Episode 3: Failure

Episode 3: Failure


In Episode 2 I talked about the euphoria among American venture capitalists for all-things-China. As I was writing it, I realized that the investors’ interest in China was a little ironic. All of these American investors want to put their money in China, even though the track record of American tech companies in China is really bad. 

To help you get a sense of just how bad, I made a big list of companies and checked to see how they’re doing in China today. Here it is:

Facebook: Blocked in China. You can’t even log on. Same with subsidiaries WhatsApp and Instagram.
Twitter: Blocked.
Pinterest: Blocked.
SnapChat: Blocked.
Dropbox: Blocked.
Box: Blocked
SoundCloud, Slack, Tumblr, Spotify: all blocked

And then there are a bunch of companies that are allowed in China, but aren’t doing as well as they’d like:

Microsoft: You can access Bing no problem, but piracy rates on Windows software in China are really high
Amazon: Less than 1% market share
Google: The Android operating system is doing well with around 80% market share. But in search, Google has less than 2% of the market.
Ebay: Exited China in 2006 after being crushed by Alibaba.
AOL: Launched a JV with Lenovo in 2001, shut down in 2004.
Groupon: A $100 million joint venture with Tencent completely flopped.

You might be wondering why these companies are failing. I hope so, because that’s the subject of today’s episode.

I think the usual reaction people have to seeing this list of failures is to blame the Chinese government in some vague way. We’re not exactly sure just what the government did, but if we had to guess we’re pretty sure that eBay would have succeeded in China had it not been for government meddling.

On the other hand, there’s another camp that hardly blames the government at all. Take our friend from Episode 2, Sir Michael Moritz. At a 2016 event, Moritz told an audience, and I quote him here: "The reason that most Western companies fail in China is because of their own doing. It's got nothing to do with the Chinese government or regulation, it's a whole series of Western imperial arrogance that brings about the downfall of many of these companies in China." 

“Western imperial arrogance” is a much different take from “government meddling.” Which one’s right?

When I made that list of failing companies, I was reminded of something called the Anna Karenina Principle, which was popularized by Jared Diamond’s Guns, Germs, and Steel. Tolstoy’s classic Anna Karenina begins with the sentence “Happy families are all alike; every unhappy family is unhappy in its own way.”

I’m going to adapt the Anna Karenina Principle for my own use. Here’s the China Bytes version: “Blocked companies are all alike; every unblocked company has failed in its own way.”

Today I’ve got a few stories about American failure in China. No, it’s not my autobiography. These are stories about how American technology companies have squandered millions trying to enter China––and despite superior technical know-how, financing, and operational experience, have failed again and again.


1:00:50 - Jerry Yang speaking in front of Congress to dissidents

That’s Jerry Yang, founder and CEO of Yahoo!, speaking in a November 2006 congressional hearing on the case of journalist Shi Tao.

Two years earlier in April 2004, the Chinese government requested information from Yahoo! China. The government wanted the account registration information for huoyan-1989@yahoo.com.cn, login times, IP addresses, and email content from February 22, 2004. Yahoo China employees complied. The account belonged to Shi Tao, who was detained 7 months later. Shi was tried in 2005 and spent nearly nine years in prison for sharing Chinese state secrets.

Here’s House Foreign Affairs Committee Chair,Congressman Tom Lantos at the same hearing:

Insert clip of Tom Lantos saying ‘It’s not that complicated.’

Actually, I have to say, it’s pretty complicated.

The complication began in 1997. Yahoo was a couple years old and Jerry Yang made his first trip to China. He was testing the water to see if it made sense for Yahoo to create a Chinese product.

In a remarkable accident of history, the Ministry of Foreign Commerce arranged a tour guide for Yang on that trip. The tour guide was Jack Ma, who would go on to found Alibaba. We’ll return to him in a moment.

It’s important to remember that back in the late 90s, Yahoo was the most promising internet company on the planet. They were dominant. And as Yang was a native Mandarin speaker, born in Taiwan but raised in the United States, prospects for Yahoo in China must have looked pretty good.

So Yang and his team launched Yahoo China in 1999, using the same web portal model that had achieved success in the US and several other markets. 

At the time, there were only about 5 million internet users in China, but Yahoo was still a bit late to the party. They struggled to compete with local competitors like NetEase, Sina, and Sohu, which were Yahoo-inspired web portals run by Chinese entrepreneurs.

When entering a new market, companies are faced with the “Build or Buy” dilemma. They can build a new business, as Yahoo tried to do. Or they can buy an existing local product. After several years of failure with Yahoo China, Yahoo decided to buy.

A quick note: it was around this time, in early 2003, that Michael Moritz left the Yahoo board.

In November of that year Yahoo announced the acquisition of a company called 3721, which was basically a web search company focused on Chinese keywords. If 3721 sounds like a horrible name for a company, you’re not entirely wrong. But, there’s a logic to it.

Arabic numeral domain names are attractive to Chinese companies. The URL for the internet portal NetEase, for example, is 163.com. Alibaba’s domestic e-commerce site is 1688.com.

There are a few reasons for this. First, there’s an abundance of homophones in the Chinese language. For example, 1688, Alibaba’s URL, is pronounced in Chinese Yao-Liu-ba-ba, which you’ll notice sounds a bit like Alibaba.

Also, and this might surprise you, Chinese people read Chinese characters, so URLs using roman letters can be difficult to read and type. You can imagine that if I told you to go to the URL “Chinese Character meaning Tree.com,” you might have a tough time typing it into your browser.

Finally, the numbers 3721 aren’t a homophone, but instead come from a common Chinese idiom that translates literally to “Don’t pay attention to 3-7-21,” but you’d use to express something like “without a care in the world.” 

The deal between Yahoo and 3721 closed in January 2014, with Yahoo paying about $120 million. 3721’s founder is an entrepreneur named Zhou Hongyi. Today Zhou is a billionaire because of the success of a different Arabic numeral company, 360.com. But during his time with 3721 and Yahoo, he wasn’t there yet. 

Sue Decker, Yahoo’s former CFO, writes of the acquisition that it “seemed like a good way to thread the needle of local product control, while also recognizing the goals of building cultural alignment and establish compliance controls where [Yahoo] needed them.”

That’s some nice sounding corporate speak, by mid-2004 it was clear that the venture wasn’t working out for reasons much more human. According to Decker, Zhou Hongyi thought that the Yahoo team was “overpaid and lazy,” while the Yahoos felt bullied by Zhou Hongyi. Zhou has since developed a reputation as China’s “Bad Boy of the Internet.” Which sounds ridiculous, I know.

Busy trying to manage its internal struggles, Yahoo was outpaced by its competitors which by then included Baidu and Google. At the same time, Yahoo had become embroiled in a Congressional investigation regarding the incarceration of Shi Tao.

But what really may have pushed Yahoo over the edge, was something that I don’t think was ever made public until Jerry Yang talked about it on the 9-9-6 Podcast in early 2018. Around 2005, Yahoo was also being threatened by the government. Here’s Yang on the 9-9-6 Podcast:

Insert Yang’s quote about being threatened by the Chinese government.

With a track record of failure and their business in China crumbling, the Yahoo team did what anyone would do in that situation: They gave $700 million as well as all of Yahoo China’s assets to their 1997 tour guide, Jack Ma. In return, Yahoo received 40% of Ma’s fledgling e-commerce startup Alibaba. As Sue Decker writes, “We realized we needed to be willing to give up all operating control.” 

The hands off strategy was a tremendous financial success. In 2012, Alibaba bought back almost half of Yahoo’s stake in the company, netting Yahoo $7 billion. But it was a failure for the Yahoo brand, which never took off in China. 

Alibaba shut down Yahoo China in September 2013. Shi Tao was released from prison the same month.

This is probably a good time to address the House Foreign Affairs Committee. Yes, Congressman, Shi Tao’s case was pretty complicated.

Part 2: The Alphabet of Success in China: Always Be Complying (with local law)

Jerry Yang was not alone at the November 2006 congressional hearing. Google was there, too.

Google provided Chinese language search beginning in the year 2000 through its US home page, Google.com, but it wasn’t until January 2006 that the company launched Google.cn. The difference between Google.com and Google.cn may seem minor, but it’s significant.

Google.com servers are located in the United States, so its services do not need to comply with Chinese law. As a consequence, the Chinese government can easily interfere with Chinese users’ access to the website. Since its earliest years, Google.com and its services have often been painfully slow when accessed from China. In late 2002, the Chinese government blocked access to Google.com altogether for two weeks.

As you can imagine, Google.com’s slow, unreliable service was unpopular with Chinese users. In the earliest years of the century, Google never breached 10% search market share in China. 

The Google eventually team realized that it wouldn’t be able to compete in China without moving its servers there and creating Google.cn. So in 2003 they sent a China-born employee named James Mi, to explore expansion. Among Mi’s contributions to Google was organizing a $5 million investment into a fledgling search engine called Baidu. Today Mi is one of the top venture capitalists in China.

Mi and the Google team recognized the upside in China, but also that launching Google.cn would require compliance with local laws, including censoring search results, something that Google had never done. A Google spokesman summed up the dilemma in a congressional hearing. He said: “The requirements of doing business in China include self-censorship - something that runs counter to Google's most basic values and commitments as a company."

At Google, there was heated internal debate about whether or not to enter China, but ultimately, the team decided to do it. 

Google CEO Eric Schmidt held a press conference in Beijing, where he delivered a quote that echoes Michael Moritz’s from the top of the program. “I think it's arrogant for us to walk into a country where we are just beginning to operate and tell that country how to operate.”

To lead the charge into China, Google poached a Microsoft Vice President named Kai-Fu Lee, who you will remember as one of the bullish venture capitalists from Episode 2. 

Lee and Google entered China with the same “Build not buy” mindset that Yahoo had. Lee hired fifty engineers in less than six months and Google launched Google.cn in January 2006. They sold their stake in Baidu a few months later.

Remarkably, before the entrance into China, Google had never translated its name. Lee pushed Google to localize and Google became Gu-Ge, a close phonetic translation made of the characters for “Valley” (Gu) and “Song” (Ge). It was a simple but necessary change: Most Chinese couldn’t spell Google.

At the time, Americans were already using Google services like Gmail and Google Maps, but Lee didn’t introduce them to China. Google’s China strategy was narrow: Build the best Chinese language search product.

Of course, that was Baidu’s strategy, too. Baidu CEO Robin Lee had already hired 1000 engineers to work exclusively on Chinese language search. And Baidu ran a television commercial, heard in the background here, with the catch phrase “Baidu Understands Chinese Better” that featured a dopey-sounding white guy speaking bad Chinese, which was pretty obviously a dig at Google. 

And in 2006, reality was that the catch phrase was based on truth. Baidu had better Chinese language search.

But Lee’s team made progress. Google China’s most impressive work was probably its implementation of Google’s now ubiquitous Google Suggest autofill feature, which was first launched in China. Google figured out–-before Baidu did, I might add––that Chinese users didn’t like to type as much as English-speakers. The number of homophones in Chinese makes typing characters kind of a pain. So autofill was especially useful.

The mechanics of Google’s self-censorship turned out to be pretty interesting. The Chinese government never handed Kai-Fu Lee a list of banned keywords. His team was just supposed to know what to censor. That’s a tough, maybe impossible task. But Google came up with a good strategy for dealing with it. They based their own censorship on what their competitors were doing. Google engineers developed algorithms that assured that if Baidu or other companies were censoring a certain type of content, Google.cn would censor it, too.

The search-first strategy had modest success. By January 2007, Google’s market share had about doubled to about 20%. 

But in the United States, Google was facing political criticism. They were accused by Congressman Chris Smith of "decapitating the voice” of Chinese dissidents. Meanwhile, Chinese activists hit Google from the other side of the Pacific. An open letter to Google from blogger Isaac Mao read, it “Seems you are adopting self-censorship which hurts those loyal users a lot [and] also devalues your motto.” Mao’s referring to Google’s “Don’t be evil.”

Despite criticism, Lee and Google stayed the course. In 2007, a year Lee called the “Year of the Product,” Google China launched 24 localized products and services, including Google Maps, a big hit. 

Things were going smoothly until 2008, when the government stirred things up. Authorities requested that Google begin to censor Chinese language search results on Google.com––in addition to Google.cn. That would have meant censoring searches by Chinese speakers worldwide. 

Google refused to comply. And perhaps not wanting to cause an international incident before the Beijing Olympics, there was no consequence. Google market share continued to creep up. 

In the first months of 2009, Google had 30% of the market, up from around 10% from when Lee joined the team in 2005.

That turned out to be the high water mark for Google in China.

The first warning sign was when the Chinese government criticized Google in January for not doing enough to censor pornography. A few months later, Chinese regulators requested to meet Google executives face to face. 

The regulators brought a laptop to the meeting, and began to demonstrate that the Google Suggest autofill feature was suggesting lewd, pornographic search queries. Not long afterward, authorities disabled some of Google’s search functions. You might remember from Episode 1 that Fanfou, the first “Twitter of China,” was shut down around the same time.

Just as tension between Google and its regulators was rising, Kai-Fu Lee resigned his role as President of Google China. It was September 2009. In Lee’s autobiography, he quotes himself in conversation with Google’s Vice President of Engineering Alan Eustace. Lee writes,  “I choose to leave now because everything is on track with Google China.”

But was it really? It’s possible that Lee believed everything was on track with Google China. But with hindsight at our disposal, it’s pretty clear that Google China was in stormy waters. And it’s also possible that Lee just didn’t want to go down with the ship. 

John Liu, who had been leading Google China’s sales efforts, took over as head of Google China. But things got ugly fast.

Just five months after Lee’s departure, Google announced that someone in China had launched a cyber attack on its servers. The attackers had tried to access the e-mail accounts of Chinese human rights activists. It was a serious, complex attack, that Google believed likely had government support. To this day, Google has not revealed the extent of the security breach. 

Google’s management team debated their response to the attack, and was actually split on whether or not to stay in China. CEO Eric Schmidt wanted to stay, while Larry and Sergey wanted out. In January 2010, almost exactly four years after the launch of Google.cn, Google announced that it would no longer censor its search results. Google.cn was shut down and traffic was redirected to Google’s Hong Kong site.

The New York Times and every other publication called Kai-Fu Lee for comment, but he declined all interview requests.

Eight years later, Google.cn is just a landing page that invites users to visit Google Hong Kong. Gmail, the Google Chrome browser, and Google search are all unavailable in China. 

Google does have a mainland presence including offices in Beijing, Shanghai, and––as of January 2018––in Shenzhen. They also recently announced an artificial intelligence research facility headed by China-born Stanford professor Fei-Fei Li.

I’ll conclude the Google story with a quote from Robin Li, CEO of Baidu. Lee said, "If the law clearly prohibits certain types of information, be it porn, or anti-government information, then I'm sure there's a reason for it. I'm an entrepreneur. I'm not a politician. I should not be in a position to make this kind of a call." Baidu has over 80% of China’s search market share.

Part 3: Two Years, Two Billion Dollars

Some people have argued that Uber is more like a Chinese company than other American tech companies. The Uber corporate culture is in some ways like typical Chinese work culture. Uber is aggressive and team members are self-proclaimed hustlers who take pride in long work hours that compare with the Chinese “9-9-6” schedule.

Still, the team was tentative when it came time to actually enter China. The recent experiences of Google and Groupon, among others, were effective caution signs. A group of Uber executives first visited China in April of 2013 to feel out the market. They wanted to make sure they did everything right.

As Uber CEO Travis Kalanick said later, “It’s just different than everywhere else… And, so, you can’t take your pattern or your model for other places and take that to China. You just can’t. You have to do it different.”

So Uber would borrow the wisdom of Kai-Fu Lee, and made sure that its China business was highly local. And it wasn’t just talk. Here’s Allen Penn, who was formerly Head of Asia Operations for Uber, speaking in 2016:


Another American Uber executive, Sam Gellman, removed himself from Uber’s China expansion team just so that the whole team would be Chinese.

This may be the kind of modesty that Michael Moritz advocated: There are plenty of talented, hardworking Chinese managers. Hire them to run your business and get out of their way.

Uber started testing in Shanghai in August 2013. They hired Davis Wang as their Shanghai business manager. Wang had previously been part of John Liu’s Google China ad-selling team back in 2006, the early days of Google China. Today he’s the CEO of MoBike, a Chinese bike-share giant. It’s worth noting that the Chief Operating Officer of Ofo, MoBike’s main competitor, also came out of Uber China.

Uber launched in Shanghai in early 2014. Today we’re used to Uber’s peer-to-peer service, where you can get a ride from anyone with a car, but Uber launched in Shanghai with a high-end service that allowed users to get rides only from licensed limousine companies.

When it came to finding partners, Uber was the beneficiary of great timing. Here’s Sam Gellman, speaking in 2014, https://youtu.be/UPr98GRhaOg?t=26m24s 

Among those partnerships was China’s top mobile payments platform AliPay. In late 2014, Baidu invested $600 million into Uber China,  and also helped Uber integrate mobile search and maps into their Chinese app. 

The partnership with Baidu was likely motivated by competition. Uber’s chief Chinese competitors at the time all had backing from Chinese internet giants. Didi, Uber’s main competitor, was backed by Tencent. Another strong competitor Kuaidi was backed by Alibaba. And now Uber was backed by Baidu. Baidu, Alibaba, and Tencent are the big three B-A-T internet giants in China, and all had an interest in rideshare.

Uber launched their peer-to-peer rideshare service in China in October of 2014. The result was explosive growth. Uber’s business in Shanghai grew faster than their business in any other city worldwide, though it was later eclipsed by tremendous growth in the western Chinese city of Chengdu.

In some ways, China was the perfect market for a rideshare business. In 2014, there were only 113 cars for every 1000 people in China, versus about 800 for every 1000 people in the United States. And there were plenty of drivers who wanted to work. The day Uber launched in the city of Zhengzhou, there were already 10s of thousands of registered drivers.

Uber China went from 1% market share to over 30% market share in just 8 months. But remember, 30% market share was the high water mark for Google. And in one critical way, China was a terrible market for a rideshare business like Uber: Rideshare wasn’t explicitly legal. 

In early 2015, trouble started. Government officials raided local Uber workplaces in Guangzhou, Chongqing, and Chengdu.

Meanwhile, industry competition consolidated. Didi and Kuaidi, Uber’s leading competitors, merged to form Didi Kuaidi.

But Uber was undeterred, and that summer Travis Kalanick wrote a blunt, optimistic letter to investors. He wrote, “Simply stated, China is the #1 priority for Uber’s global team.” 

And in April 2015, Uber made a very savvy hiring decision. They brought on Liu3 Zhen4 as Head of China Strategy.

Liu is one of a kind. She’s the granddaughter of the founding father of Patent Law in China, and an accomplished Berkeley-educated lawyer herself, with years of Silicon Valley experience at Wilson Sonsini. She is also the niece of Liu Chuanzhi, the founder tech giant Lenovo. But perhaps most relevant to this story, she is also the cousin Liu Qing, who was the President of Didi. 

If you were looking for someone to broker an eventual deal between Uber and Didi, there was nobody better than Liu Zhen.

Liu filled Uber China with an entirely bilingual team. Something that would been impossible for Yahoo 15 years ago. And Uber expanded into over 60 Chinese cities. 

But the expansion came with great cost: Uber was bleeding as much as $1 billion per year. In the rideshare industry, a company like Uber beats its competitors by offering both lower fares to riders and higher wages to drivers. For both Uber and Didi, that meant raising tremendous amounts of capital and using it to subsidize drivers and rides. In 2015, Didi had set aside $4bn just for that purpose.

Both sides were hurt in the subsidy war, but as late as July 21st, 2016, Liu Zhen was publicly dying that there were any talks about a merger between Uber and Didi. 

On July 28th, Uber had an apparent breakthrough. Liu
published a post on the Uber corporate blog titled “China Drives Forward with New Regulations.” It announced that the Chinese government had legalized rideshare. So Uber was out of China’s regulatory gray area.

But just three days later, Uber announced a merger with Didi. Uber sold its China assets to Didi for a 17.7% economic interest in the company.

Two months later, Liu Zhen left Uber. Today, she is senior vice-president for corporate development at ByteDance.

Travis Kalanick is no longer CEO of Uber, and is now managing an investment fund called Ten One Hundred, which focuses in part on emerging innovations in China.


One of Jerry Yang’s conclusions on the 9-9-6 Podcast is that, because of political issues, information services is a particularly difficult industry for American companies in China. It’s a quick, dirty––and reasonable––explanation for the troubles of Google and Yahoo. 

That’s one reason that Yahoo chose to invest in Alibaba. At the time, many people questioned why Yahoo, an information services company, was investing in an e-commerce company like Alibaba. The answer was that Alibaba didn’t receive as much regulatory scrutiny as Yahoo did. 

Another explanation is that things were more difficult for Yahoo and Google and those first internet companies to launch in China. The rulebook for the internet was still being written and nobody was really sure what the laws were going to be. That’s why regulators were able to threaten Jerry Yang with shutting down Yahoo, and why Google’s self-censorship was so murky.

But Uber’s case defies both of those explanations. Though it was a transportation business, it couldn’t escape government intervention. And its case proves that for technology companies, the rules are never set in stone. It doesn’t matter if a company arrives in China in 1999 (like Yahoo), 2006 (like Google), 2013 (like Uber), or in 2020.

My final word is this topic of failure in China is that you could argue that none of these companies really failed. Uber has nearly 20% economic stake in its biggest global competitor. Google has 80% of the mobile operating system market in China. And Yahoo has Alibaba.

So maybe it’s best to withhold judgement. As former Google CEO Eric Schmidt said in 2006, “We will take a long-term view to win in China. The Chinese have 5000 years of history. Google has 5000 years of patience in China.” I’m afraid that quote may be repeated for the next 4988 years.

Thanks for listening.

For Part 1, I relied heavily on Alibaba: The House That Jack Ma Built by Duncan Clark.

For Part 2, Steven Levy’s In the Plex was indispensable.

I also suggest checking out the 9-9-6 Podcast for interviews with Jerry Yang, Kai-Fu Lee, and Liu Zhen.

Episode 4: Technology Espionage - Past, Present, and Future

Episode 4: Technology Espionage - Past, Present, and Future

Episode 2: Red Bulls

Episode 2: Red Bulls