QDaily's 2016 Company of the Year: Didi. Six of the deepest conflicts in China's economy converged inside this one company — and it could not navigate them.

This company lived through the most dramatic events any internet company could face.

No Chinese company had a year quite like Didi's in 2016.

It acquired its biggest rival, Uber China, absorbed 150 taxi companies, and finally won the Ministry of Transport's legalization of ride-hailing. Over those summer months, old and new rivals alike fell into line one after another, and Didi came to monopolize a market worth more than 100 billion yuan.

Then, starting in October, the situation reversed sharply. Led by Beijing, Shanghai, Guangzhou, and Shenzhen, more than a hundred cities issued local rules for managing ride-hailing — most of them harsh. After months of struggle, Didi could not change the final decisions in Beijing and Shanghai to expel drivers without local household registration. The ride-hailing revolution the internet had brought was legislated down to a premium taxi service.

All the deep conflicts in China's economy converged and erupted this year through Didi. That is why QDaily has chosen Didi as its Company of the Year.

Subsidies, expansion, acquisition, monopoly — pushed along by capital from every direction, Didi needed barely five years to grow into a full-blown giant: 6,000 employees, a valuation of $38 billion, and millions of drivers ferrying tens of millions of people every day.

This capital-driven path is the template for every Chinese internet business today, repeated at different scales across movie ticketing, group buying, food delivery, and travel. Along the way, Tencent, Alibaba, and Baidu have used capital and traffic to build oligopolies in one industry after another. Traffic and capital have eclipsed technology and innovation as the most important thing in building a new company.

Bound up with all of this is Uber's defeat. The fastest-growing and most aggressive Silicon Valley company since Facebook brought a new business model to the Chinese market — and in the end could not stay. Didi's absorption of Uber China was a clash between the different paths Chinese and American companies took in China, and a sign of intensifying local protectionism and a broader push to squeeze out foreign companies.

After beating its internet rivals, Didi upended the traditional business too. The taxi industry, organized by the government, had built a few hailing apps to fight back, but none came to anything. After more than a year of protests and strikes, the industry had essentially accepted defeat and begun negotiating how to fold itself into Didi's system.

Regulation changed all of that. Local government rules on vehicle model, engine displacement, residency, and carpooling cut straight at ride-hailing's ability to compete with taxis on price. We saw far too much of this incumbent-protecting industry regulation in 2016 — it reached entertainment and cross-border e-commerce, and the draft E-Commerce Law was published in the final month of the year.

The conflicts this regulation reflects go beyond protecting incumbent interests — they point to a much older problem: household registration and population mobility. Beijing and Shanghai together hold 46 million permanent residents. People come to big cities for more opportunity and a better life, and more of them come every year. The drivers running ride-hailing, the couriers delivering food, the workers raising the buildings — non-registered residents make up 40 percent of the population of Beijing and Shanghai, and their average age is far below that of registered locals. Tens of millions of people came to the big cities to make a living and, in doing so, built out the service economy. But at the same time the friction between urban residents and migrants grew sharper. Beijing and Shanghai's blanket expulsion of every non-locally registered driver is exactly that friction breaking into the open.

Uber China's defeat, and the growing isolation of the Chinese and global internet

Didi's roller-coaster year began on August 1. That afternoon, Didi released a statement confirming the acquisition of Uber China's brand, business, data, and all other assets. In an internal letter signed jointly by Didi CEO Cheng Wei and President Liu Qing, the two called it "a great rival, an epic showdown."

That night, Uber CEO Travis Kalanick held a video call with all of Uber's staff in China, announcing the merger of its China business with Didi and the plan for placing employees. An Uber China employee who was there said Kalanick looked in real pain and choked up as he spoke.

No foreign internet company has ever succeeded in China, but Uber went further than any of them.

That was no accident. Uber is essentially the template for the new generation of Silicon Valley startups. On one hand, it is a company intensely driven by data and disciplined about costs. A team of economists ran its surge-pricing system and driver-subsidy mechanisms, and it entered each city with a three-person team.

At the same time, Uber was one of the boldest Silicon Valley companies — it championed an overwork culture, invaded user privacy freely, and pushed authority down to each city manager. That boldness led People's Uber to put private cars onto the hailing platform even though Chinese law and regulation expressly forbade it, helping to set off the ride-hailing explosion in China.

Kalanick also placed special importance on the Chinese market: he spent a fifth of his time in China in 2015, and city managers across Uber China reported directly to him every week. Google, Microsoft, Amazon, and Apple would never have done this.

And still Uber failed.

"His weapons were excellent; his backend systems were highly developed and could support him from a distance," Cheng Wei said at a closed-door talk during this year's World Internet Conference in Wuzhen. "Back then, the only advantage we had was more people on the ground."

More people turned out to be a huge advantage. Didi's ground teams — at train stations, at airports, in Beijing, in Hangzhou, across more than 400 Chinese cities — taught drivers one by one how to use a smartphone and how to take jobs on Didi.

Didi was also more aggressive than Uber: backed by capital, it opened one business unit after another with no regard for cost. Even Express and Premier — two businesses that should technically share the same plumbing — were run as fully separate units, and any collaboration meant going through each unit twice over.

But that aggressive expansion added seven new businesses to the Didi platform in eighteen months, and the number of individual drivers on the platform grew from 400,000 to 15 million over the same span.

Investors Tencent and Alibaba helped considerably too. When Uber began to threaten Didi, Tencent banned the WeChat accounts of every Uber China city operation within a single day. Kalanick said at a small media roundtable in 2015 that he had met with a Tencent vice president, but it was no use. "You know what the American media still says?" he said. "That we're the ones using unfair tactics."

After more than a decade of learning from Silicon Valley, the competitive environments on the two sides of the Pacific have become entirely different. That divergence means American internet companies can take almost any market on Earth yet cannot get into China. And the Chinese internet companies that thrived in China's market — Tencent, Alibaba, and Baidu — have been just as unable to push their products into markets abroad.

Doing business in China as a foreign company keeps getting harder. In October 2016, when China's Ministry of Transport issued a draft regulation for managing ride-hailing, one clause required that the servers of hailing platforms sit on Chinese soil, so that data could not flow out.

Data security is being raised to the level of national strategy. The Cybersecurity Law imposes strict limits on data storage. Microsoft's and Amazon's cloud services each took years to go live through partnerships with local companies, and both were forced to make major technical changes to their China products — for a business that by nature depends on the free global flow of data.

At the same time, procurement of products from Oracle, Cisco, IBM, Microsoft, Apple, and nearly every other American tech company has come under tight restrictions in China. In April 2016, Apple's iTunes Movie Store and iBooks Store in China — both running for more than half a year — were shut down.

Policy also required foreign tech companies to open their source code and intellectual property from June 30, 2017. Microsoft, Intel, and IBM all objected outright.

The Chinese internet industry is growing more cut off from the rest of the world, and for the foreseeable future that isolation will only deepen.

The growth of emerging industries turns increasingly capital-driven

Beating Uber cost Didi dearly. Uber China was folded into Didi at a value of $7 billion, and Uber received a 17.7 percent stake in Didi. According to Uber operating data that later leaked, about $2 billion of that was eventually turned into cash.

Huge subsidy wars, and when you can't kill the other side, rush to buy it out at a high price — this runs through Didi's entire history.

Wang Gang, an early Didi investor, recalled the 2013–2014 subsidy war between Didi and Kuaidi, saying Tencent at first put up only one week of subsidies, but Alibaba spent more to seize a payments foothold, and neither side could stop. The burn kept growing, until at last, on Valentine's Day 2015, Didi and Kuaidi merged, brokered by investment funds and by Tencent and Alibaba.

The Didi–Uber subsidy war essentially ran the same script again.

Didi took money from Tencent, Alibaba, and venture funds of every kind; took $1 billion from Apple; and took investment and loans from big traditional players like HNA Group, China Merchants Bank, and China Life. Its long list of 90 investors even includes the names of state-capital vehicles like CICC, China Investment Corporation, and CITIC. A, B, C, D, E, F — round after round, until the alphabet itself ran out.

With all that capital behind it, Didi grew used to burning money. Liu Qing explained the huge paper losses to investors last year: "Without burning money, we couldn't have gotten this far."

Over the eighteen months of head-to-head fighting with Uber China, Didi rolled out one business unit after another — Express, Designated Driver, Hitch, Bus, Car Rental, Enterprise, Minibus — each with its own product, technology, operations, and marketing teams. Didi grew fast, and just as fast became a bloated big company, its headcount swelling from more than 1,400 in 2014 to over 6,000.

Didi won in the end, but it faces a brutally hard internal overhaul. As Cheng Wei said at that closed-door talk, after a round of rapid growth you always have to go back and make up the lessons you skipped: scale can be stacked up by burning money, but the team and the technology can't simply be bought.

In August 2016, after Didi acquired Uber China, the mutual fund BlackRock, which held a large block of Didi shares, raised Didi's projected 2017 revenue growth in a report, from 90 percent to 150 percent. It looked as if the capital-driven cycle of burning money, expanding, and acquiring had finally reached the point where monopoly turns into big profits.

But data tracked by several third-party firms showed that even before the crackdown on out-of-town drivers arrived, the number of people using Didi each day had already begun to drop sharply, because subsidies had been cut — a 27 percent fall in three months, to be precise. At the end of October, with drivers leaving and policy uncertain, Didi restored its subsidies.

Mogujie and Meilishuo, Meituan and Dianping, 58.com and Ganji — these next-generation internet companies all followed a path much like Didi's, and all began merging in 2015. Like Didi, none of them has found a profit model that can justify its valuation. Now capital has started burning money on shared bikes.

Monopoly and the smothering of innovation have become normal in China's economy

In September 2016, China's Ministry of Commerce announced an antitrust review of Didi's acquisition of Uber China, saying it had summoned Didi for talks twice.

Didi argued that neither company's earnings had met the Ministry of Commerce's threshold for triggering review. But the monopoly had already formed. According to CNIT's first-quarter 2016 data, after the Didi–Uber merger Didi held more than 90 percent of the ride-hailing market. In the cheaper, highest-volume Express segment, Didi's share was approaching 100 percent.

As in other countries, government antitrust reviews here usually achieve little: by the time regulators wake up, the monopoly is already set in stone. The early reviews of Microsoft and Google are clear examples. China is no different.

We fear monopoly because a company holding such a concentration of resources can strangle any competition that might exist.

Didi's own success owed much to monopoly. Tencent holds China's two most important channels for social distribution, WeChat and QQ. After pouring billions into Didi, Tencent did not hesitate to turn its social monopoly into a weapon. First WeChat blocked Kuaidi, lifting the ban only after the Didi–Kuaidi merger. When Uber China, which leaned heavily on social distribution, began to threaten Didi, all of its accounts were blocked on WeChat too. By 2016, even Yidao — a company with a far smaller share than Didi — was blocked on WeChat. Every time, WeChat's stated reason was protecting the platform environment.

Baidu, Weibo, Taobao, WeChat — every product in China that calls itself a platform has shown how comfortably it blocks potential competitors.

The capital-driven subsidy-burn model deepens the effects of monopoly, because the so-called "traffic" is held by a handful of big companies.

Kuaidi offered premium car service before Didi, and built a data platform to analyze trip data before Didi. Yidao offered airport pickups before Didi and pioneered a string of related innovations. Uber put private cars on its platform before Didi, invented surge pricing, and built a carpool system that improved efficiency before Didi did. Didi caught up quickly on every feature — but, leaning on capital and traffic, it never needed to invent anything first.

For companies without traffic and capital, challenging the leader through innovation in technology or business model has become nearly impossible.

Disruption and counterattack between the old business and the new

On September 1, Didi announced strategic partnerships with close to fifty taxi companies across more than ten cities, including Beijing, Shanghai, Guangzhou, Shenzhen, and Hangzhou. In simple terms, the partnership used an "intelligent dispatch" system to connect taxi and ride-hailing orders and backend systems — a passenger's Express order could now be routed to a taxi driver.

Within two months, the number of taxi companies that had joined this shared system had grown to 150.

The taxi companies seemed to have no choice; the efficiency gap between the two sides was simply too wide. Didi made its name in a heavy snowstorm in Beijing in 2012. When the weather turns bad, hailing a cab gets very hard. On that day, Didi — which let people call a taxi by phone — took off. People discovered that getting a cab need not mean standing on the curb waiting on luck; it could become something you could estimate and plan for. Users drawn in by subsidies stayed in the end because it was simply more convenient.

Taxi drivers have long complained of low pay, because the taxi operating model is extraordinarily inefficient. A driver must hand over a daily "franchise fee," and only what is left after covering it is his to keep. But a cab cruising the streets for fares takes in whatever luck brings.

The "franchise fee" has been debated endlessly and never settled, because the meter inside the cab is the taxi company's only means of controlling its drivers. Switch to a commission model and drivers would have every reason to shut off the meter and route around the company's control. The fixed "franchise fee" is the solution for a problem no one could solve. Once the matching of drivers and passengers moved onto the phone, Didi had no need to install a meter in a private car — every job a driver took was logged automatically.

Making every private car a potential taxi let Didi marshal a pool of drivers far larger than the taxi fleet. Carpool technology cut each driver's per-trip cost; in the most mature US carpool markets, San Francisco and New York reached a ceiling of $5 for in-city carpooling without subsidies.

Surge pricing, in turn, mobilized part-time drivers: when passengers are too many and drivers too few, higher fares pull more drivers onto the road. Ride-hailing is a new model far more responsive to supply and demand than the traditional taxi business.

In barely three years, Didi destroyed the taxi business. Everyone and every company in it was replaced. There were countless incidents of taxi drivers and Didi drivers converging on headquarters to protest. After Express and Premier launched in 2015, taxi drivers' resentment of Didi grew sharper still. But the old business had no power to fight back; now more than 80 percent of taxi drivers open the Didi app while cruising, listening for orders dispatched from the Didi platform.

Then regulation stepped in, and the great majority of cities placed various restrictions on vehicle model, driver identity, and operating permits.

Local governments and taxi companies also saw the opening that smartphones created. Local governments in Guangdong, Zhejiang, and Guizhou began backing homegrown phone-hailing apps.

Shouqi Ride, the state-backed car-hailing service that had just announced a new funding round, unveiled an expansion plan in the final week of December — opening a local-brand franchise system, aimed mainly at companies backed by local governments.

Regulation grows ever stricter — and what it cares about most, oddly, is not fairness

On October 8, the first-tier cities of Beijing, Shanghai, Shenzhen, and Guangzhou published, one after another, draft or public-comment versions of their detailed rules for managing ride-hailing. All four cities raised the bar for a premium car through requirements on household registration, license plate, and vehicle type.

Didi's worst fear had come true. Under the drafts published that day, Beijing and Shanghai — both cities with strict household-registration management — required drivers to hold local registration, a requirement that directly excluded the great majority of current premium-car drivers.

That day, Didi published an open statement that spelled out the consequences of strict ride-hailing policy: "Millions of ride-hailing drivers may lose their current jobs and income, which also means millions of families will lose a vital source of earnings. Large numbers of ride-hailing drivers will face the blow of unemployment, become idle members of society, and have to look for new work — which could trigger mass crises and social instability."

A few hours later, the statement had been deleted in most places, and the few copies that remained had this passage cut.

For an enormous industry touching the daily lives of hundreds of millions of people, regulation is necessary.

But this round of regulation ignored many practical questions — for instance, if an accident happens during a carpool, how is responsibility shared among the platform, the driver, and the fellow passenger? Even concrete matters, such as exactly how to test and certify drivers and where to apply to change an operating permit, were left unspecified by many cities.

What was spelled out clearly was mostly aimed at limiting ride-hailing's growth.

Local governments showed no shortage of creativity in raising the bar for ride-hailing vehicles through rules on price, engine displacement, wheelbase, and body length. Most cities pushed the threshold up to expensive, fuel-hungry Class B sedans, making sure a ride-hailing car would cost more to run than a taxi.

Most cities also capped the number of carpool trips per day, anywhere from two to eight. This restriction — which does nothing for the environment or for congestion — made sure ride-hailing could not use technology to lower the per-trip operating cost of a vehicle.

Many industries have been through similar regulation.

Internet television was barred last year from aggregating content from multiple video sites. A rule that hurts the consumer experience and does nothing to prevent industry monopoly — but does help the traditional broadcasters.

Cross-border e-commerce went through the new cross-border e-commerce tax regime and the retail import product catalogue, introduced one after the other. The categories hit hardest were mother-and-infant and household goods, which now faced an added tax of 11.9 percent. Low-priced cosmetics that had been tax-free now faced a rate of 32.9 percent.

Whether certain goods could still be imported and sold the old way also changed, because of an import product catalogue jointly issued by the Ministry of Finance and other departments. Although by year-end cross-border e-commerce got word of a one-year grace period for the new rules, more than half a year of upheaval had already cooled investment and consumer spending in the field.

The whole e-commerce industry faces stricter oversight. The Draft E-Commerce Law, three years in the making, was formally submitted for review by the Standing Committee of the Twelfth National People's Congress in December 2016. The draft sets rules for every link in the e-commerce chain. One proposal is that, to address the information asymmetry in e-commerce, a third-party platform should first compensate consumers directly and then recover the cost from the sellers on its platform. This is something platforms had been handling through their own ratings systems.

The conflict more sensitive than commercial competition: migrants

"Look — those are all Didi cars. When the plate starts with Ji, Yu, Lu, or E, we can spot it right away." On Christmas Eve 2016, Express driver Xu, behind the wheel, pointed at the out-of-province cars passing us on the road.

She had been working in a factory in Wuhan. The factory closed six months earlier, so she started driving for Didi. At the time, the Ministry of Transport had just issued rules confirming the legal status of ride-hailing and setting out the requirements for an operating permit.

Driver Xu was not someone who tracked the news closely and chased subsidies from platform to platform. In early October, she drove the family Jetta from Wuhan, Hubei, all the way to Beijing, becoming one of the 600,000 out-of-town drivers working in Beijing on the Didi platform. By then Didi had essentially paused subsidies, but Xu was already fairly content with her income — she said that in Wuhan she had to drive ten hours a day to make 200 yuan, while in Beijing she could make two or three hundred without trying hard.

Unemployment and the chance to earn a bit more in a big city — these were the main reasons drivers registered elsewhere came to Beijing. Didi's existence gave these several million people a better way to make a living.

In earlier interviews we did with dozens of drivers, most of the out-of-town drivers, even after subsidies were taken away, still felt that driving for Didi was a good job. Some had driven freight trucks before, often going a whole month without a single day off; some had worked at a bathhouse, owed months of wages they could never collect; some had driven for a rental-car company in Beijing, working two shifts and sixteen hours a day for 6,000 yuan a month.

The decline of manufacturing has been discussed for years. Some factories in Guangdong that once ran constant overtime began holding to a rhythm of idling from New Year's Day through to Chinese New Year for lack of orders, as early as 2013. The China Labour Statistical Yearbook 2015 reported 4.864 million unemployed across 17 provinces targeted for capacity cuts.

But these people's plans for making a living conflict with the big cities' plans. Beijing has set an explicit target of cutting the permanent population of its six central districts by roughly 15 percentage points by 2020. Medical, education, and government-service institutions have been moved out, step by step, from the city center to outer counties and even to Hebei and Tianjin.

On December 21, Beijing and Shanghai announced the final version of their ride-hailing rules — put in drivers' own words, "Beijing plates for Beijing people" and "Shanghai plates for Shanghai people."

Not just ride-hailing: manufacturing, wholesale markets, warehousing and logistics bases, small-commodity markets, and other industries and areas with concentrations of out-of-town residents are all priorities in Beijing's population "relief" plan.

"On the Beijing-registration question, the first round of public comment asked me about it. My view is that it isn't only ride-hailing that needs a registration rule — in fact many industries restrict the work to Beijing-registered residents," said Wei Dong, CEO of Shouqi Ride, who took part in the consultation on Beijing's ride-hailing rules, at a media briefing at year-end. "It isn't ride-hailing alone restricting Beijing registration. Many industries are doing the same. In a sense, all of these baseline thresholds serve one purpose: controlling the total urban population. Ride-hailing is just falling in line with that."

Big cities are too congested; more towns and smaller cities worry about jobs. China's overall GDP growth has held "precisely" at 6.7 percent for three straight quarters. The conflicts will only grow more severe.

Judging by how ride-hailing was handled, the government still prefers the blunt, across-the-board approach: cap the jobs directly to stop people from moving. After ride-hailing, which industry will be restricted next?

Driver Xu did not learn of Beijing's policy restrictions until Christmas Eve. The Wuhan drivers in her group chat had been too busy driving to hear about the new rules issued three days earlier. She said she would go and look into it when she got back. "They probably won't stop us from driving, right?" From start to finish, Xu never once complained about earning too little.

That is how 2016 ended for the more than a million Didi drivers registered outside Beijing and Shanghai.

Cover image from Visualhunt.com