iQiyi Is Only Seven Years Old, and It Has Already Lived Through Nearly Every Big Change in China's Internet Business

Burning cash became a habit, but no industry has found its way out by burning cash.

In late March 2010, the video site "Qiyi" went live. It was Baidu's newly formed video subsidiary, and after a few weeks of trial operation it began advertising tens of thousands of hours of free, licensed, high-definition video.

In a country long known for piracy, where licensed content was thought to have no future, a copyright war worth more than 10 billion yuan got underway. Ku6 said it would put up 300 million yuan, Sohu said it would invest US$10 million, and 56.com said it was spending a third of its monthly revenue on copyright.

A year later Qiyi was renamed iQiyi, but the fight was not over. Baidu's grip on search could not help it beat every rival; some video sites were bought up, and the ones left standing raised even more money to keep going.

Just as the video sites' copyright war was starting, Meituan took US$12 million from Sequoia Capital and, along with the thousand-odd companies that followed, kicked off the group-buying war. Within a short time the five biggest of them had raised US$500 million between them.

Most Chinese internet companies had missed the bursting of the Nasdaq dot-com bubble in 2000, and had never seen tens or hundreds of millions of dollars poured into startups that made no money.

But before long, outspending the competition became the growth model Chinese internet companies knew best. Ride-hailing, O2O, food delivery, one after another. Now even bike-sharing companies can pull together hundreds of millions of dollars of venture capital in a single year.

The changes were not only about burning cash. In 2010, the year iQiyi was founded, the Chinese-language internet was a completely different world.

Back then most people still did not go online on their phones. Nokia was still the world's best-selling phone maker, and it sold more than three times as many handsets in China as the iPhone. OPPO was still an MP3 player company, Xiaomi had not been founded yet, and Lei Jun was hanging around forums discussing beta builds of MIUI with users.

Internet companies at the time still followed Silicon Valley's lead. In March 2010, Google said it was leaving the Chinese market, citing hacking attacks and censorship. Disappointed users laid flowers outside Google's Beijing headquarters.

A little earlier, Kai-Fu Lee, who had run Google China, left to start the incubator Sinovation Ventures, hoping to bring the model of YC, Silicon Valley's biggest incubator, to China. Innovation was something everyone wanted to talk about.

Back then, founders of new companies were routinely asked, "What will you do if Tencent copies you?" Now, when Didi swallowed Uber China, what people said was that it was "the first company to take investment from all three of BAT."

Seven years is not long, but China's internet business has changed beyond recognition. Burning cash became the basic way to grow; capital-driven mergers and monopolies happened in every industry; startups went from challenging the three giants to competing for their money; the Strategic Emerging Industries Board, once seen as the way out for internet companies, was killed off. And all the while regulation kept tightening.

iQiyi has been caught up, to some degree, in every one of those changes. After going the full circle, it is still losing money badly and struggling to turn a profit — like Didi, Meituan, 58.com, and almost every big Chinese internet company that grew up over these seven years.

iQiyi's history is close to a miniature of the whole Chinese-language internet: these companies start a new business full of hope, meaning to overturn the old one, but the competition shifts quickly from innovation to money, and money then pushes them to end the war by merging. To this day, not one of them has used profit to prove it deserves the 10-billion-yuan valuation it carries on paper.

From Video to Music: Internet Companies Cared About Copyright for the First Time

China's video industry began with piracy.

Youku, Tudou, Ku6 — the earliest video sites all got their first batch of content by encouraging users to upload and share. What followed was an endless shouting match over copyright, with the video sites suing each other constantly.

The conflict peaked in 2009. That September, Sohu Video, Joy.cn, and Voole formed the "China Online Video Anti-Piracy Alliance" and sued Youku and Xunlei.

A month later the "Anti-Piracy Alliance" and Xunlei both turned up at the same hotel in Shenzhen, holding press conferences in meeting rooms one wall apart to denounce each other. Xunlei hung a banner, red characters on white, outside Sohu's door: "Exposing the Sohu-Voole Sham Alliance." Tempers ran high, there was a scuffle, and in the end the police had to be called to settle it.

The farce ended in 2010. First, the State Administration of Radio, Film and Television (SARFT) issued a document saying it would tighten control over film and TV content on video sites. According to The New York Times, video sites in Beijing had entered a "fairly systematic state of regulation."

The video sites began dumping pirated content as if to prove their loyalty. In 2010, Ku6 deleted unlicensed European and American films and shows, said it would spend 300 million yuan on licensed content, and joined Sohu, which put up US$10 million, to set up a "copyright acquisition fund"; Youku and Tudou opened their content libraries to each other and deleted everything pirated; 56.com said it was spending a third of its monthly revenue on copyright.

This was the setting iQiyi arrived in, and it said from the first day that it would run a licensed-content business. The press release on its launch day made a point of its deals with rights holders including China Film Group, Huayi Brothers, Zhejiang TV, Hunan TV, and Beijing TV, saying its "content purchases had reached tens of thousands of hours."

It was close to a gamble, meant to set the company apart from the UGC (user-generated content) mainstream of the time. Gong Yu said in an interview: "If we do whatever everyone else does, we will never be number one." And: "Other companies have 500 people, 200 of them on long-form video; our company has more than 300 people, all on long-form video, so we have a chance to do it better."

It was not only video. Online music started clearing out piracy that year too. In October 2010, 564 music sites were reviewed by the government and ordered to delete banned and pirated content; by December, dozens of music sites had been shut down, among them Haoting.com, qq163, Knight Music, Tingge.com, and Cocoxi Music.

What pushed these companies to start talking about licensed content was, above all, direct commercial interest.

Copyright worked like a weapon, letting internet companies dig a moat and squeeze rivals. The "rightness" of copyright itself did not seem to matter much — China's internet in the 2000s was, after all, built on piracy. In 2004, the piracy rate for Microsoft's Windows in China was 90 percent; by 2006 it had improved to 82 percent.

For video sites, long-form video fetched higher ad prices. The story everybody repeated was that Hulu, the American service that carried only licensed content, turned a profit in its first year of operation and took 33 percent of the U.S. video advertising market with 1 percent of video traffic, while YouTube, founded three years before Hulu, was still deep in the red.

"I ran the numbers and found that a licensed-content model can make money now, so that is why we are doing it," Gong Yu said just after the site launched. YouTube, he noted, could turn only 4 percent of its traffic into advertising, while the other 96 percent of its bandwidth costs produced no direct value.

So the fight changed from infringing on each other to bidding against each other for rights. That shift cleared out a batch of pirate sites that were short of money to begin with, and it opened a new round of ruinous competition.

Capital Became the Weapon That Decided Who Won

iQiyi's story rested on using licensed content to draw users and then win over advertisers. It and its rivals soon had a new problem: the price of copyright kept climbing.

In a year or two, video rights prices "rose more than tenfold," and the online rights to a hit TV series went from one or two thousand yuan to 50,000 yuan.

That was only the start. Youku and Tudou went public one after the other, raising the cost of entry to the video business further, and prices climbed without limit in a seller's market.

By 2011 the average price of a single TV episode reached 700,000 to 800,000 yuan, peaking at 1 million yuan. That year Youku lost 172 million yuan, Tudou lost 511 million yuan, and iQiyi lost 500 million yuan.

The fight over online music rights followed much the same path, beginning around 2012 and 2013; over the next few years music rights went from tens of thousands of yuan to millions.

Sites without deep pockets were hit first. Around 2007 China had more than 300 video sites; by 2011 about 20 were left. Ku6, which had promised to pour money into buying rights, announced that same year that it was giving up on long-form video.

The money the video industry burned was nothing next to other new business models.

The earliest and most extreme case was group buying. Groupon, the American pioneer, went public in 2011, and China's group-buying industry peaked at the same time. By Analysys's count at the time, the number of group-buying sites reached 5,000 in a little over a year.

Their models were much alike: start with restaurants, negotiate a discount price with the merchant, then let users order through the site, which took a commission and a cut, plus a sizeable pool of customer money sitting in its accounts.

The business looked easy to enter, but it leaned heavily on venture capital to pay for marketing and subsidy wars. Between April and July 2011 alone, five group-buying sites — Meituan, Dianping, 55tuan, Manzuo, and F-tuan — raised close to US$500 million between them.

That money went into frantic subsidies for users and merchants, and the "war of a thousand group-buying sites" began. As more sites appeared, merchants ran short and started demanding higher commissions. Lashou, then the site with the largest market share, used its war chest to go to zero commission to grab merchants.

As in video, a business model propped up by spending could not survive without a steady supply of capital, and the small companies were washed out. After Lashou's U.S. listing failed in 2011, its market share slid to third in the industry, and it faded away along with many other badly run group-buying sites that had run out of cash. By early 2013, more than 80 percent of group-buying companies had folded.

Burning cash became a habit from then on. Every time capital found a new industry, it spent round after round on a loud cash-burning game, trading money for user growth.

Dianping and Meituan survived the group-buying war and then competed in O2O the same way; in 2012 Didi and Kuaidi started hailing taxis online; in 2013 Ele.me, Meituan, and Nuomi started food delivery; in 2016 Mobike and ofo brought bike-sharing.

Group-buying sites mostly subsidized merchants, to get more restaurants to offer discounts. Food delivery and the various at-home service companies subsidized merchants and users at the same time — every day you got coupons of one size or another. Ride-hailing began the same way, using money to attract drivers and riders: every trip came with a bonus for the driver and a discount for the passenger.

By the time bike-sharing arrived, the rides were essentially free.

One result of burning cash is that the company providing the service never really takes charge. Video sites are the clearest case: the ones who gained from the content business were the content makers, not the sites — in theory, whoever produces the content need not depend on any single platform.

In September 2015, iQiyi's chief content officer Ma Dong left to start his own company. His new firm, Miwei Media, took the U Can U Bibi team with it, and the show went from being iQiyi's "only child" to its "partner."

The same held in food delivery and O2O at the height of the competition: restaurants and tradespeople had no reason to sell through one platform only, since the platforms were not that different.

Burning cash killed the companies that could not raise money, but it did not end the competition.

Capital-Driven Mergers Became the Last Step, and Behind Them Were Those Three Companies

Monopoly did not come easily. The survivors began absorbing their rivals to improve the odds of getting one, and video sites went through this early.

In March 2012, Tudou and Youku announced a merger in an all-stock deal. The next year Baidu bought PPS for US$370 million in cash and folded it into iQiyi, joining up their membership and advertising businesses. What Baidu wanted was a bigger user base, more channels, and more ad inventory. iQiyi said the merger had "built China's biggest video advertising platform."

The merger lifted market share directly. Youku Tudou and iQiyi took first and second place in the online video advertising market, with Sohu Video third; together the three held more than half the market.

In 2014, Tencent Video locked up exclusive video playback through WeChat, and its share quickly passed iQiyi's to reach second place.

In 2015, Alibaba paid about US$4.5 billion for Youku Tudou. By then the video business had become a contest among Tencent, Alibaba, and Baidu alone; Youku Tudou, Tencent, and iQiyi held 60 percent of the market between them.

But video was not where mergers ran hottest.

Major mergers in China's internet industry over the past three years

2015.2

Didi Dache merges with Kuaidi Dache

2015.4

58.com merges with Ganji.com

2015.10

Dianping merges with Meituan

2015.10

Ctrip merges with Qunar

2016.6

Meilishuo merges with Mogujie and Taoshijie

2016.8

Didi merges with Uber

Mergers were the inevitable end of a cash-burning game that could no longer go on. For video companies, the cost of rights, of traffic, and of acquiring users kept rising, and a transfusion from a big company was the only way to guarantee the spending could continue.

The mergers in ride-hailing and O2O worked the same way: by the end, the two or three companies left made no real difference to users, and buying each other out was the only route to a monopoly.

Worse, there was barely an industry left that could stay outside the control of Tencent, Alibaba, and Baidu. Behind every merger was the shadow of one of the three. Didi, Meituan, Dianping — they all took the same path, trying to corner their market on long-term life support from the three giants, and in the end nothing much came of it.

The three are skilled at financial engineering, and they can feed traffic and users to the companies they back through different channels.

This is what happens when innovation disappears. Baidu's search, Tencent's social network, and Alibaba's e-commerce all built walls as they grew, and they throw off steady profits that no other company can challenge. The new industries have not managed that. Ride-hailing, at-home services, bike-sharing — even after merging their way to a monopoly, none of them can generate its own profit, and all still depend on the big companies to keep them supplied.

Before mergers came into fashion, the question startups asked each other was "What if Tencent copies you?" Now the question is "How do I get Tencent to invest in me?"

When the Main Business Does Not Make Money, Every Company Starts Talking About "Ecosystems"

After all the merging and consolidating, the main business still was not making money, so the new story was given another name: "ecosystem."

To get its content into more places, iQiyi picked up the word too.

In 2014, Xiaomi invested US$300 million in iQiyi, a deal described as "building a more complete and vital video ecosystem." iQiyi brought the video content; Xiaomi said its 85 million MIUI users could lift iQiyi's share on mobile. Xiaomi TVs and phones became iQiyi's priority distribution channels, and iQiyi's shows and films began appearing on Xiaomi televisions, set-top boxes, and handsets.

Hardware as the entry point, with the profit coming from the internet services that follow — that was the story Xiaomi told first. The app store built into every smartphone was a revenue channel, with game and app developers paying for better placement. Xiaomi even, for a while, put ads into the system software on every user's phone.

In 2014, Xiaomi started investing in "ecosystem-chain companies." The so-called ecosystem chain was really a way of making products by investing in and incubating them, tying more hardware to Xiaomi.

iQiyi joined in this kind of "ecosystem building" too. In September 2016 it invested 150 million yuan for a 5 percent stake in Coocaa, a Skyworth subsidiary, and Skyworth televisions came with iQiyi's VIP membership preloaded.

Then more businesses arrived. In April 2015, five years after its founding, iQiyi said it would go into video commerce, and the iQiyi Mall opened.

Add the VR, livestreaming, games, animation, and education mentioned later, and iQiyi recently laid out its ecosystem story in full, built around IP and billed as a "pan-entertainment ecosystem": "In our business model we correspond to Disneyland or Universal Studios. We have video, we have online games, and we have the iQiyi Store, though it is small for now. Together they form a healthy ecosystem."

Alibaba pulled its film, television, music, and news assets together into a "big entertainment and media ecosystem."

But the company that took the word "ecosystem" furthest was LeEco. After working with Foxconn to make the LeEco Super TV and the LeEco set-top box, LeEco said it was building a whole-industry-chain ecosystem of "platform plus content plus devices plus applications."

In 2014, LeEco decided to build cars as well, unveiling its "SEE Plan" and promising a super electric car and an internet-connected car ecosystem.

The LeEco phone of the same year turned the company's ecosystem talk into a vocabulary no one could parse, such as "ecosystem chemical reaction." LeEco even changed its name to "LeEco Ecosystem" and put it on the wall outside the Wukesong Arena in Beijing.

"Ecosystem" was a basket that could hold anything, letting internet companies pile unrelated businesses together and claim they were connected. This freshly wrapped business story came, at bottom, from the fact that the main business made no money.

Xiaomi was watching its phone sales fall, out of China's top five by shipments, and its original logic — sell phones cheap, make the money later on internet services — had come apart.

LeEco's cash crisis last year set off a chain of damage: the e-commerce business was abandoned, staff were laid off, buildings were sold and wages went unpaid, and the investment project in Russia was shut down. Jia Yueting admitted that phones, cars, televisions and the other businesses meant to be key links in the ecosystem had run into trouble because "running blindfolded" left the need for money running ahead of the ability to raise it.

Xiaomi, which used to talk about its ecosystem chain, has lately moved the conversation to "becoming a MUJI," and back to the business of making money selling hardware.

From Free to Paid, a New Story Opens Up

Profit still had not arrived, so iQiyi found itself a new story.

Spending on rights kept growing. From 2013 to 2015, the "content costs" in Baidu's financial reports grew by more than 90 percent every quarter, and losses passed 2 billion yuan in 2015, three times what they had been two years earlier.

Original content became the new weapon. Youku Tudou and iQiyi both said more than once that they wanted to be — or to beat — Netflix, which charges users a monthly fee directly for content it makes itself.

iQiyi got a taste of the benefits. After paying users could watch all of The Lost Tomb in 2015, iQiyi, which had signed up only 5 million paying users in its first four years, grew that to 20 million in the year from 2015 to now.

Market share rose as well. By monthly active users on mobile, iQiyi still ranked third in 2014; thanks to its own dramas, it now ranks first.

Gong Yu's plan is that once the model matures over the next three or four years, advertising, memberships, and games and other content will each supply about a third of revenue, filling iQiyi's pockets together.

For more than a decade, China's internet has been one enormous experiment in free. You are used to video, music, and news costing nothing; a news paywall, for one, has never really been tried here. Internet companies did not charge users, and covered their costs and made their profits some other way, usually advertising.

But the new business models showed no sign of profit, and now, after working through a whole range of models, internet companies have come back to the simplest one: charging the user.

Figures from iResearch say people have started paying for online music too. In 2015, China's paid online music market was worth 1.05 billion yuan, up 121.8 percent from the year before.

Weibo's revenue in 2013 was only 66 million yuan, less than one yuan per user. After it moved into livestreaming and e-commerce, the figure reached 1.18 billion yuan in 2016.

Another new phrase is the "knowledge economy." Zhihu, Fenda, and Dedao all began in 2016 letting users pay for content made on their platforms.

But this trend is only beginning, and the revenue from paying users comes nowhere near covering the money burned. What iQiyi's memberships bring in is still very limited, and Gong Yu's target remains "turning 10 percent of users into paying users."

Not one company in the "knowledge economy" business has disclosed how well it retains paying users either, and with all the promotions and subsidies mixed in, only the service providers themselves know how much revenue they are really making.

The End of the Strategic Emerging Industries Board Left Internet Companies Short of a Finish Line

Building an ecosystem and starting new businesses both cost money, and to solve that, internet companies generally wanted to go public.

What showed Baidu a new opening was the Strategic Emerging Industries Board that the Shanghai Stock Exchange began preparing in 2015, which set no profitability requirement.

Before that, listing on China's A-share market required approval from the China Securities Regulatory Commission, and the process was complicated and carried a profit threshold. The A-share main board required an applicant's cumulative net profit over the most recent three fiscal years to top 30 million yuan; the ChiNext board required at least one profitable year, with net profit of no less than 5 million yuan and revenue of no less than 50 million yuan.

Far more companies wanted to list than the narrow channel could handle, a serious mismatch between supply and demand. Even companies that cleared the profit bar had to queue for years to have their applications processed.

That created problems of its own in the A-share market: a high bar and strict supervision limited its size, which made it easy for speculators to manipulate and produced still greater disorder.

So loss-making internet companies had little choice. More than a decade earlier, Baidu had listed in the U.S. and Tencent in Hong Kong.

The Strategic Emerging Industries Board offered a new route, and companies listed abroad started thinking about coming home. By March 2016, 20 companies had announced plans to go private: Qihoo 360, Dangdang, Jumei, Youku Tudou, Momo.

iQiyi was among them. In February 2016, Robin Li and iQiyi CEO Gong Yu, on behalf of a buyer consortium, offered Baidu US$2.8 billion for 80 percent of iQiyi.

iQiyi was losing about 2 billion yuan a year, but on measures such as market value, net profit, and cash flow, it met the requirements of the Strategic Emerging Industries Board.

For the investors who had taken control of these new giants, the money burned still had to be earned back eventually, on the public markets.

It all looked settled. Shanghai had even picked three internet companies and one state-owned enterprise — Dianping, iQiyi, COMAC, and an internet finance company — as the first batch to list on the Shanghai Stock Exchange's Strategic Emerging Industries Board.

Then the plan collapsed. In March 2016, the draft outline of the 13th Five-Year Plan was revised in 57 places, and the line about "establishing a strategic emerging industries board" was deleted; it would not be set up. This had to do with a turbulent stock market: from June 2015, Chinese stocks had gone into a crash no one could control.

iQiyi's privatization was called off in July 2016. With the listing delayed, raising money cost more. This February, iQiyi completed a US$1.53 billion debt-to-equity subscription, of which Baidu took US$300 million. Subscribers usually attach hard conditions, such as a profit target or an IPO deadline; if iQiyi cannot deliver, it has to give the money back, and it turns into real debt.

The latest word is that iQiyi may list in Hong Kong. Ant Group executives also made a trip to Silicon Valley last year for a round of meetings and to size up potential investments, and a U.S. IPO is possible too.

Since Last Year, Every Internet Company Has Started Making "Content"

Among the new businesses iQiyi has tried, there is one few people notice: iQiyi Toutiao. If it were not for the iQiyi name and the green color scheme, you would never guess it had anything to do with video.

It does not, in fact. It is a pure news app, much like Toutiao, pulling in articles from media outlets and ranking them by click rate, interest, and other factors. The only feature that connects it to iQiyi is that you can follow a celebrity's news page.

This matches a new fashion: content has become a hot business.

The source and the template is Toutiao, founded in 2012 without a single news editor. Two years later it raised US$100 million, and it now has more than 500 million users.

The traditional news portals started "Toutiao-fying" themselves. Tencent, NetEase, Sina, and Sohu all added content from other outlets on top of their own, and used interest-based recommendation by artificial intelligence to display the news.

News sites and browsers were not the only things going "Toutiao." The "Taobao Toutiao" section on Taobao's home page also aggregates content from third-party partners to drive shopping. JD.com and Dianping both have their own Toutiao sections; even China Merchants Bank's Palm Life app added a "Xiaozhao" Toutiao feed.

This, too, turned into a money fight among the big companies.

Alibaba's UC Browser announced last December that it would put up a 1-billion-yuan fund to encourage creators making content on its platform, and Alipay added a news section all but indistinguishable from Toutiao's.

Tencent said it would step up its financial support for content entrepreneurs, with 1 billion yuan in cash subsidies aimed at creators of original and short-video work; Baidu's Baijiahao said it would share out a cumulative 10 billion yuan with content producers in 2017.

The heart of the content business is that a sensational headline lifts user engagement and pulls in more traffic, even if it has nothing to do with revenue.

Regulation Keeps Tightening, but What It Cares About Most Is Not Fairness

Every new industry, when it is born, probably talks at some point about "disrupting the traditional industry."

iQiyi too once said online video would "disrupt television." But more often the last word belongs to policy, even when the new industry has replaced the old.

Video sites are regulated by the government, and being ordered to take a show down is nothing unusual.

After the ban on Korean content in July 2016, every Korean film and TV program was pulled from online platforms. Sister Is So Hungry (《姐姐好饿》), a variety show iQiyi launched in November, was accused by SARFT of carrying unacceptable content and taken down; it came back only after being re-edited.

In December, every web movie, web series, and web variety show on iQiyi's platform had to fill in a registration form for key online original programs, under a filing and registration system, with iQiyi stamping them all and submitting them to the provincial bureau for the record.

But what video faces is nothing next to what hit ride-hailing.

Didi used capital to outlast and swallow every rival. In three years it wrecked the taxi industry: any private car could become a taxi, and Didi could mobilize a pool of vehicles far larger than the number of licensed taxi drivers.

Taxis themselves were remade too. A car cruising the street for fares could pick up passengers more efficiently, instead of parking somewhere and waiting for customers to find it. More than 80 percent of taxi drivers now keep the Didi app open as they cruise, listening for orders the platform sends them.

But direct regulation can knock all of that down.

At the end of last year, cities across the country found creative ways to raise the bar for ride-hailing vehicles, through rules on price, engine displacement, wheelbase, and length. Most set the bar at expensive, thirsty mid-size (B-class) cars, making sure a ride-hailing car cost more to run than a taxi.

Most cities also capped the number of carpool trips per day, at between two and eight. This limit does nothing for the environment or for congestion, but it does make sure ride-hailing cannot use technology to lower the cost of running each vehicle per trip.

There are plenty of other industries whose creativity regulation has strangled.

Internet-connected televisions were barred in 2015 from aggregating content from more than one video site. That rule hurt the viewing experience and did nothing to prevent monopolies in the industry, but it did help the traditional broadcasters.

Cross-border e-commerce went through a new tax regime and then a new list of permitted retail imports. Mother-and-baby goods and daily necessities were hit hardest, facing an extra 11.9 percent in tax. Although the new rules came with a year's grace period, after more than six months of upheaval the sector had all but vanished.

Regulators also react faster now. Hailing a cab by phone had been around for three years before policy stepped in last year.

And now regulation hangs over bike-sharing, the newest industry of all. From the day Mobike began operating to the day local governments started working out how to manage bike-sharing and drafting rules for it, only about 10 months passed.

Internet companies have made little progress in finding a way out. The people who regulate them get more practiced by the day.

Cover image from Pixabay