为何大数据会扼杀企业

大数据被很多人吹捧成了大企业的救星：有人说它能预言未来，照亮我们的道路，给古老的商业模式带来新的生机。但是在现实世界中，数据是会杀人的。它能杀死项目，杀死金钱，甚至杀死时间。25年前，数据的增长速度大约只有每天100GB，而现在，数据的增长速率差不多已达到50,000GB每秒。随着数据量的海量增长，企业也越来越难以凭借自身的能力进行数据分析，从而加大而不是减小了企业战略决策的难度。

时间是我们最宝贵的资源，而数据偷走了我们大量宝贵的时间。我们的感观早已被各种各样的数据淹没。每天我们都会收到数不清的电子邮件、手机短信和提醒消息，每一条信息都会让人分心，降低我们的工作效率。它们将我们抽离了原本该做的事情，迫使我们将注意力放在也许重要、也许不重要的事情上。同理，企业的业务数据也同样多得令人窒息，牵扯了我们的大量精力，已经成了影响企业高效决策的拦路虎。

不妨想象一下，如果有一天，你只会收到对你来说真正重要的信息，而且这些信息还能在正确的时间、在正确的地点找到你，世界将是什么样子。那么你每天至少能多做多少事情？我们将大量的时间耗费在被动消化这些海量信息上，真正用来主动谋划企业发展的时间少之又少。这样既令人心力交瘁，又削弱了企业效能。

更重要的是，数据会令企业丧失精准度。光靠捕捉更多信息并不会自动使企业产生更多价值。有人可能会想，我们收集的数据越多，就越能从中获得好的见解。这种自欺欺人的心态是很危险的。只有当数据能带来准确而重要的见解时，它才是好的数据。

另外，只有与你息息相关的信息才是有用的信息。好的信息必须具备时效性和真实性。然而不幸的是，当企业想从大数据中提取有用的见解时，却经常会起到反效果。举个真实的例子，美国有一个叫麦克·西伊的人是办公用品超市OfficeMax的常客，他的女儿不幸和男友死于一场车祸。OfficeMax不知怎么得知了这个消息，在发给麦克·西伊的自动促销邮件中竟然出现了这样的抬头：“麦克·西伊（女儿死于车祸）。”这并非大数据有意作孽，而是它的相关性（和适宜性）的问题。一个企业要想只收集其确实需要的数据几乎是不可能的，很多时候你收集到的是那些原本不该看到的东西。对于一家公司来说，你收集到的数据很可能是误导性甚至是毁灭性的。大数据虽然能将很多不相关的点连接起来，呈现一幅完整的图画，但是要确保数据的相关性、及时性和真实性，你首先还要正确理解它的背景。

现在，全球每天的数据总量都能达到250万的三次方字节，要想通过大数据获得全面的见解是很难的。你要么会陷入无力分析的境地（因此无法获得见解），要么就更糟糕，你可能会在有限的甚至是被错误解读的数据基础上获得错误的见解。如果没有正确地理解数据的背景，将不啻于椽木求鱼。一些看似有希望改变游戏规则的见解，在实际中却很有可能导致你从游戏中出局。

数据也会扼制你的灵活性。传统的数据分析方法，是将交易系统中的所有数据存放到一个数据仓库里（也有的叫数据湖或数据池），然后运行几套业务智能系统，叫几个或十几个分析师分析上一周的时间，然后把数据导到Excel里，或者做一个PPT。周而复始，得到的见解始终是滞后的。这种数据处理方法其实是一种浪费。由于要处理的数据很多，你得需要很长的时间才能获得有用的或是有可操作性的见解。你需要找到一种透过能繁杂的数据，得到为你的公司量身定制的信息的方法。

当我开车进城的时候，我想知道路上的交通堵不堵，需要多久才能达到目的地。如果有人给我的建议跟我同事上次开车走这条路时一样准确，那我就会不那么依赖GPS应用了。Waze就是这个领域的一款非常强大的应用，因为它截取了所有司机的一个巨大的时间断面的信息。这种全球数据的集中化使得所有用户都能获得与背景环境相关的见解。大数据也需要采取类似的做法。企业现在应该停止在自己公司的范围内积攒业务数据了，而是应该真正利用云计算的规模经济效益，不仅仅做到基础设施与应用的共享，更重要的是做到数据的共享。

如果你想将大量数据变成有价值的见解，你就应该利用一个集中化的全球性平台，因为这样一个平台可以借助大量内部和外部资源消化海量信息。企业将数据收集、管理和分析工作外包出去，就可以使这种通用平台专心研究数据科学，而你只需要集中精力，将它为你量身打造的见解应用在提高企业核心能力、强化企业竞争优势上。

20年前的一场“无软件”运动将世界从线下带到了云端。而今天，我们也需要掀起一场“数据有罪”运动。现在已经到了从收集数据转向让这些数据切实发挥作用的时候了。这将的话，在别人还在空谈“大数据”或疲于内部业务智能项目的时候，我们就能够解放精力进行创新。（财富中文网）

本文作者Nader Mikhail是Elementum公司的创始人兼CEO。

译者：朴成奎

Big data has been anointed the savior of big business: it divines the future, reveals our path, and breathes new life into our venerable business models. But in reality, data kills. It kills projects, it kills money, and it kills time. Twenty-five years ago, data was growing at a rate of 100GB a day. Now, data grows at a rate of almost 50,000GB a second. And as the volume of data grows, the ability of companies to make sense of it diminishes, confounding rather than illuminating strategic decisions.

Time is our most valuable resource, and data drains it. We are on sensory overload. Every one of the thousands of emails, text messages, notifications, and alerts we receive daily are a distractionand therefore kills productivity. They inherently take us away from what we’re doing and force our attention to issues that may or may not concern us. In the same way, our business data is overwhelming and distracting us—throwing up barriers to productive decision-making.

Imagine a world in which every piece of information you receive would not only be relevant to you, it would find you at the right place and right time. How much more would you be able to get done every day? We expend massive amounts of energy just trying to keep up with all this information, leaving little time or energy for us to actually move the needle for our organizations. It’s overwhelming, and it’s crippling.

What’s more, data kills accuracy. Capturing more data will not automatically generate more value for a company. The more we collect data, the more we convince ourselves that we will be able to glean good insights from it. This modern take on the sunk cost fallacy is corporate quicksand. Data is only good when it results in accurate and relevant insights.

To be useful, information has to pertain to you, it has to be timely, and it has to be true. Unfortunately, when it comes to gleaning insights out of big data, the odds are stacked against you. Take for example the OfficeMax coupon that was addressed to “Mike Seay, Daughter Killed in Car Crash.” It’s not the quality of data that lies at the source of the blunder, but it's relevance (and appropriateness). It’s virtually impossible to collect only the data you really need—and therefore, you are much more likely to be using data that you shouldn't. Data that, in the context of what you’re trying to do, is mistaken or even damaging. Big data is good for connecting dots that would otherwise go unconnected. But in order for information to be pertinent, timely, and true, you need to understand its context.

And with 2.5 quintillion bytes of data accumulating every day, the likelihood of achieving a broad purview is low. You will either fall victim to analysis paralysis (and therefore, never unlock insights), or worse, you will glean false insights based on limited or misunderstood data. Without context, you run a high risk of chasing red herrings. Insights that seem game-changing can, in reality, be game-ending.

Data also kills agility. The traditional approach: suck all the data from your transactional systems into a data warehouse (or data lake or data pond), slap a few business intelligence systems on top, throw a few (dozen) analysts at it for a week, and dump everything back into Excel and Powerpoint. Rinse, repeat, and continue to fall behind. This type of data processing is a waste. With so much data to handle, it takes way too long to get any useful or actionable insights. There’s simply too much irrelevant data sitting between you and your decisions. You need to find a path through all that data to receive information that is tailored and customized to your business.

When I get in my car to head to the city, I want to know if there’s traffic on the way and how long it will take to get to my destination. I’d be a lot less inclined to use GPS apps if the recommendations were only as accurate as the last time one of my co-workers drove that route. An app such as Waze is powerful because it pools information from a large cross-section of all drivers. This centralizing of global data allows for contextual insights that benefit all users. Big data requires a similar approach. It’s time to stop accumulating business data within the four walls of your company and to start taking advantage of the true economies of scale of the cloud: not just shared infrastructure and applications, but shared data.

If you want to turn data points into valuable insights, you need to leverage a centralized, global platform that can ingest information from a multitude of internal and external sources. Outsourcing all this data collection, management, and analysis will allow this common platform to focus on the data science, while you focus on applying its tailored insights towards strengthening your core competencies and sharpening your competitive edge.

Two decades ago, there was a “No Software” movement that took the world from on premise to cloud. Today, we all need to embrace the “Data Kills” movement. It's time to transition from collecting data to making it useful. It will free us to innovate while others are tangled in internal business intelligence projects, drowning in their own data lakes and "big data" water cooler prattle.