Meituan Is Finally Going Public. What Have Round After Round of Huge Internet Investments Done to Consumption, Work, and Capital Itself?

For Meituan, going public is nowhere near the end of the story.

On Monday, the Hong Kong Stock Exchange published Meituan-Dianping's prospectus. Eight years after it was founded, Meituan is finally going public.

The group-buying wars, fought at one point by more than a thousand websites, ran all the way to the merger of Meituan and Dianping. This year Meituan and Alibaba — which bought Ele.me and Baidu Waimai — came to split the food-delivery market between them. And Meituan's $2.7 billion purchase of Mobike made it the biggest player in bike-sharing. Outside e-commerce, nearly every new kind of internet business that has stuck in our daily lives now sits inside Meituan.

Over 2017, 310 million users and 4.4 million merchants generated 357 billion yuan of transactions on Meituan. By taking commissions from merchants, selling marketing and advertising, and charging delivery fees, Meituan earned 33.9 billion yuan in revenue — and lost 3.8 billion yuan.

The food-delivery war cost it 14.2 billion yuan over the past three years, though the losses are shrinking. Most of the money it raised over the past year is still unspent. The prospectus shows that Meituan holds 45.2 billion yuan in cash — enough to last several more years even if it burned all of it on food delivery at the pace of the past three years.

But food delivery is not the only thing Meituan has to burn cash on. Its fight with Didi over ride-hailing broke out this year. Mobike is not a monopoly either. Alibaba has taken a fresh interest in Koubei, its in-store dining and review platform, naming it one of the "divisions" in its "New Retail" campaign. And Ele.me and Baidu Waimai, now inside Alibaba, may find new openings of their own.

A successful listing will bring Meituan a lot of money, but it is unlikely to end eight years of cash-burning, round after round.

How that cash-burning ends decides what Meituan's investors get back. It decides whether eight years of investment and subsidy add up to one enormous bubble.

Meanwhile, it has genuinely changed how a great many people live. More than a million riders move up and down the roads every day and fill office-building elevators at lunchtime. Restaurants now have to worry not only about people walking past the door but about traffic online.

For merchants, for users, for riders, the billions Meituan raised eight years ago have been disruptive beyond any doubt.

This prospectus is a chance to see how each of Meituan's businesses actually developed, and what the models it stands for — O2O (online-to-offline), the sharing economy — have really done to one industry after another.

After Three Years of Paying Users to Order, Picking Up the Phone for Delivery Became a Habit

In 2010, Meituan was a boutique group-buying site that pushed a single discounted deal each workday. Four years later, more than 5,000 competitors had disappeared and Meituan-Dianping was the group-buying site still standing.

Today group buying is a minor part of Meituan. Founder Wang Xing defines the company as an "e-commerce platform for life services." In the prospectus, Meituan says it uses technology to connect consumers and merchants, provides services that meet people's daily need to "eat," and extends from there into a range of lifestyle and travel services.

In Wang Xing's words: "People who want to eat out, order delivery, watch a movie, travel, or rent a car are basically the same group."

Meituan's business is now split three ways: food delivery; in-store, hotel and travel; and new initiatives and others. Food-delivery revenue was 1.75 billion yuan in 2015, 5.3 billion yuan in 2016, and 21.03 billion yuan in 2017. By 2017, delivery accounted for 61.9% of Meituan's total revenue.

We may all feel familiar with food delivery, but fewer people treat it as a habit than you would guess. The prospectus shows that in 2017 the 10% who ordered most often placed an average of 98 orders over the year. Measured against Meituan's user numbers, that is 30 million people out of 1.4 billion — and Meituan already holds half the delivery market.

Spread across all users, it works out to 18.8 delivery orders per person per year.

That is not a high frequency. Meituan has long compared its revenue model to Taobao's, but as far back as 2014, a third of China's internet users were buying on Taobao, and each of them placed 52 orders a year on average.

The prospectus shows that in 2015, the year Meituan started delivery, the business ran a gross margin of -123.7%. For every 100 yuan it collected from users on paper, it kept nothing and paid out another 123.7 yuan on top.

That amounts to a full year of free delivery, courtesy of Meituan. Subsidies on that scale, with Ele.me and Baidu Waimai joining the war, finally got people into the habit of ordering in.

Meituan's delivery operation grew steadily more efficient, and its losses eased over the next two years. By 2017 the business turned a gross profit, with a gross margin of 8.1%, and carried more than 170 billion yuan of transactions for the year.

That efficiency also shows in how much of each transaction Meituan keeps as revenue. From 2015 to 2017 that share climbed steadily, from 3% to 9.5%.

Next to Ele.me, now owned by Alibaba, Meituan is without question the far more efficient company: stop subsidizing delivery and it can make money, however thin the margin.

530,000 People Deliver for Meituan Every Day, and the New Job Pays No Better Than Foxconn

Food delivery created the rider, a job that did not exist at all before, and it has created a great many of them.

A 2017 report on the sharing economy from the Internet Society of China and the State Information Center counted 40 million people working in this kind of job — 4% of China's labor force.

They have become the most basic and most important part of the new economy. Rider numbers grew alongside the frenzied subsidy wars of the early years: subsidies brought more orders, and riders took home more money.

As with every business built on subsidies, the good days did not last long. QDaily has reported that as competition escalated and the giants merged, riders' base pay, commissions, and bonuses kept shrinking. By the end of 2017 the average in Beijing had fallen to 5,000-6,000 yuan a month.

The prospectus says Meituan had 531,000 active riders in 2017. Many more have registered but taken few orders or none; Meituan describes this as "providing employment opportunities for 5 million people," while Ele.me's official figure is 3 million. Add the two together, counting every duplicate registration, and China has at most 8 million delivery riders — and the great majority no longer do the work.

Meituan's rider costs came to 18.3 billion yuan in 2017, which works out to about 34,000 yuan a year per rider.

By comparison, a new worker at Foxconn's Zhengzhou plant earns 45,000 to 54,000 yuan a year including overtime, and that leaves out the dormitory and the year-end bonus Foxconn provides.

Meituan's riders may also earn less than parcel couriers. A ZTO Express job posting shows a ZTO courier in Beijing earning 60,000 to 96,000 yuan a year.

That 34,000-yuan average has something to do with the fact that riders in second- and third-tier cities have multiplied and now outnumber those in first-tier cities.

The shift from in-house riders to outsourced ones matters too. A rider survey Meituan published this year says in-house riders earn the most, usually 6,000 to 8,000 yuan a month. Among riders recruited by agencies, who do not sign with Meituan directly, nearly half make less than 4,000 yuan a month. Crowdsourced riders who sign up on their own earn less still, because they take fewer orders.

The prospectus does not disclose the mix of in-house, agency, and crowdsourced riders. But reports and protests show that once the competition settled, both Meituan and Ele.me began shrinking their in-house fleets. In March it was reported that Meituan-Dianping was carrying out a plan to dismiss all 15,000 of its in-house riders nationwide and fill the gaps with capacity hired through agencies. Meituan replied that moving some stations in a few cities to a mixed model was a normal management measure.

Riders and delivery stations told QDaily in May that Meituan has stopped hiring in-house riders in Shanghai.

The reason is easy to see: an in-house rider is effectively an employee of the delivery company, and handing the work to agencies saves a great deal in management costs.

But the job has no skill barrier, so riders who found work in the sharing economy can only accept whatever arrangement they are given, fair or not.

530,000 people is not a small number, and that is Meituan alone. Foxconn employs 1.3 million in all of China.

Yet in this new industry, pay and protection turn out to be worse than what people expect of a "sweatshop."

Group Buying and In-Store Business Still Bring People to Restaurants, but Delivery Subsidies Have Cut Restaurant Foot Traffic

Meituan does not run restaurants. Both its group-buying and its delivery business depend on local merchants, but — and this is the irony — the two mean completely different things to those merchants.

Group buying (and reviews) means restaurants, attractions, and hotels display their information on Meituan; users spend there and leave reviews, and other users then decide where to spend based on those reviews. A shop gets more customers through online exposure, and Meituan, having built this "marketplace" of reviews and listings, collects advertising fees.

By the figures Meituan cites in its prospectus, China has 11.8 million local service merchants; its 4.4 million active merchants are 37% of them.

Delivery works on a different logic: those 4.4 million restaurants look more like Meituan's suppliers. People order on Meituan, Meituan's riders collect and deliver the food, and the whole process trains users to enjoy a meal without ever appearing in the restaurant. The "offline traffic" is intercepted by Meituan.

According to research firm Trustdata, delivery coverage in Chinese restaurants — the number of merchants offering delivery divided by the total number of restaurants — rose from 17% in 2017 to 37% in the first quarter of this year, and 48% in first-tier cities. Delivery has laid down a whole layer of infrastructure that changes everything from how restaurants win customers to how customers reach them, and no restaurant brand, large or small, can ignore the channel.

The hard years for brick-and-mortar business are not unrelated to delivery's quick rise. A report from Winner Data shows that growth in average daily foot traffic at physical stores slowed month after month from 2013 to 2016 — exactly the years when delivery companies began burning cash to buy the market.

In 2017, as the market steadied and the big players pulled back their subsidies, the numbers for physical retail started to recover.

A report from the Meituan Research Institute says restaurant closures nationwide in 2017 came to 91.6% of openings. For every 100 restaurants that opened, about 92 closed; Sichuan restaurants alone fell by 40,000 in a single year.

This reshaping brings new market rules. Users have far more choice: on a phone they can scroll through dozens or hundreds of restaurants instead of only the ones in sight down the street.

Merchants can cut their store costs a little too — you have surely seen a delivery-only restaurant brand tucked into some out-of-the-way spot.

Because delivery breaks the limits of a restaurant's floor space and location, a new restaurant in theory gains a chance to grow. But when people browse restaurants on a phone, which one they pick depends entirely on how visible it is, and that weakens brands.

To pull customers back, a restaurant needs better food and also advertising fees, so users keep seeing it. In that contest the merchant is plainly the weaker side.

People Still Use Meituan and Dianping to Find Places to Go, but the Revenue Comes More and More From Funneling Traffic to New Businesses

Meituan's delivery business carried 171 billion yuan of transactions over 2017.

Group buying, which began as a way to display restaurant information, looks more and more like a vast traffic gateway for Meituan.

In the prospectus, Meituan says 80% of the new users of its hotel-booking business in 2017 came from people who had used its delivery and in-store dining services. Those new users helped Meituan's hotel business sell 205 million room-nights over the year, second only to Trip.com Group.

But Meituan does not disclose the average price of a hotel booking, and it sidesteps the question of how much revenue those 4.4 million merchants actually earn through the platform.

How much difference Meituan makes to a merchant's income is exactly what Meituan will not discuss. E-commerce platforms — Alibaba, JD.com, even Amazon — are quite happy to say how much business they have brought third-party sellers.

Meituan and JD.com have similar cost structures. Meituan is valued at about $60 billion, close to JD.com's $58.3 billion. In 2017 JD.com's transactions passed 1 trillion yuan and it netted 5 billion yuan; Meituan's transactions came to 357 billion yuan and it lost 2.8 billion.

Meituan's New Bike-Sharing Business Has Already Broken the Bicycle Industry's Offline Network

In the "Risk Factors" section of the prospectus, Meituan says: "We have incurred substantial losses historically and may continue to incur substantial losses in the future."

And right after that: "Mobike, which we acquired in April 2018, has incurred losses since its inception. We cannot guarantee that Mobike or our business as a whole will achieve profitability in the future."

Bike-sharing may not make money, but it has also cut the money the bicycle industry makes.

Once shared bikes filled every street and lane, people no longer needed a bicycle to commute on. After ofo and Mobike moved into most first- and second-tier cities, chain bicycle stores — Giant, Xidesheng, Merida — closed in large numbers; one brand reportedly shut "more than a thousand" stores in a year.

Giant wrote the phrase "affected by bike-sharing" straight into its annual report. In 2017 its output fell 10.5% from the year before and its sales fell 14% — the old bicycle maker had to cut prices to hold its volume. Its factories ran at 82% of capacity, down from 93% a year earlier.

Giant never won a bike-sharing order. The bicycle companies that did won a crisis of their own.

No industry has ever gathered such enormous capital so fast. ofo raised more than $4 billion in all, and Mobike raised about as much before Meituan bought it — all in a little over two years.

Mobike's supplier is Foxconn, which said in mid-2017 that it had reached an annual capacity of 36.5 million bikes, close to half of world production.

In May 2017 the veteran manufacturer Shanghai Phoenix took an order from ofo for 5 million bikes.

"It was like a flood — however many I could make, they wanted," said the owner of one contract manufacturer. Parts makers were "taking orders until their hands went numb."

ofo became Phoenix's largest customer, and making cheap bikes for it brought Phoenix a lot of revenue fast. Phoenix's 2017 results show revenue of 1.428 billion yuan, up 126.63% year on year. Bicycle manufacturing and sales made up 1.142 billion of that, 80% of the total, and nearly 60% more than the year before.

Phoenix earned 8 yuan of profit on each shared bike it built. Not much, but it added tens of millions of yuan of profit.

The good times were short. The bike-sharing bubble burst just as fast, and Phoenix's first-quarter 2018 results show revenue of 164 million yuan, down 42.75% year on year.

ofo had promised to buy 5 million bikes over the year and took delivery of only 1.78 million; the rest fell through. By then Phoenix had hired workers, opened production lines, and bought equipment for the big order. In the end the bicycle company had to eat that investment itself.

And that is only ofo. There are plenty more second-tier bike-sharing companies that never settled their bills and cannot return users' deposits. Bike-sharing brought the bicycle industry some sales, and in the end it was still a blow.

Bike-Sharing Eased Congestion and Made Getting Around Easier, but as the Cash-Burning Slowed, New Problems Appeared

At the start, bike-sharing, like ride-hailing, told a story about a new form of transport.

In 2017 ofo said its bikes had cut carbon emissions by more than 2.16 million tons in total, saving society 615.15 million liters of gasoline and cutting PM2.5 emissions by 1.035 million tons.

Mobike said that after its bikes appeared, 82% of Guangzhou residents gave up unlicensed cabs and Shenzhen's public-transit blind spots shrank by 99.34%. Across the 50 cities where Mobike put out bikes, it said, the share of trips made by car fell by more than 3%.

A report last April from Alibaba's Amap said trips under 5 kilometers fell 3.8% in Beijing and 3.2% in Shanghai; shared bikes had taken the place of other transport over the "last 5 kilometers," and congestion on the worst roads had eased somewhat since they appeared.

Bike-sharing looks like a change for the better. But with no detail on how any of it was calculated, the real effect is hard to measure.

As the investment dried up, meanwhile, broken bikes left lying in the street began to eat into public space. People have grown used to sidewalks packed with shared bikes. Bike graveyards have appeared on the edges of every large city.

Starting late last year, Beijing, Shanghai, Guangzhou and other cities banned bike-sharing companies one after another from putting out new bikes. By one estimate, close to 20 million shared bikes were deployed over 2017, which will become nearly 300,000 tons of scrap metal once they are junked.

The bike-sharing business model never counted the costs it might impose on city management, road traffic, and public safety. Losing money is one thing — but if Mobike ever does turn a profit, plenty of taxes and rules are waiting up ahead.

Until then, society carries the cost of bike-sharing's problems.

ofo began as an idea for collecting abandoned bikes on campus and putting them back to use. Mobike began as an idea for a shared bike that needed no maintenance for four years. Both were once clever answers to real social problems.

Under round after round of capital, all of that changed.

Meituan's Eight Years: Startup and Capital Remaking Each Other

When Meituan was new, it was not a company that liked to solve problems with money.

Meituan launched in March 2010, and Wang Xing then described it as a "new kind of group-buying site": one item a day, such as a 100-yuan wine-tasting set sold at half price. While rivals were already burning cash on advertising and street-level promotion, Wang Xing had not yet raised new money and insisted on doing everything "in-house" — visiting merchants one by one, running a single deal a day but guaranteeing its quality, sending traffic to one merchant at a time.

By the time Meituan raised its Series B, it had not spent all of the Series A.

But capital pushed it forward in the end. Four years of burning cash on group buying ended with Meituan paying an enormous sum to merge with Dianping and taking over the market. Capital was behind that.

After that, to support its huge valuation, Meituan could not stay in group buying, so it went into food delivery.

Then it burned cash in delivery for three years. It got its costs under control, but going by the prospectus, that thin profit is not the number you accept after burning 14.2 billion yuan.

Meituan raised $4 billion in October 2017. Two months later it announced it was entering the ride-hailing market.

Didi, likewise, burned cash on ride-hailing for three years and then moved into Meituan's delivery business. In April 2018 Didi said it was piloting food delivery, handing users 100-yuan coupons outright in cities such as Nanjing — enough to bring lunch down to 3 yuan.

Both companies built their core defenses out of capital, not technology. When growth ran into trouble, each threw itself at the other's market.

Enormous young companies like Meituan and Didi are also changing where capital goes: more and more money concentrates in industry leaders like Meituan, force-growing them and then finding Tencent or Alibaba to take a stake in support.

It has become harder for small companies to raise anything. SoftBank once put a few tens of millions of dollars into Alibaba and turned it into tens of billions. Today a few tens of millions is nowhere near enough to carry a new company.

Figures from the ChinaVenture Research Institute show China's investment volume hit $9.436 billion in April 2018, more than four times the year-earlier level. Over the same month, though, the number of companies funded fell from 256 in April 2017 to 240. The money is piling up in big companies.

Internet companies chasing growth and valuation, unable to slow down or stop losing money, is not only a Chinese problem. Most high-valuation Silicon Valley startups are the same, to the point that a dog-walking app can raise $300 million in a single round.

Masayoshi Son's $100 billion SoftBank fund and Sequoia's new $8 billion fund are both answers to this.

Companies grow by burning cash; capital gets bigger and more concentrated. This is what happens when companies and capital change each other.

Once It Entered Ride-Hailing, Meituan Had No Choice but to Go Public

From group buying to delivery, Meituan has copied the same story: spend a mountain of money to build a market habit, then own the market.

The entry barriers in these businesses are relatively low. There is no need to build warehouses across the country the way JD.com did, or to spend years building data centers for cloud computing the way Amazon did.

There is not much to say about how the businesses connect to one another, though the prospectus does try to show that delivery and reviews send traffic to hotels.

Judging by the move from group buying to delivery, how strong those links really are is hard to say. Meituan had years of experience with its "ground-troops" sales force, and it was working inside the same restaurant industry, and it still had to start subsidizing from zero, losing 223.7 yuan for every 100 yuan it took in.

Ride-hailing went the same way. Late last year Meituan promoted ride-hailing on the home pages of both its main apps — in six cities, Beijing, Shanghai, Hangzhou, Xiamen, Fuzhou, and Wenzhou, it promised to launch the service in any city where 200,000 people "signed up." Meituan, which claims more than 300 million users a year, still needed two full days in Beijing and in Shanghai to get 200,000 people to open the page and tap it once.

That people who order delivery on Meituan will therefore hail cars on Meituan was strange logic to begin with.

But entering ride-hailing did change some things — for one, it made Meituan go looking for money in the public markets.

After round after round of cash-burning and merging, Didi has taken money from nearly every source in the market: Alibaba, Tencent, SoftBank, China Merchants Bank, Ping An Bank, Temasek and others. Those investors are unlikely to put large sums into Didi's rival. Once capital has billions riding on Didi, it will not also fund Meituan to attack Didi and endanger its own investment.

Nor is it easy for Meituan to borrow from banks. As of April 2018 it had borrowed 917 million yuan in total. One loan of 162 million yuan required Meituan to pledge a $250 million wealth-management insurance product as collateral — a book value roughly 10 times the loan.

Meituan still holds 45.2 billion yuan in cash, but ride-hailing is a bottomless pit. The prospectus shows that, against Didi's huge subsidy counterattack, Meituan spent nearly 900 million yuan on sales expenses including subsidies in the ten months after its ride-hailing service launched in Nanjing in 2017.

That is ten months in Nanjing alone, a second-tier city of just over 8 million permanent residents. Beijing and Shanghai each have roughly three times as many, by conservative official counts.

If Meituan wants to fight Didi for market share across dozens of second-tier cities, 45.2 billion yuan is not enough.

Meanwhile Tencent's WeChat, Baidu Maps, and Alibaba's Alipay all give Didi a prominent place to hail a car from. It is hard to imagine any of them swapping Didi out for Meituan.

For Meituan, listing is a chance to raise money and keep fighting Didi. And there are many other places that need money: Alibaba is still there in delivery, and ofo, too, may end up bought by a competitor.

Meituan also plans to move into physical supermarkets. The prospectus's list of intended investments includes an unnamed "Company D"; from what is known so far, it is a listed supermarket chain in which Meituan will take a stake of no more than 5%. Meituan's own Xiaoxiang Fresh opened last month to compete with Hema — nearly a year after Zhangyu Fresh, its first offline fresh-food supermarket.

Among the many companies Tencent has invested in, Meituan is an unusual one.

WeChat's "Wallet — Third-Party Services" panel now carries four Meituan services. Yet Meituan is a rare Tencent-backed company that does not lean on Tencent's traffic: the prospectus shows 89% of delivery transactions come through the Meituan, Meituan Waimai, and Dianping apps, and less than 11% through WeChat and QQ. By comparison, Tongcheng-eLong, which listed recently, gets two-thirds of its users from Tencent.

Squeezed between rival pools of capital, Meituan now competes with several of the big companies Tencent has backed. That map did not exist a few years ago; it appeared as Meituan grew and pushed into hotels, ride-hailing, and other fields where Tencent already had investments.

A person close to Meituan told QDaily that data from its ride-hailing service in Shanghai shows the business clearly pulls new users onto the platform, and that converting those riders into users of other services works quite well. The company does not see its ride-hailing subsidies as money poured into a new business alone; they also bring fresh traffic to the businesses it already has.

Wang Xing has explained it himself: Meituan wants to be a "star," because "a meteor burns out and that is that; a planet lasts a long time, but it does not shine on its own... only a star shines by nuclear fusion, so a star has to be big enough."

In its years as a fast-growing internet company, Meituan built a relatively efficient organization, which is why it came out on top in several heavily subsidized industries. But the money burned along the way also means Meituan has to keep chasing bigger businesses.

The path of a company like this may show what eight years of capital-driven internet entrepreneurship really produced: the Amazon-style platform Wang Xing talks about, another Taobao covering every service in daily life, or a collection of enormous bubbles in China's internet.

This trip to list in Hong Kong is not the end of the story, only the start of the next round of competition.

Cover image from Visual Hunt