究竟什么是大数据,如果它不是关于大,也不需要考虑数据?

很多人问，为什么没有鉴于庞大的媒体接收出现关于大数据的公认的定义。鉴于瞄准这一行列（与创业投资的创业）全球IT厂商在营销上所花费的努力，他们对此产生疑惑也是可以原谅的。

就像大部分的消费者营销好奇地的重点是青少年买家的购买力只是他们的父母和祖父母的一小部分，大数据和它的推动者也是一样。

他们用流行语向他们的传统客户—-IT部门展示，而未能参与（事实上，完全疏远）业务经理的业务问题，这些问题通过分析可以解决在并且能直接控制预算。

这是为什么，这对采用大数据有什么意义？

IT部门的目标源于懒惰和极客。懒惰在这个意义上来说是指，除了极少数的例外，这些大公司就不需要直接关注企业管理者的需要，他们的IT部门会处理这个转化工作。

而极客在这个意义上来说是指，较小的专业厂商是由技术人员来进行管理的，因此他们的市场讯息 （背景的原因）也往往集中在“速度和补给”上，而不是在解决实际业务问题。

这对大数据项目有着负面后果，并且风险将成为减少“新油”吸引力的沙子。

大量的炒作大数据的承诺的文章已经出现在媒体之中，这既提高了企业管理人员的期望，同时也使他们感到困惑。

厂商希望IT部门率先启动大数据项目是在自欺欺人。今天的IT部门的状况如下：（一）日复一日的工作使其过于忙碌（BYOD，移动接入，安全等）；（二）引导他们选择有前途的领域进行引航分析和职能经理关注的领域相差太远；（三）不熟练的数据科学家。

所以，回到了何谓“大数据”的问题，其实并没有公认的定义 ，然而，它不是’大’，它不是关于’数据’（大部分）。

首先，它意味着收集比该组织的当前或传统的分析需要更广泛的数据。因此，除了运行在金融系统上的相同月度报告，还有例如，在网站上统计的数据，以及今天分析和它没有已知关系的业务整合的数据， 其次，它意味着通过结合本不同的数据（也可能是外部的，如地理或社会）的新方法推导了新的见解。相对于“每周报告包”，这里更多的是“数据发现”

第三部分是关于更有效的可视化，可视化可以帮助利益相关者吸收、共享和利用见解新的数据分析。

如果第一支柱（’问题’）是为帮助管理人员更好地了解他们的业务驱动，第二个支柱（’加速’）是关于业务响应的速度和精度。

一旦发现部取得了一些见解，大数据的第二个方面是关于把它投入到应用工作之中。这可能是通过帮助管理者更迅速做出决策的按需报告，也可能是一组到输入数据自动作出决定（又名“算法”）的应用规则。

第三个支柱是“转化”，这是它部署的过程。许多厂商将大数据视为技术浪潮，但大数据更应被视为业务转型浪潮。这正是为什么直到业务变化问题得到妥善解决其也没有成为主流产品得原因。

第三个支柱的寓意是指需通过实验学会开发数据的机会。“数据作为企业资产”的原则与此相关，并推而广之到各个业务已达到的成熟水平。好消息是，随着越来越多的企业部署，有越来越多的协议定义各个阶段的成熟标准。

有意思的是，科技企业是最后开始转向业务人员承认的大数据定义的。匹维托公司首席执行官保罗•马瑞兹日前表示：“今天，多亏了被称为‘大数据’的技术，计算机才可以捕获事物的发生，而且可以影响事件的展开”。

因此，和所有媒体的炒作相反，大数据确实是既不大也不是关于数据，它是关于质疑、加速和转化。

作者：Mike Fish是BigData4Analytics（一家大数据管理咨询公司）的董事

英语原文：

Many people ask why there doesn’t seem to be an accepted definition for big data in view of the massive press it receives. And given the amount of marketing effort expended by the global IT vendors who are targeting this bandwagon (and the venture-funded startups), they can be forgiven for being confused.

Just as most consumer marketing – curiously – is focused on adolescent buyers with a fraction of the purchasing power of their parents and grandparents, so it is with big data and its promoters.

They shower their traditional customers – the IT departments – with buzzwords, while failing to engage with (or indeed, completely alienating) the business managers who have both the business problems that analytics can solve and direct control over the budgets.

Why is this, and what does it mean for big data adoption?

The targeting of IT departments stems from both laziness and geekiness. Laziness in the sense that, with few exceptions, these large vendors got to be large without needing to address directly the concerns of business managers – the IT departments they deal with have until now performed this translation task for them.

And geekiness in the sense that the smaller specialist vendors are often managed by technologists, and so their market messages – for reasons of background – also tend to focus on ‘speeds and feeds’, rather than on solving practical business problems.

This has negative consequences for big data take-up and risks becoming the sand that makes ‘the new oil’ less attractive.

The massive hype – the promise of big data – that has appeared in the press has served to both raise expectations among business managers, and also to confuse them.

Vendors hoping that the IT department will take the lead in initiating big data projects are deluding themselves. IT departments today are a) far too busy with day-to-day work (BYOD, mobile access, security, etc) to do this, b) too distant from the concerns of functional managers to be able to guide them in selecting promising areas for piloting analytics, and c) often unskilled as data scientists.

So, returning to the question of what the term ‘big data’ actually means, there is no accepted definition – however, it’s not about ‘big’ and it’s not about ‘data’ (mostly).

First, it means collecting a wider range of data than the organisation’s current or traditional analysis requires. So instead of running the same monthly reports on the financial system, for example, or on website stats, the business incorporate data that may have no currently known relationship with how it analyses its data today.

Second, it means deriving new insights by combining this disparate data (that may also be external, such as geospatial or social) in new ways. This is more about ‘data discovery’ than about ‘weekly reporting packages’.

The third part is about more effective visualisation – helping stakeholders absorb, share and exploit insights from new data analyses.

If the first pillar (‘question’) is about helping managers better understand the drivers of their business, the second (‘accelerate’) is about speed and accuracy of business response.

Once the discovery part has yielded some insight, the second aspect of big data is about putting it to work. This might be via on-demand reports that enable managers to make decisions more rapidly, or it might be the codification of a set of rules that are applied to the incoming data to make decisions automatically (aka ‘algorithm’).

The third pillar is ‘transform’, the process by which it is deployed. Many vendors see big data as a technology wave, but it should be indentified more as a business transformation wave. This is precisely the reason why there is not mainstream adoption until the business change issues are properly addressed.

Implicit in this third pillar is the need to learn to exploit the data opportunity by experimentation. Related to this is the principle of ‘data as a corporate asset’, and by extension to the level of maturity that each business has achieved. The good news is that as more businesses deploy, there is increasing agreement in what defines each stage of maturity.

It is interesting that the technology business is at last starting to move toward definitions that business people would recognise. Paul Maritz, CEO of Pivotal, recently said: ‘Today, thanks to the technologies known as “big data”, computers can capture things as they are happening and can affect events as the events are unfolding.”

So, contrary to all the media hype, big data is really neither big nor about data – it is about questioning, accelerating and transforming.

由36大数据合作伙伴 北理大数据教育 翻译自bigdata-madesimple，并由36大数据编辑。

转载请标明