桑迪来袭，Uber陷入舆论风暴

    就在我的老家纽约正从飓风桑迪的余波中逐渐恢复过来时，硅谷却掀起了一场完全不同的舆论风暴：大家纷纷发出质疑，电子叫车应用Uber在纽约的“峰时价格”（surge pricing）到底是漫天要价，还是在试图平衡供需。

    Uber是靠算法立家的公司，善于用数据和事实来回应舆情。不过，尤其是在灾害肆虐的非常时期，人们更想听到的是同情，而不是数据。我不觉得Uber会趁着飓风刚过就准备大发利市，但我确实认为，在以后类似的非常情况下，它应该考虑采取几个举措。

    Uber的企业文化中有些方面让我想起了早期的谷歌公司（Google）。Uber也十分推崇数据和算法。就像其首席执行官特拉维斯•卡兰尼克会强调的那样，Uber的秘密武器就是能把响应时间缩短到最小值的算法。能做到这一点，靠的是“路径规划”（routing，即司机应该赶到何处）和一个定价模型。需求大增的时段，比如节假日和周末的晚上，这个模型能让价格“迎峰上涨”（surge）。

    Uber的说法是，司机都是自由人。这点我觉得说的在理。如果他们私下接客能赚得更多，他们当然会为这些客人服务，而不会成为Uber的用户。但Uber显著提高了他们在高峰期的收入，因此激励司机成为Uber圈子里的一员，甚至让这些司机延长已经够长的工作时间，多拉几位客人（可以把这看成是加班费）。这就是供需规律怎么作用的原理，跟一个人是否读过《阿特拉斯耸耸肩》（Atlas Shrugged，兰德最著名的小说，是“继《圣经》之后对当代美国人影响最大的一本书”—译注）无关。

    Uber不是一种非盈利的公共服务。它是一家为愿意埋单的客户提供增值服务的公司。因此，它需要在其运营的城市遵循不断变化的各种规章制度。迄今为止，它已经在波士顿、华盛顿特区和芝加哥惹出过一些事端。这一点值得另文阐述，不过我的总体感觉是，有些管理工作是合理合法的，旨在保障乘客安全，维护公平的竞争环境。另外一些则纯属一些交运公司放出的烟幕弹，它们一直无需面对竞争，用差劲的服务也能混得不错。

    所以，作为一家私营企业，“峰时价格”主要是为了让Uber的用户在路上能拦到更多的车。而Uber自己则既能从中直接（因为他们能从车费中分成）受益，又能间接获益（提高其服务对司机和乘客的吸引力）。这是一个具有两面性的价值主张，Uber需要好好管理：司机需要觉得付给他们的车费够多，而乘客则要求获得的舒适度和速度物有所值。如果需求足够大，Uber就能成功。否则它就算失败了。我本人经常使用Uber服务，多数是用来在旧金山和纽约叫车从市区去机场。但还是先回过头说说桑迪吧……

    飓风刚过，纽约的Uber就开始执行一套旨在增加供应（司机）以满足需求（乘客）的措施。可能对Uber来说，它就没有一本“人类灾情应急措施”手册。所以灾难当头之际，它还是照老一套办。其中就包括了实行2倍的峰时价格。由于公众对这种漫天要价的做法大为不满，公司只得一边照旧付给司机2倍报酬，一边只收乘客1倍的费用，致使公司每天要多花10万美元的成本（实际上也就是补贴了市场）。随后它又恢复了峰时价格，不过表示不会拿自己那份分成——多收的钱也都归司机。

    While my hometown of New York City recovers from the aftermath of Hurricane Sandy, there's been a different verbal maelstrom here in Silicon Valley: The question of whether Uber NYC's "surge pricing" is gouging or simply an effort to balance supply and demand.

    Uber's an algorithm-driven company which responds to emotion with facts. However, especially in times of distress, people want to hear empathy, not data. I don't believe Uber set out to maximize profits in the wake of a hurricane, but I do think there are several steps they should consider in future extraordinary situations.

    There are aspects of Uber's culture which remind me of early Google (GOOG). Uber believes in data and algorithms. Its secret sauce, as CEO Travis Kalanick will note, are algorithms that attempt to minimize response time. It does this through routing (where should drivers be) and a pricing model that allows prices to "surge" during periods of intense demand such as holidays and weekend evenings.

    The company's argument, which I believe to be valid, is that its drivers are free agents. If they can get better rates with their private clients they will seek to service that market and not Uber patrons. Uber effectively increases their take-home pay during surge moments in order to encourage a driver to work within the Uber pool, or even extend their already long day to pick up a few more passengers (think of it as the price of overtime). This is how supply and demand works, regardless of whether you've ever read Atlas Shrugged.

    Uber is *not* a nonprofit public service. It is a company selling a premium service to customers who choose to pay. As such it will need to play within the evolving rules and regulations of the cities in which they operate. To date this has caused some issues in Boston, DC and Chicago. It's worthy of a separate post, but my general feeling is that some of the regulatory efforts are legitimate and aimed at passenger safety and a level playing field. Others are purely smokescreens from incumbent transport companies who have been able to get away with providing crappy services in the face of no competition.

    So being a private company, of course, surge pricing is generally about getting more cars on the road for Uber customers. And Uber benefits from this directly (since they take a cut of fare) and indirectly (strengthening the general appeal of their service for drivers and riders). It's a two-sided value proposition that Uber needs to manage: Drivers need to feel like they're getting paid enough and passengers want enough comfort and speed for their dollars. If there's enough demand, Uber succeeds. If not, Uber fails. Personally I use Uber occasionally and mostly for city-to-airport travel in SF and NYC. But back to Sandy....

    In the storm's wake, Uber NYC implemented a set of decisions meant to increase supply (drivers) to meet demand (passengers). Uber probably didn't have a "what to do in a human tragedy" playbook and instead ran its normal operating procedures. This included putting 2x surge pricing into effect. In response to public outcry over gauging, it continued to pay drivers the 2x but charge passengers 1x, costing the company $100k/day (effectively Uber was subsidizing the marketplace). Then it put the surge back in place, but said it wouldn't take their share of profits -- all money would go to the drivers.