Uber为什么成不了垄断者？

原文来自Tiller Partner的合伙人Tory Green。

Tory认为：Uber不可能成为像Facebook等互联网公司一样的垄断公司，原因在于：1)Uber的网络效应是供给端的，靠规模可以被打破 2)Uber缺乏互联网公司的零边际成本效应

当市场前三的 Sidecar 今年早些时候关闭并出售其资产给通用汽车以后，很多人认为共享出行市场的战争似乎结束了，普遍的看法是 Uber 注定要成为一个市场垄断者。

但是最近， Lyft 和滴滴获得了数十亿美元的投资， 很多人已经质疑这种看法，并开始思考：这个市场是否可以有多个玩家?

纽约客今年早些时候有一篇文章认为技术驱动的行业无法支持多个玩家，硅谷竞争趋势都是指向一个垄断的赢家。

造成这种情况的主要原因来自网络效应经济 — 即一个产品或服务的价值随使用人数的增加而增加。网络效应在 Facebook 、 eBay 和 Skype 等公司随处可见，每一个都是他们自己领域的垄断者。

互联网垄断企业存在的第二个原因是零边际成本分布 —- 提供单位额外的商品或服务不会增加总的成本。除了增加带宽有一点点可忽略不计的成本外， Google 和 Snapchat 服务一个新用户的成本基本为零。

这两种力量在一起，形成良性循环。一旦服务流行起来，它创造更多的消费需求。由于提供服务的成本在很大程度上是零，很容易吸引新的用户使用服务。这些新增的用户使产品更有价值，反过来，吸引更多的用户。如此循环，使互联网成为不公平竞争的滋生地。

提及科技公司，法学教授Tim解释说： “ 从长远来看，竞争是异常，垄断才是常态 ” 。这就是为什么我们看到互联网领域这么多近乎垄断的公司，如 Facebook ，谷歌，维基百科， LinkedIn ， Craigslist ，亚马逊和 Twitter 。事实上，乍一看，技术行业真的是 “ 赢家通吃 ” 的市场。

这种观点很大程度上支持了为什么Uber的估值是其竞争对手Lyft的12倍，而其收入只有Lyft的4倍。

随着近期 Lyft 和滴滴融资总额超过 20 亿美金，似乎投资者并不准备认输，这或许是明智的，上面讨论的支持 “ 赢者通吃 ” 理论的两个关键点上Uber都有缺陷。

如同Quartz网站的文章指出的， 第一个错误在于不理解有两种不同的网络效应。

第一种网络效应被称为“需求端网络效应”。在需求端网络效应中，每新增加一个用户，产品或服务的价值会因为这个用户的增加而直接增加。 Facebook 就是一个完美的例子—你加入 Facebook ，因为你所有的朋友都在那里。从经济角度来看，需求方的网络效应是游戏改变者，因为即使一个小小的竞争优势也能迅速滚雪球成一个根深蒂固的竞争优势。

另一种类型的网络效应被称为“供给侧网络效应”。在供给侧的网络效应里，产品或服务的使用增加，对用户的直接效用没有影响，但它产生有价值的免费产品和服务。这方面的一个很好的例子是手机运营商。一个运营商的用户越多，他们越可以负担得起基础设施的投资。基础设施越好，服务质量就越好。在这种情况下，用户的数量是间接地影响客户的选择：人们加入一个运营商不是因为他们的朋友都在那里，而是因为哪个运营商的服务最好。

Uber的例子也是同样。公司的竞争优势主要来自于路上的司机数目和经营城市的数量。这两者都是供给侧的好处。

不过，虽然供给侧的网络效应有自己的优势，这不是一个无法战胜的竞争优势。在这种特殊情况下，挑战它的方法是采购规模 — 吸引更多的司机，开到更多的城市，等等 ...

竞争对手如Lyft，正在做这样的事情，而且已经融了20亿美金。

事实上，Lyft的总裁 John 认为，“在运输行业，尤其是我们的业务，有很强的网络效应，但只是一定的程度上。 ”具体来说，John发现，“一旦你能达到3分钟内接客，网络上有更多的人并没有什么额外的好处。”

“赢者通吃”理论的第二个错误是，与许多科技公司不同，Uber并没有零边际成本的效应。虽然公司的商业模式已经消除了很多传统的出租车业务相关的成本，比如车辆和车辆维修、牌照、保险、汽油和司机的薪酬福利，边际成本仍然存在。

看一看 Uber 泄露的财务数据，我们发现，在 2014 年第二季度他们 “ 销售成本 ” 和 “ 运营和支持 ” 合计 4900 万美金，而对应收入只有5800万美金。传统的会计准则会把这些作为可变成本，这和额外计为固定成本的的 1.15 亿美金的营销成本、研发和一般开销是分开的。

虽然文件没有逐行列出费用的明细，但是很容易推测，这些费用的很大一部分都直接关系到他们的服务产品的拓展。例如，Uber在每一个新的城市启动时，它必须创建一个本地的团队来处理特定的政治、法规和消费者的喜好。Uber也有可能花很多钱获取新的司机，包括市场营销、奖励和奖金、筛选和背景调查。最后，Uber要为司机提供保险，这也是直接的边际成本。

而如果像加利福尼亚州一样的裁决 — 法官认定Uber的司机实际上就是员工—在全国范围内被采纳，那么边际成本可能会飙升。据一些分析人士，如果提供医疗、失业、工伤，工资税、 401K 、休假和报销行使的里程、汽油和过路费，边际成本会提高一个数量级，公司为每个司机要额外支付1.3万美金(即 41 亿美金一年)。

所以，虽然零边际成本是技术公司获得垄断地位的主要要素之一，它在共享出行行业里并不存在。

综上所述，投资者开始意识到硅谷所认为的共享出行市场是一个 “ 赢家通吃 ” 的市场这个观点有两个缺陷： 1 )没有需求端的网络效应 ; 2 )实际边际成本不低

但Uber远不是一个出租车服务公司， Uber可以做更多!

一些硅谷的专家会说，盯着出行共享市场是近视的，Uber将演变成一个通用的 “ 技术工具 ” ，将重新点燃网络效应的良性循环，一直达到全球垄断。

例如，许多业内人士认为，Uber可以创建一个全方位服务的 “ 城市物流网络 ” ，完全拥有配送服务令人垂涎的 “ 最后一英里 ” 。随着如 UberRUSH (快递)， UberEATS (送外卖)和 UberCargo 服务(搬运)服务的推出，该公司似乎在向这个方向前进。

但即使是Uber能够实现其拥有 “ 生活方式和物流的交汇 ” 的梦想，这是否改变底层的经济学原理呢?

Uber会立即获得需求端的网络效应吗?大多数人并不关心谁提供了服务，只要他们能准时赶到那里。所以这把我们带到了我们开始的地方 ...... 有规模优势，但他们来自供给端。考虑到市场的已有玩家如FedEx和 UPS 在这个市场的经验和规模，竞争可能不会像很多Uber多头相信的那样很快消失。如同FedEx首席执行官弗雷德 · 史密斯说： “ 我认为这只是一个城市神话，说Uber在某种程度上改变了物流行业的基本成本结构。 ”

Uber会立刻获得零边际成本吗?不会，它会面临如上面讨论的同样的成本结构问题。

还有人认为，自动驾驶汽车将会成为获得垄断地位关键，但即使这也不是什么万能药。如果Uber在这条路上继续前行，并且拥有这些资产，那么成本可能反而会增加。如果它采取 Lyft 正在走的路，把汽车的所有权外包给别人，某些成本仍然不可避免的出现在价值链的某个地方。用米尔顿 · 弗里德曼(注：诺贝尔经济学家)的说法来讲，天下没有免费的午餐，不管你用什么类型的车辆来运输都和零边际成本分布这个原理相悖 — 即使是太阳能供电的自动驾驶汽车。

一旦Uber扩大当前的的服务品类，竞争对手将有时间和空间来完事自己的物流框架，并且在Uber错过的地区建立优势。是不是FedEx， UPS ， Instacart ， Deliv ，亚马逊或 Lyf t无关紧要 — 真正的问题是仍有竞争的空间。

总之，技术的成熟和其他的衍生服务反而会为其他更多的玩家进入这个行业提供机会，而不是更少。

当Uber的口号从 “ 每个人的私人司机 ” 改变为 “ 生活方式和物流的交汇”，毫无疑问，它这么做是把自己定位成科技公司，就像 Facebook ， LinkedIn 和 Twitter 一样，认为自己可以很容易利用网络效应的良性循环来垄断市场。

但是，让我们认真的思考一秒钟：这是 2016 年，任何一家有机会获得长期成功的公司，这样或者那样，都是一家技术公司。如果你不利用技术，你就已经死了。

因此，尽管共享出行市场和物流可能已经进化到不再是严格的模拟市场，(比如，它不会经历历史上我们在酒店、航空公司和 / 或汽车租赁公司市场看到的竞争程度)，但它也不是严格的 “ 数字化 ” 市场。只要网络效应留在供给侧，只要扩展需要边际成本，只要监管的威胁仍然存在，那么Uber就不可能成为一个真正的垄断者，像一些硅谷内部人士希望的那样。

事实上，正如John指出的， Lyft “ 在所有排名前 20 位的市场正获得越来越多的份额 ” ，如果一个玩家有完全的垄断地位，这是不可能发生的。

因此，考虑到所有这些潜在的竞争 — 事实上，共享出行和物流市场都可能可以支持多个竞争对手，这对Uber有什么影响?是否值得700亿美金的价格?

好吧，我现在不会对这个发表意见，但我要说的是，很多投资者，甚至是出了名的保守的纽约大学教授 Damodoran ，似乎也这么认为。

但是，也许这一切里更重要的问题是：如果你相信Uber估值合理(或至少接近)，那么为什么 Lyft ，滴滴和 GrabTaxi 估值比它低很多?

毕竟，这些公司都建立了类似的基础设施建设， Lyft 在 Uber 的老巢旧金山控制了 40 %以上的市场份额，收入是其1/4而估值是其 1/12; 滴滴号称(注：因为愉悦资本投资了神州专车，为了保持中立，我只能呵呵一下，“号称”)在利润丰厚的中国市场拥有 87 %的市场份额，估值是Uber的1/3 ; ，Ola，在印度市场号称拥有80%的市场份额，估值才和 Lyft 差不多。

考虑到这些差异，也许是老派的价值投资在硅谷这个地方也有用武之地 …

译者：戴汨 愉悦资本创始合伙人(midai@joycapital.com.cn)微信公众号：ThinkingSlow。愉悦资本是新一代的VC基金，由刘二海、李潇、戴汨创立，我们是创始人也是投资经理;愉悦资本，创始人和创始人对话。

附英文原文：

When top 3 player Sidecar closed earlier this year and sold its assets to GM, many pronounced the ride-sharing wars over and the prevailing wisdom seemed to suggest that Uber was destined to become a monopoly.

But recent multi-billion dollar investments in Lyft and Didi Chuxing have challenged that perception and caused many to wonder whether this market has room for more than one player…

The Silicon Valley Argument

An article written earlier this year in the New Yorker argues that industries driven by technology can’t support multiple players, and that competition in Silicon Valley trends toward one monopolistic winner.

The primary reason for this stems from the economics of network effects – a phenomenon where the value of a product or service increases with the number of people using it. Network effects are seen in companies such as Facebook, eBay and Skype; all virtual monopolies in their field.

A second reason for the existence of internet monopolies can be found in the economics of zero marginal cost distribution – a situation where an additional good or service can be produced without any increase in total cost. With the exception of the minimal cost of increased bandwidth, it’s largely free for Google or Snapchat to host another user.

Together, these two forces create a virtuous cycle. Once a service becomes popular, it creates additional consumer demand. Since the cost of distribution is largely zero, it’s easy to attract and onboard new customers to the service. These additional users make the offering more valuable and, in turn, attract even more users. This cycle continues and makes the internet a breeding ground for unfair competition.

When it comes to technology companies, law professor Tim Wu explains: “over the long haul, competition has been the exception, monopoly the rule”. That’s why we see so many near monopolies in the space, such as Facebook, Google, Wikipedia, LinkedIn, Craigslist, Amazon and Twitter. Indeed, at first glance, it seems that technology industries may be “winner-takes-all” markets.

This view is likely a large part of the reason that Uber is valued at 12x more than its nearest competitor Lyft, despite only having 4x the revenue.

The Emerging Counterpoint

With recent investments in Lyft and Didi Chuxing totaling over $2 billion, it seems like investors aren’t ready to throw in towel yet, and that’s probably wise given that there are two key faults underpinning the “winner-takes-all” theory discussed above.

As pointed out in this article on Quartz, the first mistake is not understanding that there are two different types of “network effects”.

The first is known as a “demand-side network effect”. In a demand-side network effect, the value of a product or service is directly increased by each additional user solely due to the addition of that user. Facebook is the perfect example of this – you join Facebook because all of your friends are on there. From an economic point of view, demand-side network effects are a game-changer, as even a small competitive lead can rapidly snowball into an entrenched competitive advantage.

The other type of network effect is known as a “supply-side network effect”. In a supply-side network effect, increased usage of a product or service has no influence on the direct utility for users, but it spawns the production of valuable complimentary goods and services. A great example of this can be seen with cell phone carriers. The more users a carrier has, the more money they can afford to spend on infrastructure. The better the infrastructure, the better the quality of service. In this case, the number of users is indirectly influencing the customer’s choice: although people may not join a carrier because their friends are on there, they are likely to join the carrier with the best service.

And that’s the case with Uber. The company’s competitive advantage largely stems from the number of drivers they have on the road and the number of cities they operate in. Both of these are supply-side benefits.

But while supply-side network effects definitely have their advantages, this is not a competitive edge that can’t be overcome. In this particular case, the way to challenge it is by Purchasing scale – attracting more drivers, opening in more cities, etc...

With $2 billion in funding to date, that’s exactly what competitors such as Lyft are doing.

Indeed, Lyft President John Zimmer agrees that, “in a transportation business, specifically our business, there are very strong network effects, but only to a point.” Specifically, Zimmer found that “once you hit three minute pickup times, there's no benefit to having more people on the network."

The second mistake in the “winner-takes-all” theory is that, unlike many technology companies, Uber does not experience zero marginal cost distribution. While it is true that the company’s business model has eliminated a lot of the costs associated with a traditional taxi business, such as vehicles and vehicle maintenance, licenses, insurance, gas and driver compensation and benefits, there are costs to distribution that still exist.

Taking a look at Uber’s leaked financials, we see that in the second quarter of 2014 they listed $49 million in “cost of revenue” and “operations and support” against $57 million of revenue. Traditional accounting standards would assume that these are variable costs, which are separate from the additional $115 million in “fixed costs” of sales and marketing, R&D and general overhead.

While the document doesn’t break out expenses by line item, it’s easy to speculate that a large portion of these costs are directly tied to the expansion of their service offering. For instance, each time Uber launches in a new city, it has to create a local team to deal with the particular politics, regulations and consumer preferences of that environment. Uber also likely experiences substantial costs in the acquisition of new drivers, including marketing, incentives and bonuses, and screening and background checks. Finally, Uber now provides insurance to its drivers, and there’s a cost to that as well.

And if rulings such as the one in California - where a judge determined that an Uber driver was, in fact, an employee – gain national traction, then distribution costs could skyrocket. According to some anaLysts, offering health insurance, unemployment, worker’s compensation, payroll taxes, 401K, vacation time and reimbursement for miles, gas and tolls, could raise distribution costs by an order of magnitude, and cost the company an additional $13K per driver (or $4.1 billion per year).

So while zero marginal cost distribution is a key ingredient in the recipe for making a tech monopoly, it simply doesn’t exist in the ride-sharing industry.

In summary, investors are beginning to realize that the Silicon Valley view that ride-sharing is a “winner-take-all” market is flawed for two reasons: 1) the absence of demand-side network effects and 2) tangible marginal costs to distribution.

But Uber Has the Potential to be So Much More Than a Taxi-Service!

Some Silicon Valley pundits will argue that focusing on ride-sharing is myopic, and that Uber will evolve into an all-purpose “tech utility” that will re-ignite the virtuous cycle of network effects and pave the way to global domination.

For instance, many insiders believe that Uber can create a full service “urban logistics fabric” and completely own the coveted “last mile” of distribution. With services such as UberRUSH (courier), UberEATS (food delivery) and UberCargo (moving), the company seems to be heading in this direction.

But even if Uber is able to realize its dream of owning the “interdiv of lifestyle and logistics”, does this change the underlying economics?

Would Uber instantly gain demand-side network effects? Most people don’t care who delivers their packages, as long as they get there on time. So that leaves us where we started…there are advantages to scale, but they come from the supply side. And given that incumbents such as FedEx and UPS have both experience in this market AND scale, competition might not disappear as quickly as some Uber bulls would have you think. As FedEx CEO Fred Smith says “I think there’s just an urban mythology that [Uber] somehow changes the basic cost input of the logistics business”.

Would Uber instantly gain zero marginal cost distribution? No, it would face the same costs as discussed above.

Others believe that the emergence autonomous cars will hold the key to monopolistic power, but even that isn’t the panacea that some hold it out to be. If Uber continues on the path it seems to be heading now and takes ownership of these assets, then costs would likely increase. Even if it take the road that Lyft is travelling, and outsources the car ownership to someone else, there are still costs that are going to present themselves somewhere in the value chain. In parlance popularized by Milton Friedman, there’s “no free lunch” and delivery via any type of vehicle – even a solar powered self-driving one – is the exact opposite of zero marginal cost distribution.

If anything, as Uber aims to expand beyond its current offerings, competitors will have the time and space to perfect their own logistical framework and establish an advantage in the areas the Uber misses. Whether that’s FedEx, UPS, Instacart, Deliv, Amazon or Lyft is largely irrelevant – what matters is that there’s still room for competition.

In short, the maturation of technology and proliferation of options should serve as an opportunity for more players to enter the industry, not fewer.

Valley Value Investing?

When Uber changed its slogan from “everyone’s private driver” to “where lifestyle meets logistics” it undoubtedly did so to position itself as a “technology company” that, like Facebook, LinkedIn, and Twitter, could easily dominate its market with the virtuous cycle of network effects.

But let’s get serious for a second: It’s 2016, EVERY company with a chance of long-term success is, in one way or another, a “technology company”. If you’re not leveraging technology, you’re already dead.

So while the market for ride-sharing and logistics may have evolved to the point where it’s no longer strictly analog (i.e. it probably won’t experience the same level of competition that we’ve historically seen in hotels, airlines and / or rental car companies), it’s not strictly “DIGital” either. As long as the network effects remain on the supply-side, as long as there are costs to expansion and as long as the threat of regulation remains, then it’s unlikely that Uber will ever become a true monopoly in the way that some Silicon Valley insiders hope.

Indeed, as John Zimmer points out, Lyft is “gaining share in all top 20 markets” and that’s not what happens when one player has a complete monopoly.

So given all of this potential competition – the fact that both the ride-sharing and logistics market can likely support multiple competitors, where does that leave Uber? Is it worth the $70 billion price tag?

Well, I won’t opine on this right now, but what I will say is that a lot of its investors, and even the notoriously conservative NYU professor Aswath Damodoran, seem to think so.

But perhaps the more important question in all of this is the following: if you believe that Uber is fairly valued (or at least close to it), then why are Lyft, Ola, Didi Chuxing and GrabTaxi valued so low?

After all, these companies are all building a similar infrastructure, and Lyft controls over 40% of the market in Uber’s home turf of San Francisco and has a fourth of the revenues of Uber at 1/12th of the valuation; Didi Chuxing claims to control over 87% of the lucrative Chinese market and has 1/3rd of the valuation of Uber; and Ola claims 80% of the Indian market and has a value similar to Lyft.

Given these discrepancies, perhaps old-school value investing has a place in Silicon Valley after all…