Uber's bold move

Image: Ethan Pines for Forbes

Dressed in a grey polo, matching grey chinos and black sneakers, Travis Kalanick tunes out the noise and locks in. He paces the length of a conference room like a basketball coach, alternating between sips of coffee and bits of trail mix he picks out of an espresso-size paper cup. Around the table, six young men present the Uber co-founder and CEO with the early results of a critical initiative: A new version of the Uber app that launched three weeks earlier. What users see as simple tweaks in the app’s design have profound impacts on app downloads, usage, ratings, pickup times, retention rates, load times, the distribution of users who choose UberPool over UberX and much more. Those impacts vary by country and by the type of phone the rider uses.

It’s the ultimate logic quiz. For 80 minutes Kalanick scrutinises every chart, questioning assumptions. “This could be a measurement issue or a real problem,” he says, pointing to a seemingly arcane metric. He repeatedly pulls out his iPhone to check for himself how such details might affect actual users, alternating between satisfaction and mild annoyance. At one point he says that without “real data” on a specific feature, “emotion will rule the day”. At Uber that would be a very bad thing.

This process, one of the fundamental building blocks of the Uber machine, is a called a jam session. Jams determine how problems get turned into ideas, how ideas get turned into products and how products get reviewed with an eye to their impact on Kalanick’s overriding obsession: The efficiency of Uber. Jams are also how Kalanick touches almost every important aspect of the Uber experience.

One of the most talked-about people in Silicon Valley, the 40-year-old (No 64 on Forbes’s World’s Most Powerful People list) has been described as many things—most of them unflattering. Ruthless and unethical, an evil genius and a loose cannon, a “bro” and a “douche”. There’s a seed of truth in all of those. But the terms miss the special sauce, the unifier that explains how Kalanick has driven Uber to become the richest startup in history, with a valuation of $68 billion.

As the jams demonstrate, Kalanick’s ultimate professional trait—the part that channels the hyper-competitiveness, harnesses the intensity and mitigates any personal flaws—is troubleshooting. He likes to call himself Uber’s problem-solver-in-chief. When you watch him jam, it’s easy to see the joy he takes in that role, as he compares a small and particularly successful tweak to a “super-gangster move”, his face lighting up with a kid-in-a-candy-store smile that further narrows his quasi-permanent squint. Many founders, Kalanick included, have vision. Others, Kalanick not included, are fanatical, public evangelists. Kalanick views his role as driving Uber forward through a series of logical hurdles, which must be jumped, endlessly.

“Every problem is super-interesting and has its own nuances, and you solve it today, but you try to solve it with an architecture. You build a machine to solve the problems that are like it later,” Kalanick says. “And then you move on to the next.”

There’s always a next—and those nexts keep getting bigger. That’s what makes Uber one of the world’s most interesting companies and explains why Kalanick can raise as much money as he wants—$16 billion in equity and debt so far—with neither a profit nor an IPO in sight. Uber’s business model as a frictionless middleman (for ride-hailing, in its case) is so powerful that it’s become a cliché. Well over 100 startups have been described as the Uber of something, from Honor (Uber of home care) to Wag Labs (Uber of dog walking). One of the few Valley companies that are not trying to be the Uber of something? Uber. The more apt comparison: Amazon, which started as a company synonymous with online bookselling and morphed into the internet’s retail megastore and more. Kalanick is no longer interested in just getting you a ride: He’s positioning Uber to be at the centre of mobility. If it moves, Kalanick wants a piece of it. Less than seven years after launch, Uber is already reshaping how cities think about public transit, parking and congestion, and how Millennials think about car ownership. And it has barely scratched the surface in terms of moving physical goods. “The car market, the transportation market, consumer ground transportation—you might think of it as $5 trillion or $6 trillion globally, but honestly I don’t know if it really matters,” Kalanick says. “The point is it’s in the trillions.”

Over the past two years Uber has not only dramatically expanded its ride-sharing offerings globally but also pushed into delivery ser­vices like Eats (food), Rush (anything, quickly) and Freight (watch out, long-haul trucking). It has played with marketing stunts like UberChopper (helicopters), UberSeaplane and UberBoat, usually around specific events. It’s also investing seriously in self-driving cars and self-driving trucks, and it has even proposed, Elon Musk-like, a far-fetched blueprint for flying cars.

If Uber’s core ride-sharing market is any example, each of those toeholds could soon be a full-fledged foot in the door. Uber offers a whole menu of ride-sharing options like X, Pool, Black, Select and, in some countries, Moto (motorcycles), spending billions to subsidise drivers as part of what critics say is a deliberate strategy to put rivals out of business—and freely raise prices after that. Forbes estimates that Uber’s losses could reach about $2 billion in 2016 on revenue that could easily top $5 billion. Profits today are secondary to market dominance tomorrow. Jeff Bezos, eat your heart out.

“When I first backed this company, I had no idea that one day I would be listening in on auto-industry earnings calls,” says venture capitalist Bill Gurley, of Benchmark, one of Uber’s first investors and a board member. “There is so much anxiety about what this company could do to one of the biggest industries on the planet.”

Travis Kalanick has always seen patterns. As a teenager he analysed traffic on southern California’s famously congested freeways with an eye to finding the optimal lane for every possible situation. His computer-science education at UCLA, even though he eventually dropped out, helped him hone his problem-solving mind-set. While still in school, he started Scour, a multimedia search engine and file-sharing exchange that flopped. At his next startup, Red Swoosh, which made it easier for media companies to deliver video files online, Kalanick endured bad investment proposals, a lowball acquisition offer, leaks suggesting the company was toast, staff desertions, investor disaffection and more before he finally managed to sell it to Akamai Technologies for $23 million. Kalanick has called these his “blood, sweat and ramen years”.

Kalanick founded Uber in 2009 with Garrett Camp, a fellow entrepreneur who had recently bought back his old company, StumbleUpon. It was built around an app, called UberCab, that took the hassle out of booking a black car. The service, launched in 2010, initially was little more than a toy for Kalanick and his friends to get around San Francisco. But pretty quickly, Kalanick began to understand how math could help Uber disrupt not just limousine services but urban transportation itself. If ­prices went down, more riders would be interested, which in turn would lure more drivers onto the platform, which meant wait times would drop, bringing even more riders on board and helping drivers earn more. Fine-tuning that virtuous cycle has become Kalanick’s obsession ever since.

The result has been the fastest ascent in Silicon Valley history, as Uber outpaced even rocket ships like Google and Facebook, with revenue that exceeded $1 billion in the second quarter and a workforce of more than 9,000 employees and 1.5 million drivers. (More people earn a paycheck—or part of one—from Uber than from any other private employer in the world except for Wal-Mart and McDonald’s.) Uber has rolled out its app—often bulldozing regulatory hurdles and vocal opposition from taxi drivers—in more than 450 cities across 73 countries. In any given month 40 million people will take an Uber ride, and its drivers will collectively cover 1.2 billion miles, or about 35 times the distance between Earth and Mars. Kalanick’s goal now is to “make transportation as reliable as running water.”

The primacy of efficiency has only increased with this gargantuan scale. Kalanick, whose net worth Forbes estimates at $6.3 billion—he says he hasn’t sold any of his Uber shares—manages this by breaking every component into a problem to be solved. “You create a system, which could be code or process, and guess what—we are in the world of bits and atoms, so it’s process and code, and honestly, that’s every problem,” Kalanick says, before correcting himself. “It’s people, process and code.” Phew.

“Travis likes to encourage this philosophy of rigorous experimentation and testing,” says Brian Tolkin, the 26-year-old head of UberPool, which groups riders into shared vehicles. Jams can extend across days and even weeks. Recently, a cross-department set of questions—how Uber could dole out incentives more efficiently—turned into a series of gatherings involving engineering, finance, data science and operations executives. “He jammed for two weeks straight, and sometimes it was three or four hours a day,” says Thuan Pham, Uber’s CTO. All that effort resulted in a relatively small outcome: A fine-tuning of departmental responsibilities, allowing quicker marketplace responses.

Given Uber’s mammoth size, aggregated tweaks can make a huge difference. On the technical side, jam sessions have led to more than 1,000 separate but integrated “services”—bits of technology that encompass the Uber apparatus. “The more efficiently we can do this, the more money the drivers make,” says Pham, 48, who fled Vietnam as one of that country’s “boat people” after the Vietnam War and has been at Uber for nearly four years. The promise and the experience—push a button, get a ride—have become commonplace, but delivering on them demands myriad bits of code operating in perfect harmony.

Consider some of what happens when a user opens the app: Her location gets beamed to Uber’s servers, and dispatching software begins searching for pricing algorithms and scouring the map for nearby vehicles; all of that gets updated every few seconds, before the user even requests a ride. Once a driver accepts a ride, the routing software directs the car to a location, updating ETAs based on GPS readings every four seconds. That near-continuous tracking goes on during the ride, and often before it is over the driver is predispatched to the next trip. Then there’s billing, processing, ratings and the company’s own analytics to evaluate the quality of the ride. The complexity of it all gets amplified with Pool, as new ­pieces of software get activated to determine who else may want to ride along a similar route and to calculate possible matches that don’t prolong a ride by much—a calculation that changes continuously with traffic conditions. Even some matches that meet the criteria have to be discarded because they require drivers to backtrack. “Psychologically, our riders don’t like to go backwards,” Pham says.

That code—and more important the process (and, yes, people) that create it—forms the backbone of Uber’s forays into the larger world of mobility. The company has devised a method to consider investments in new systems and technology. “We focus first on the existential threats, or we won’t have a business,” Pham says. The most notable in that category—and likely the biggest in dollar terms—is self-driving cars, but countless others involve analytics, surge pricing, routing and data centre technology.

Next comes nice-to-have features. And given Uber’s growth, everything becomes something of a game of lather, rinse, repeat. Even when Uber builds a system that can cope with 10 times its existing capacity, it has to scrap it within 12 to 18 months. “In 3.5 years we have rewritten our dispatch system three times,” Pham says. As we talk inside a small conference room, his phone rings with the sound of an emergency-vehicle siren, and he excuses himself briefly. When he returns, he tells me only two events trigger the jarring ringtone: A “level 5” outage, when Uber’s system grinds to a halt—it happens very rarely—or a call from Kalanick. This one was the latter.

A couple of years ago, when Kalanick became convinced that mapping was one of the “existential” technologies Uber had to own, he poached Google’s Brian McClendon, a co-creator of Google Earth who had led Google’s mapping efforts for years. Few people understand better the massive investment required—Google has spent many billions—and why it’s worth it. “Everything about the business depends on maps,” McClendon says.

While Google Maps—and navigation services like Waze—have helped Uber, there are many things they can’t do or do poorly, McClendon says. Some are merely annoying. When a user in San Francisco begins typing a destination, say “den”, Google Maps might suggest Denver, which isn’t a realistic result for an Uber ride. Other issues are subtler but essential for an efficient service. When Uber decides which driver to match to a rider, it’s not enough to know where the drivers are—the company also needs to know where they will be a few seconds later when they may be ready to accept a ride. If by then they’ve passed an intersection or, on a freeway, gone by an exit, they might no longer be the best match.

As it maps and remaps the world, Uber collects data to improve pickups and drop-offs—say, figuring out ideal locations in a mall with multiple entrances—using machine learning and statistics to analyse historical data. This is key as Kalanick experiments with a new idea in efficiency: Advising riders to walk to an optimal pickup spot to avoid a particularly congested street or a needless drive around the block. At the same time the company’s camera-equipped vehicles are also busy capturing street signs that Uber’s system can detect and, using machine learning, understand (“No Stopping,” for example) so it can avoid getting its drivers into trouble.

And, of course, maps are essential to Uber’s two-year-old effort to develop self-driving cars. Kalanick’s most celebrated moves in that area came in August, when he simultaneously announced that some Uber rides in Pittsburgh would be in self-driving cars (there’s still a driver behind the wheel, to prevent mishaps and to comply with regulations) and that the company had paid $600 mil­lion to acquire Otto, a startup by veterans of Google’s pioneering autonomous vehicle group, which has made quick progress with self-driving trucks. The moves gave Uber a strong position in a critical area that over time will determine how people or things move from A to B. Says Kalanick: “We’re working really hard to make this a reality as soon as possible.”

In February Kalanick took the stage in Vancouver for his first-ever TED talk. It didn’t go well. Kalanick was introduced with the “evil genius” tag, and his speech—an argument against regulation that could quash Uber’s potential to improve the life of cities—was delivered nervously and defensively. After retreating backstage, Kalanick ran into Target’s chief marketing officer, Jeff Jones, widely credited with revitalising the retailer’s business. Kalanick asked Jones what he thought of his talk. Jones’ reply? B-minus. It was the kind of tough-love answer a logician could appreciate. Within a few days Kalanick, along with Gurley, was recruiting Jones in Minneapolis, where Target is based; by August he was on board with the title of president, ride sharing. “Travis focuses on who he needs on his team and goes after them,” says Ariana Huffington, who joined the Uber board of directors earlier in the year.

As Uber seeks to become the planet’s operating system for transportation, it remains, by most accounts, a punishing workplace. But those around him maintain that Kalanick is mellowing out and moving past his scorched-earth insurgent days. Strong hires and a stable management team are one sign. His China defeat is another. While no one inside Uber would use that term—the company has turned a $2 billion outlay into a $7 billion stake in its Chinese competitor, Didi Chuxing—the stunning turnabout indicates a more mature Kalanick.

The Uber CEO, after all, was obsessed with China—“The more people said, ‘Don’t do it,’ the more I’m like, ‘Wow, I think there is something here that people don’t know’”—and he went all in. Uber’s China adventure quickly became an amplified microcosm of the company’s story. The service caught on like wildfire, expanding to 60 cities as the team grew to 800 employees. Within two years China accounted for a whopping one-third of Uber’s trips; of Uber’s top-10 cities by number of rides, eight were in China. And as in many other countries, the massive growth came with massive losses—the consequence of a brutal subsidy war with Didi. Perhaps worse, China was sucking up too many resources and too much attention from Uber’s engineers, product developers, businesspeople and executives, including Kalanick.

So this past summer Kalanick made peace, agreeing to fold Uber China into Didi. “It was really impressive to see how Travis had recognised this not as a personal defeat but as what was best for Uber,” Huf­fing­ton says. “He saw the opportunity cost of continuing to fight in China was very significant.” Kalanick himself is sanguine about the episode: “When people say, ‘You didn’t get what you want,’ well, hey, that’s fair. But we never claimed that we are always winning 100 percent,” he says. “We had a valiant effort, good purpose, and we showed people that something could be done that they didn’t think was possible.”

What’s possible for Uber now is pretty unlimited. The company retains an unprecedented cash hoard, and with the bruising China battle over, Forbes believes Uber’s losses may have peaked. Kalanick can now focus on winning in critical markets like the US, Brazil and India, even as it branches out into other industries within transportation. “When you go and raise $10 billion, you are doing that to invest it,” Kalanick says.

Lyft claims its US service matches Uber’s in major cities, “which is where 90 percent of all rides happen,” says Lyft president and co-founder John Zimmer. That’s where Kalanick’s efficiency obsession comes in. If he can’t offer a clear differentiation in terms of when a driver shows up, then he can leverage his machine to put more people in each car—delivering cheaper rides for customers, bigger paychecks for drivers, more revenue for Uber. That’s why Kalanick has been jamming so much around Pool, which is growing like a weed, in hopes of beating Lyft’s carpooling option. In its hometown of San Francisco, 40 percent of Uber rides are now shared.

Pool is also critical for Kalanick’s goal of improving the life of cities, as Uber disrupts public transportation and even urban planning. The city governments that decried Uber, and its ability to break into government-sanctioned taxi monopolies, love the idea of taking cars off roads.

And Kalanick is already selling them hard on the benefits. Summit, New Jersey, for example, recently decided to subsidise Uber rides for commuters rather than build additional parking at its train station. The long-term reputational payoff—Uber suddenly becomes the good guy—is obvious.

Efficiency-first thinking dominates how Uber sees the transition to self-driving cars. It’s a when, not an if, and it will radically change the economics of Uber from an asset-light business to one where it would have to commit loads of capital to roll out a vehicle fleet. Critics say Uber’s lack of a dedicated vehicle manufacturing unit could be its Achilles’ heel, as carmakers like General Motors (in partnership with Lyft), Ford and Tesla—and perhaps others like Google or Apple—deploy purpose-built autonomous vehicles for ride-share services in the coming years. But full autonomy is likely years away, and when it finally arrives, carmakers that want to compete will have to supply large fleets to make on-demand services viable, whereas Uber will be able to introduce autonomous vehicles gradually.

It’s an advantage that could endure, as traditional and autonomous vehicles are expected to coexist for years. And in any case, when self-driving cars become a reality, the operating system for mobility that Kalanick is building—that perfect coordination of bits and atoms—will become more critical than ever. “You have to quantify humanity, human action in the physical world,” says Kalanick. It’s the kind of cold, beautiful and challenging problem that keeps Uber’s in-house logician motivated.

Additional reporting by Alan Ohnsman and Brian Solomon