推荐序1

If companies were people, then we would be in the middle of one of the greatest health crises of the modern age. Once, the lifespan of a company on the Fortune 500 index of large businesses was 65 years. Today, it's only 20.In the last decade, most of the world's big, reliable firms have been displaced by digital upstarts: Apple, Amazon, Tencent, Google, Baidu, and Alibaba.

It gets worse. The chances of a company reinventing itself are low. The Corporate Strategy Board says efforts at digital transformation fail 95% of the time; Clayton Christensen,author of The Innovator's Dilemma,puts the number at 99%.

But there's some good news, too. Because technology has given us the ability to measure everything, accurately, better than ever before. We can know ourselves.

A 2011 MIT studyfound that companies that use data-driven analytics instead of intuition have 5%-6% higher productivity and profits than competitors. Over a few years, data and analytics is the difference between success and obscurity.

Data, it is often said, is the new oil. Data replaces opinions with accuracy, letting us know our customers, our suppliers, and ourselves with unprecedented clarity. And data is the food of artificial intelligence, because it's how we train machine learning algorithms.

On its own, oil isn't very useful.It just sits in the ground. To put oil to work takes an ecosystem: Refineries, gas stations, motors, regulations, roads, and more. And so it is with data. Simply collecting it won't help you; you need to extract it, clean it, analyze it, execute on what you learn, and feed that learning back into your systems.

As technology replaces many traditional tasks through automation and machine learning, we may wonder what is left for humans to do. The answer is simple: Think critically about what we want those machines to do for us. The most important skill for a human, whether they're a startup, an analyst, or a manager, is to ask the right question.Asking good questions is harder than it seems.It requires an understanding of the existing business model, the competitive landscape, and the resources at your disposal. But it also requires that we know that the existing business model is outdated, vulnerable, and ready for change.

A world powered by real-time information creates two roads. One road is littered with the bodies of companies that couldn't make the transformation, preferring anecdote over fact. The other road is paved with the profits of those who learned to harness data and embrace analytical thinking.

You're at a fork in this road. And right now, you're holding the map that will steer you down the right path.

如果我们将企业比作人类，那么许多企业正处于壮年的巨大健康危机之中。曾经，世界500强企业的生命周期是65年，而现在仅有20年。近10年来，众多规模大、可靠的企业已被“数据新贵企业”所替代，例如苹果公司、亚马逊、腾讯、谷歌、百度、阿里巴巴等。

更糟糕的是，企业进行自我重塑的概率变得越来越低。公司战略委员会指出，95%的企业数字化的转变是失败的。《创新者的窘境》一书的作者克雷顿·克里斯滕森认为这一数字已达到99%。

当然也有好消息：科技赋予我们衡量一切事物的能力，我们能更好地认识自己。在这点上，曾经的任何时代都难以企及。

麻省理工学院的一项研究表明，相比依靠直觉来实现决策的企业，那些通过数据驱动实现决策的企业拥有更高的生产效率和利润。这类企业的生产效率和利润普遍高于竞争对手5%～6%。显然，未来是否拥有数据分析能力，将决定一家企业是成功，还是逐渐销声匿迹。

我们经常说，数据是新石油。数据的准确性代替了“意见”的主观性，让我们更好地了解我们的供应商、我们的顾客以及我们自身。同时数据也是人工智能的基础，因为我们正是通过数据的运用来实现机器学习的。

对石油来说，一直被埋藏在地下的石油并无价值。它的价值在于应用，石油开采需要一个“生态系统”：炼油厂、加油站、汽车、规则、道路等。数据也是如此，仅仅收集数据并无价值，你需要提取、清洗、分析，让分析结果得以执行与运用，并反馈至“生态系统”中。

随着自动化操作和机器学习代替了部分传统工作，我们为此很疑惑：还有哪些工作需要人类来做？答案其实很简单：我们需要辩证地思考究竟人类需要机器来做什么。无论是初入职场的新人、分析师，还是企业管理者，提出正确的问题是他们最重要的能力。

但是，这实现起来很难。提问者既需要了解企业当前的商业模式、竞争格局以及可控资源，也需要意识到现有商业模式已经变得过时、不稳定，而且亟待改变。

信息随时随刻在产生，它为世界指出两条路：一条路布满着那些故步自封、因循守旧企业的“尸体”；另一条则为拥有数据思维和掌握数据驾驭能力的企业铺就康庄大道。而此时此刻，你正处于交叉路口，手中恰好握着一张指引正确路径的“地图”。

Alistair Croll

哈佛商学院访问执行官，Coradiant公司联合创始人

《精益数据分析》一书作者