Along the way, though, Uber will no doubt get closer through this exercise to honing the kind of back-end logistics platform that could eventually move just about anything.