大数据企业想要成为行业巨头的5个要素

编者按：Navin Chaddha是早期阶段风险投资公司Mayfield的总经理。这家公司目前正在投资的一些公司包括Gigya、Elastica、Lyft、MapR和Poshmark。

随着2014年下半年的到来，大数据俨然已经成为了一种社会主流，它影响了我们的休闲读物、多个产业的格局和面向消费者的应用等各方各面，同时也左右了大批资本的流向。风险投资行业在过去45年的时间内已经见证过许多技术周期——从PC时代的诞生，到主从式架构计算和基于网络计算的发展，还有云端和SaaS模式的崛起，我们对一家公司从创业阶段发展成行业巨头的模式已经形成了一种固有的认知。

根据我们的观察所得，成为一家基业长青的大数据公司需要具备以下的条件：

1. 从平台向生态系统的转换

要了解一个技术平台是否掌握主导地位，最清晰的方式就是看看这个平台的生态系统建立速度有多快。例如在SaaS时代，Salesforce能够快速成为业界领军的原因正是它拥有一个庞大的生态系统。大数据时代也是一样。

在大数据领域有一家叫做MapR的公司发展十分迅速，它就是一个从平台转换成生态系统的例子。作为一家Hadoop平台的服务商，它是唯一能够将开源（社区创新、便携性和灵活性）的优势体现在独特的平台架构升级的公司，为客户提供企业级的可靠性、安全性和性能。

MapR的生态系统不仅融合了新兴的Hadoop开源社区，而且在MapR App Gallery中迅速扩展合作伙伴的解决方案组合。企业客户可以在这个生态系统当中利用现成的大数据工具和应用轻易地部署和扩展大数据方案。

另外一个例子是MongoDB，这是一个业界领先的开源NoSQL数据库，被多家公司用于各种类型的应用当中。MongoDB正在为各行各业的合作伙伴建立一个大规模的生态系统。

2. 解决没有人愿意处理的棘手问题

这并非大数据世界当中最光彩的部分，然而我们相信这种类型的工作会造就许多大公司。在主从式计算的时代，数据整合先驱Informatica在解决复杂的数据整合难题的过程中逐步成为业界巨头，而且在Gartner Data Integration Magic Quadrant当中占据了连续八年领导地位

在这个领域值得留意的另外一家公司是Trifacta，它的平台可以帮助技术类和非技术类的分析师将原始数据转换成可执行的数据。

3. 在大数据时代彻底改造商业智能，在获取数据的同时提供分析结果

像Business Objects能够帮助行业管理人员获取数据分析的结果，于是它成为了主从式计算时代的行业巨头。我们认为一部分的大数据公司也正在成为像Platfora这样的公司，后者能够在本地部署Hadoop，实现快速获取实时可视化的分析结果。

4. 深入运用专业领域的知识

确保专业领域的宝贵知识能够运用到你的分析应用当中，这样你才能立于不败之地。SAP就是利用这个策略成为了软件行业的巨头。

我们从Palantir这样的大数据分析公司当中看到了这种宝贵的专业知识，这家公司专门为反诈骗和网络安全这些特殊领域提供由人力驱动和机器协助的解决方案，它服务的垂直行业包括国防、保险、医疗和执法等。将机器数据转化成分析结果的Splunk也能体现出这种特质。

5. 利用直观的界面取悦客户

为你的IT和行业客户提供赏心悦目的数据交互界面；理解用户与应用进行交互的方式，不断改进用户体验的细节，做出直观和美观的界面。例如Dropbox在实现了一种简单直观的文件共享方式之后就迅速成长为一家行业巨头，现在它在世界范围内已经拥有超过2亿用户。

能够提供直观界面的大数据公司还包括Tableau，这家公司通过生成可视化内容 查看和理解数据，并从中得出分析结果；还有Elasticsearch，这是一个能够提供快速丰富搜索体验的开源解决方案。

大数据时代的未来

我们还需要关注的另外一个领域是物联网，因为它将会以各种全新的方式提供数据，从而改变技术产业的格局。现在这些数据的来源可以是恒温器、手机和手表，甚至是水杯这样的物品……以后的数据将会来自我们从来没有想过的地方。关于数据的所有权、生命周期和提取的全部观念都要经过重新定义，届时将会催生出一大批新的公司。这将会掀起新一轮的创新大潮，公司会推出一些以前从来没有想象过的全新产品和服务，而现有的产品和服务将会改写。（译：consideRay）

英语原文：

As we enter the second half of 2014, it would be fair to say that big data has gone mainstream, attracting coffee table books, multiple industry landscapes, consumer applications, and large amounts of funding. Having seen many technology cycles during our 45 years in venture capital — including the birth of the PC era, the transition to client-server computing and then web-based computing, and the emergence of the cloud and SaaS models — we have pattern recognition on what it takes for a company to go from startup to leader.

Here are some observations we’ve made about what it would take to build a lasting big data company:

1. Transition from a platform to an ecosystem

One of the clearest ways to see whether a technology platform is taking hold is to look at how fast the ecosystem is growing around it. For example, in the SaaS era, Salesforce rapidly became a giant because of its expansive ecosystem. Big data will be no different.

One thriving big data company that is transitioning from platform to ecosystem is MapR. It is the only distribution for Hadoop that combines the benefits of open source (community innovation, portability and flexibility) with unique architectural enhancements that provide enterprise-grade dependability, security, and performance.

The MapR ecosystem embraces both the flourishing Hadoop open source community as well a rapidly expanding portfolio of partner solutions in the MapR App Gallery. This enables enterprise customers to easily expand and implement their big data initiatives with ready-made, big data utilities and applications.

Another example is MongoDB, an open-source and leading NoSQL database used by companies for a wide variety of applications. MongoDB is building a significant ecosystem of partners across industries.

2. Solve the messy, hard problems no one wants to touch

This is not a particularly glamorous part of the big data world; however, we believe that many big companies will be built doing this work. In the client-server era, data integration pioneer Informatica became a giant by tackling tough data integration challenges and has maintained its edge by being positioned as a leader in the Gartner Data Integration Magic Quadrant eight years running.

An example of a company to watch in this space is Trifacta, which enables both technical and non-technical analysts to access and transform raw data into actionable data.

3. Reinvent business intelligence for the big data age by providing insights, not just data

Companies such as Business Objects that empowered line of business executives to gain insights grew into giants in the client-server era. We believe that a similar class of big data companies are in the making with companies such as Platfora, which are built natively on Hadoop and rapidly deliver insights visually and iteratively.

4. Embed deep domain expertise

Ensure that valuable expertise from your specific domain is embedded into your analytics application so that it cannot be dislodged. SAP became a giant in the software industry using this strategy.

We see this valuable domain expertise in big data analytics companies such as Palantir which provides human-driven, machine-assisted solutions for specific use cases like anti-fraud and cybersecurity as well as to vertical industries like defense, insurance, healthcare, & law enforcement; and Splunk which transforms machine data into insights.

5. Delight your customers with an intuitive interface

Give your IT and line of business customers compelling interfaces to interact with their data. Understand how users interact with your application and invest in the details of the user experience to make it intuitive and delightful. For example, Dropbox became a giant after creating a simple, intuitive approach to file sharing that is now shared by more than 200 million users around the world.

Big data companies with intuitive interfaces include Tableau, which can create visualizations to help enterprises easily see, understand and derive insights from their data, and Elasticsearch, the open-source solution that offers a fast and rich search experience.

And what’s coming next?

And one more thing, keep your eye on how the Internet of Things will transform the landscape by serving up data in all kinds of new forms. Today it’s thermostats, phones, watches, even drink glasses… tomorrow data will come from places we have not yet dreamed of. The whole idea of data ownership, lifecycle and ingestion will have to be rethought, spawning new companies. This will give rise to a wave of innovation and companies creating new products and services never thought of or possible before, and existing ones re-imagined.