Twenty Years From Joyo to Amazon China: When a Few Computer Hobbyists Tried the Internet | 1999 Memoir ⑤

Twenty years of Joyo Amazon

Amazon China’s e-commerce business is basically finished.

Caijing, citing people familiar with the matter, reported that NetEase and Amazon were in talks to fold Amazon China’s overseas-purchase business into Kaola, and that the two sides had signed an agreement by the end of 2018. Both NetEase and Amazon China said they had “no comment.” Neither denied it.

Since Amazon bought Joyo in 2004, the site has gone from China’s second-largest e-commerce company to a niche business with less than 1% of the market. Overseas purchases were about the last bit of life Amazon’s head office put into it. Amazon has kept shrinking in China: its national network of warehouses went from 13 in 2015 down to two — Beijing and Kunshan — by the end of 2018. There is not much hope left.

Twenty years have passed since Joyo, Amazon China’s predecessor, went online on January 3, 1999. In its earliest days Lei Jun was its chairman, and its investors were China’s largest software company, Kingsoft, and its largest hardware company, Lenovo. Joyo embodied the fearlessness of 1990s Chinese internet ventures, lived through the bubble bursting and swelling again, and saw foreign companies in China rise and then fall.

It all started with a few young computer hobbyists who hung around CFIDO.

Selling to Computer Hobbyists

In March or April 1998, 23-year-old Gao Chunhui came to Beijing and met 29-year-old Lei Jun, CEO of Kingsoft, face to face for the first time.

Gao Chunhui told QDaily he had just quit his job, because he could not stand the internet where he lived, in Shenyang. “One time I went to update my website at three in the morning and couldn’t even get the update through. That’s when I decided I absolutely had to get out of Shenyang.” The two or three thousand yuan a month his personal site brought in was also part of why he could afford to quit.

Gao Chunhui had been using computers since 1989, his first year of middle school, and by 1998 he was a well-known webmaster and hacker. His site, www.paulgao.com.cn, was the only personal website CNNIC included in its internet survey that year. It called itself a personal website, but it carried all sorts of things — above all downloads of the keygens he wrote himself to crack software.

Cracking software was also how Gao Chunhui and Lei Jun came to know each other.

Both hung around CFIDO (Huiduo Net). Started in 1991, CFIDO was not a single website but a collection of more than 100 BBS forums across the country. Calling them forums is generous — there was no graphical interface like Tianya’s; they worked more like a set of email discussion groups.

Reaching other people on this network took a certain amount of skill, so the circle was small and mostly used real names. People usually put their phone numbers in their signature lines so others could call them.

Lei Jun had just been through the failure of Kingsoft’s Pangu project, and mostly hung around the CFIDO boards where people discussed programming and the programmer’s life. Back then he did not yet tell the story of how Fire in the Valley (《硅谷之火》) had inspired him to start a company; he thought he could be a programmer for life. In a set essay on the site called A Programmer’s Life (《程序人生》), he wrote: “After I graduated from university I was assigned to a research institute, and I didn’t really fit the atmosphere there. I jumped to Kingsoft and began my career as a professional programmer. Later I became a manager at Beijing Kingsoft, but I was still a programmer.”

Lei Jun’s boss, Qiu Bojun, who wrote WPS 1.0 in 1989, was an even busier CFIDO user, and spent 150,000 yuan to run the West Point BBS. This forebear of every software entrepreneur in China, whose ID was Vincent Q — everyone called him VQ — posted on every subject there was, from the China-Japan dispute to Kingsoft to a new download of Duke Nukem 3D. A Kingsoft employee from those years recalled: “I’d get to work in the morning and find Qiu Bojun already there. He said the station had gone down around 4 a.m., so he’d hurried in early to restart it — otherwise nobody else could dial in.”

Lei Jun had been writing an encryption program called BITLOK since university. Gao Chunhui was active on the software-cracking BBS CRACK.CHINA, where he shared cracking techniques. Around 1996 he found a hole in BITLOK and emailed Lei Jun about it. Lei Jun asked him on the spot whether he wanted to come work in Beijing, but Gao had already taken a software development job in Changchun and had not yet started his personal site, so he did not go.

Few Chinese people were online at all then — 620,000 by the first half of 1997 — and fewer still could work CFIDO. Quite a few internet entrepreneurs who later became famous hung around there, among them Pony Ma, who ran the Ponysoft BBS in Shenzhen and turned up now and then on CRACK.CHINA.

In 1997 Pony Ma wrote an excited, long post about going to Hong Kong for an Oracle development course. He listed every computer book and software CD he had bought, and said he had not been able to bring himself to spend 1,100 yuan on Microsoft’s newly released FrontPage 97 web tool — he would wait for a pirated copy.

This 26-year-old engineer, who called himself by a homely northern word for “I” and who, when he was not digging into development, liked to share Red Alert tips, was still a year away from starting his own company — something his CFIDO friend Ding Lei had already done. But one line in Pony Ma’s signature at the time he did later make real: “A world with no strangers.”

In 1999 Gao Chunhui went to Beijing and met Lei Jun, and the two decided to work together and build a website as a company. That was the earliest Joyo.

As Gao Chunhui remembers it, he wanted to build a site in an area where he already had something to show for himself, so he and Lei Jun discussed a computer news and downloads site modeled on CNET.

At first Lei Jun treated Joyo as a way to test the water. Gao Chunhui told QDaily: “He hadn’t figured it out either — he was just going to spend 300,000 to 500,000 yuan a year and give it a try. That cost was acceptable to him.”

On January 3, 1999, Joyo (joyo.com) went online. Gao Chunhui even renamed his own, more influential personal site JOYO & PAUL GAO to send users to Joyo. Eight months after launch, Joyo’s homepage passed 1 million visits.

But the disagreement over what to do, and how, was there from the start.

Gao Chunhui saw himself as a partner — the plan had been to set up a company from the beginning. That came to nothing. In February 1999 he joined Kingsoft instead, running the Joyo website division inside the company, and working out of Kingsoft’s offices.

Joyo already drew plenty of attention as a downloads site, but e-commerce was becoming the bigger internet business. In the summer of 1999 CCTV staged the “72-Hour Internet Survival Experiment,” and people started buying things through 8848, an e-commerce site founded that same year. A few months later Alibaba and Dangdang went online one after the other.

Gao Chunhui had used 8848 and thought the infrastructure for e-commerce was not there yet.

But Lei Jun wanted to do e-commerce. That takes money, so Lei Jun spun Joyo out as its own company, with Kingsoft and Lenovo each taking a stake. Gao Chunhui says he met Lenovo CEO Yang Yuanqing a few times and found him quite hesitant about internet business; Lenovo’s later internet attempts likewise put in a little money and pulled out at the first setback.

Even so, Joyo’s relaunch as an e-commerce site went ahead.

Conflicts like this were common at Chinese internet companies of that era. Inside Sina there was the same tension: serve the “web junkies” better, or go after a mass audience. Sina founder Wang Zhidong insisted on the mass market: “I don’t want web junkies spending five or eight hours a day glued to Sina. I want more people spending half an hour a day on Sina. Those users have the highest commercial value and the greater social influence.”

“From far away they look like gods; up close they’re all just people. Everyone’s weighing their own interests. You need some bargaining power — only when both sides are equals can you sit down and talk. Walking away means you had no power to bargain with him,” Gao Chunhui told QDaily. In April 2000 he announced on his own website that he had decided to leave Joyo over disagreements about strategic direction.

Joyo Turns to E-Commerce

Gao Chunhui’s departure was no small matter in the IT world of the day; China Business Journal used it as a hook for a piece titled The Internet’s First Layer of Bubble Has Burst (《网络的第一层泡沫已破裂》). But the e-commerce bubble in China was only just starting.

Gao Chunhui’s replacement, Wang Shutong, was one of the first generation of managers foreign tech companies trained in China. She had run sales teams at Microsoft and then Cisco from 1993 on, until Lei Jun poached her for Joyo in early 2000. Gao Chunhui says that in his last few months at Joyo some of his communication with Lei Jun went through Wang Shutong.

At the end of that year, Wang Shutong announced that the goal for 2001 was to pass 100 million yuan in revenue and turn a profit. Gao Chunhui thought that was a stretch — Joyo was spending 20 million to 30 million yuan a year at the time.

Wang Shutong formally took over Joyo in April 2000. By the next month the site had been overhauled: no more IT news on the homepage, and in the top left corner a line reading “If you’re a longtime Joyo user, you can still go straight to it.joyo.com for IT content!”

In their place came books, video discs, and CDs. Wang Anyi’s I Love Bill (《我爱比尔》) sat dead center in the homepage’s promotion slot at 3.8 yuan — cheaper than a pirated copy. Next to it was a 10-yuan VCD of A Chinese Odyssey Part Two: Cinderella (《大圣娶亲》) and its famous line: “Once a true and sincere love was placed before me, and I did not treasure it…”

The man actually in charge of books was Chen Nian, whom Lei Jun brought in — a book-trade veteran with no internet experience at all. As Chen Nian later remembered it: “Nobody knew how to do e-commerce; everyone was clutching a copy of The Amazon Legend (《亚马逊传奇》).” That book topped Joyo’s bestseller list in May 2000, which tells you something about how small the group of people buying books online still was.

It is not hard to see why Chinese internet people were so taken with Amazon’s story. In January 2000, just as the bubble reached its widest, Amazon was worth $30.85 billion despite huge losses — far more than almost every big Chinese state enterprise listed overseas except China Mobile. Amazon was less than five years old.

As Julia Kirby, an editor at Harvard Business Review, put it, “Amazon was born of strategy.” Jeff Bezos, a Wall Street technology executive in his early thirties, saw that a website’s endless choice of books was its advantage over a physical bookstore, and before launch he had already laid out a long plan: sell books to catch the first users, then, once they had a shopping habit, sell them other things — anything at all. He wrote that plan into the shareholder letter for the IPO. By 2000 Amazon was already selling CDs, videotapes, DVDs, toys, and more.

Joyo, turning into a shop, and Dangdang, born at the end of 1999, both hoped to build a shopping habit through the sheer breadth of books, CDs, and video discs, which no offline store could stock in full. “A Chinese Odyssey (《大话西游》) at five yuan a disc sold out on the spot.” No one had expected a licensed disc to sell in that volume.

To get Joyo’s name known, the company ran ads in the magazine Reader (《读者》) for several months running.

But China’s e-commerce infrastructure was very thin then. Bezos said more than once that Amazon took off so fast because of credit cards and the U.S. Postal Service. The first solved paying online; the second made sure goods reached customers in a reasonable time. China had neither. UnionPay did not even exist in 2000, and China Post still worked at a pace where a parcel could take weeks.

So the e-commerce sites of the day served little more than Beijing, Shanghai, and Guangzhou. As late as 2003, Dangdang could handle only a small share of deliveries itself, and handed 85% to Xinhua Bookstore, both partner and competitor.

Joyo had opened only three sites — Beijing, Shanghai, and Guangzhou. Because the infrastructure was so weak, the homepage still carried, in large type, instructions for ordering by phone.

By 2000 8848 had launched flower delivery — an “O2O” service — covering every category and advertising for users as it prepared to push toward an IPO. Then in the summer of 2000 the bubble burst, Nasdaq internet companies collapsed in bulk, and the investment banks that had taken them public were punished for misleading investors. 8848, meanwhile, split in two when its shareholders could not agree, and over the following years kept changing direction until it vanished.

8848 disappeared early from the history of the Chinese internet. Dangdang, by comparison, had an easier time.

Joyo’s fundraising did not go as smoothly as Dangdang’s, and its team was not stable either. Wang Shutong left after less than two years and later founded the cross-border shopping site DHgate.

By 2004 Joyo’s main push was no longer books but DVDs, CDs, and computer software discs, which had a wider audience in China. The top search terms on the site on June 16 that year included Normandy, Stefanie Sun, The Lion King, and Transformers.

Amazon Buys Into China

On August 19, 2004, Amazon announced it was buying Joyo for $75 million.

By then Amazon had come through the bubble at home, proved it could turn a profit, and stretched its business from books to any category a person could buy online. It was also investing in more: the Kindle e-book project and AWS cloud computing both began that year.

“This acquisition is recognition of what Joyo has achieved over more than four years.” In the coverage at the time, this line, spoken by Lei Jun as Joyo’s founder, was used as a headline again and again. A few months later, though, Lei Jun admitted in an interview that Chinese e-commerce was still in the cash-burning era, and that Joyo did not have many options.

Handing over control was no bad thing for Lei Jun himself. From May 2000 to the sale in 2004, total investment in Joyo — from Kingsoft, Legend Capital, Tiger Technology Fund, and private money from Hong Kong and Taiwan — came to more than 97 million yuan. Every investor, Lei Jun included, cashed out through Amazon’s purchase. Lei Jun used that money to become a successful angel investor, and the decade of investing that followed had a lot to do with his later ability to found Xiaomi.

But before Amazon moved, an August 2003 issue of The Economist had already handed the title of “China’s Amazon” to Dangdang. Two days before Amazon announced it was buying Joyo, Dangdang gave China Youth Daily a vivid story: “an Amazon board member, a famous Wall Street investor,” had put a copy of The Economist in front of Bezos; Bezos read the cover story and contacted Dangdang at once to arrange a meeting; and he had finally offered $150 million.

That piece about Dangdang never actually ran on the cover of any regional edition of The Economist that year. How much else Dangdang embroidered, no one can say. But Dangdang really was China’s largest e-commerce company at the time.

Li Guoqing, for his part, said publicly that Dangdang had a “welcome gift” ready for Amazon: before the end of the year it would at least double its range of Chinese books and audiovisual titles, and it would launch an even fiercer round of price cuts. New Dangdang warehouses in Shanghai and Guangzhou were about to open, too, as the company moved to expand faster in East and South China.

Joyo’s managers, meanwhile, left one after another in the year after the sale. Vice president Chen Nian went on to found Vancl. He said: “The Americans basically didn’t listen to you. Putting together a China plan that matched your own understanding of the market was just too hard.”

Yet neither Li Guoqing nor the newly arrived Amazon took the two rivals that would prove fatal at all seriously. In 2004, JD.com, which sold computer parts in Zhongguancun, had only just started selling online. Jack Ma’s Taobao had been running for a year and a half, but the e-commerce reports of the day filed Taobao under a different category altogether, and did not think it would compete with B2C players like Dangdang and Joyo.

Bezos Comes to China to Put Out the Fire

In its first years in China, Amazon tried to copy the path that had worked at home.

Amazon did not advertise. It put most of its money where customers could barely see it — warehousing and logistics — and built Amazon’s largest operating network outside the United States.

In its third year in China, Bezos visited in September 2007 and gave interviews to a group of Chinese and foreign media, The Wall Street Journal among them.

Bezos announced that Amazon would bring algorithmic recommendations to China, keep investing there, and build new data centers — all technology none of its Chinese rivals had.

The warehouses in particular needed no one to work out where goods should go. Amazon’s management system set the best picking route from data such as the average efficiency of the staff at a fulfillment center and each product’s shelf life. Workers did not have to know where any item was stored; in a fulfillment center run by algorithms and forecasts, how much a worker’s own experience and learning could affect warehouse management was cut to a minimum.

But technology was not Joyo Amazon’s biggest problem.

In its fourth year online, without a cent of revenue, living on money from an Alibaba that had itself only just survived the bust, Taobao reached nearly 100 billion yuan in transactions in 2008. JD.com took $10 million from Capital Today, and Richard Liu announced that JD would expand from 3C products into every category and become a one-stop shopping platform, and that it would build its own combined warehousing and delivery network.

Capital started pouring into Chinese e-commerce.

Amazon, meanwhile, kept to its global habit of not advertising its retail business, counting on low prices, service, and word of mouth to bring new users in.

That year Amazon ran its first-ever ad in China, promoting the short domain z.cn. Joyo Amazon was formally renamed Amazon China. Amazon China’s CEO at the time, Wang Hanhua, said: “From now on Amazon China will be an operations center in the global market, not a decision-making center.”

But the brand campaign brought no matching rise in sales, and for years afterward Amazon’s international arm held it up as an example of money wasted.

The managers in China saw the branding problem — Wang Hanhua had made his name on product placement at Motorola — but he later said he simply could not make the call.

Bezos’s enthusiasm for China did not last much longer. That year the Kindle went on sale, and AWS was already commercial.

In 2011 JD.com founder Richard Liu put his finger on it in an interview: “If the head of the China business can’t decide anything, what is there to say about execution? Go ask Hanhua — can he say that everything he wants to do succeeds? I can.”

In the end, Amazon had better technology than JD.com or Alibaba, but those two, each led by its founder in person, raised billions and, next to Amazon China, burned cash almost without limit, making up for what they lacked in efficiency.

Unlike in the United States, Japan, Britain, or Germany, Amazon ran into powerful rivals in China: backed by capital, they did not have to care so much about efficiency, and the gap in flexibility and autonomy between a startup run by its founder and one overseas market of a large corporation could by itself decide who won.

By 2012, when the e-commerce war broke out, nobody was thinking about Amazon anymore.

A former e-commerce reporter told QDaily: “I remember the wild-west days of e-commerce — every morning when I opened my eyes and every night before bed I’d nervously scroll Weibo, never knowing what motor-mouths like Richard Liu and Li Guoqing would say next. You’d set the lowest price on the whole internet, I’d go under it. JD.com, Alibaba, Dangdang, Suning, Gome, Yixun, Yihaodian, Paipai… reporters were going out of their minds every single day.”

The “wild-west days of e-commerce” she meant were 2011 and 2012, when several mall-style shopping sites fought fierce price wars.

In August 2012, starting with JD.com, sites including Suning.com, Gome Online, Dangdang, and Paipai launched the “8/15” price war. On August 14, 2012, JD Mall CEO Richard Liu suddenly posted on Weibo: “JD will make zero profit on major appliances for the next three years.” Suning then said it was joining in, and Richard Liu’s Weibo feed turned into a live broadcast of price cuts.

At the time JD.com had 8.7 billion yuan of cash on hand, and Suning was preparing to issue 8 billion yuan of corporate bonds.

That price war burned on until the regulators stepped in, scolded everyone equally, and banned any repeat. But the e-commerce shopping festivals — JD’s 6/18, Suning’s 8/15, even Tmall’s “Double 11” — also became accepted and familiar in those same years, through exactly this kind of price war.

“Joyo Amazon actually took part in those price wars too, but they were a foreign company and the China CEO had no real say. It wasn’t like JD.com or Suning, where the executives themselves were posting on Weibo nonstop — so it just looked as if Amazon had no voice at all while the local players made all the noise,” she said.

In China, JD.com told an Amazon story of its own. By investing in logistics again and again and burning cash on subsidies to win new users, JD kept drawing investment. In the years when capital was fighting its way into the Chinese internet, JD grew from small to large on the strength of that capital.

In 2015, speaking at an e-commerce summit, Li Guoqing took a shot from a distance: “Very few people in China have President Richard Liu’s luck — investors will hand him $2 billion just to lose it. He’s far too lucky!”

Once the Shop Failed, the Spiral Turned Vicious — and Everyone Involved Moved On to Something Else

The Kindle went on sale officially in China in 2013, and Amazon even broke its own global rule against paying advances for e-book rights, spending 1 million yuan to take the exclusive first e-book release of Yu Hua’s new novel The Seventh Day (《第七天》) from rivals such as China Mobile and iReader.

The Kindle’s arrival in China was treated as a turning point. Tao Weipeng, a planning editor at CITIC Press Group’s e-book center, told QDaily that in the month the Paperwhite went on sale, sales of the e-books CITIC had on offer quadrupled.

More important was Amazon’s overseas-purchase service, which launched on “Double 11” in 2014. For the retail business, overseas purchases were the last chance.

“It took Amazon a long time to realize that copying the American Amazon model wouldn’t work in China,” Doug Gurr, who became president of Amazon China in August 2014, said in an interview with the Chinese edition of Bloomberg Businessweek.

On August 20, 2014, Amazon announced it would set up an international trade headquarters in the Shanghai Free Trade Zone to develop cross-border e-commerce, cross-border trade, and cross-border finance. Over the next three years Amazon connected its U.S., British, Japanese, and German storefronts to the overseas-purchase service one by one.

Overseas goods were plugged into China, and the marketing picked up. But Amazon China’s rivals were no longer what they had been.

JD.com and Alibaba both went public in 2014, raising tens of billions of dollars between them.

At the same time phones replaced computers as the main way people got online, and the e-commerce war became a zero-sum game: people stopped ordering through web pages and started installing apps. Most people will not install many apps, so in e-commerce the big kept getting bigger and the small kept getting smaller.

On top of that, Amazon’s head office began to rein in how fast its overseas operations burned cash, putting more money into its video rights war with Netflix. Amazon China entered a vicious spiral: fewer customers, so it cut the products that did not pay and lost the leverage to get good prices, so it had fewer customers still.

Cross-border shopping was never going to carry the traffic a full-line retailer needs, and in the end it could not save Amazon China’s e-commerce business. There is no hope for it in China.

In China, Kindle is a hardware business that makes money, and its research and development leans on Amazon’s global operation. Away from Amazon’s retail channel, it might actually do better.

As for AWS, the other business with enormous potential, it was handed to a Chinese partner to run back in 2016.

Twenty years after it was founded, Joyo’s business is coming to an end. And the people who took part in it, who shaped it, have each gone on to different futures.

Li Guoqing, founder of Dangdang, Joyo’s main early rival, published an open letter this week announcing he was leaving the company. He said Dangdang should think hard about not having sided with either Alibaba or Tencent, but that turning Amazon down was definitely right.

Qiu Bojun, Joyo’s largest shareholder and the founder of Kingsoft, retired in 2011.

Lei Jun, Joyo’s chairman, spent eight years turning Xiaomi into the only publicly listed Chinese phone maker.

Wang Shutong, Joyo’s first CEO, left to found the cross-border shopping site DHgate, and is its CEO to this day.

Chen Nian, general manager of Joyo’s books division and later vice president, is still running Vancl.

Wang Hanhua, who ran Amazon China for eight years, went to several different places after leaving; he is currently president of Sonos Greater China, and says he has found a balance between work and life.

Gao Chunhui, who started Joyo, now has a Weibo bio that reads: “Founder of Joyo, iMobile, ECSHOP software, and the iWallpaper HD (《爱壁纸HD》) app.” Open Paulgao.com, the site that once made Gao Chunhui famous, and it redirects to www.ipip.net. This is Gao Chunhui’s new venture — a geolocation database for companies that finds a visitor’s location from their IP address. Its customers include Xiaomi, Toutiao, Didi, and NetEase.

Cover image from Scottish Government on VisualHunt.com / CC BY-NC