For many companies, the Big Data frontier, while promising, is fraught with pitfalls.
Like any other new and powerful business tool, Big Data analytics can be dangerous if it isn’t managed properly. On the other hand, not using the growing information at your business’ disposal could prove even more dangerous to your market success in a fast-moving, data-driven world.
I’d like to reassure you that there is plenty of room for middle ground. Your organization can safely gain great insights from Big Data by leveraging the right combination of people and skills—some of which are likely already part of your operation—and by tapping Data-Science-as-a-Service resources as you need them.
In a recent Wall Street Journal article, Professor John Jordan warned of the Risks of Big Data for Companies (http://online.wsj.com/news/articles/SB10001424052702304526204579102941708296708). Faced with access to more information than organizations ever dreamed of, he said, “There’s a greater potential for privacy invasion, greater financial exposure in fast-moving markets, greater potential for mistaking noise for true insight, and a greater risk of spending lots of money and time chasing poorly defined problems or opportunities.”
Certainly, these are potential perils from Big Data analytics gone wrong. Just as certainly, there are steps you can take to effectively harness your company’s Big Data at your own pace without getting lost “in the noise that can obscure the basic forces represented by customers, value and execution.”
The biggest step your organization can take toward mastering Big Data is building the right team of people and talent to meet your analytics needs. There are three sets of people or skillsets, needed to manage Big Data.
- The first, not surprisingly, is the data scientist. This person has a statistical analysis, computer science and programming background. He or she makes sure the data is clean, the model is effective, and the results are accurate. You may not have a data scientist in your organization, but you don’t necessarily have to hire a dedicated staff member for this role. You can opt for using Data-Science-as-a-Service resources as needed to provide the statistical and computer expertise for your project. At EMC IT we have created a Data Science Competency Center with a team of data scientists that our business units turn to for such services.
- Next, the business domain expert knows the domain inside out. While this person may have some statistical knowledge, his or her crucial ability is to provide answers on what the data means to that particular segment of the business (customer service, marketing or sales). This person can say if the required data is all there and what factors or predictors are potentially most relevant to their business focus. This is the person who prevents data analytical misinterpretations like the instance cited in the WSJ article in which a pharmaceutical company analysis misread a spike in aspirin sales as an escalating flu outbreak when it’s really due to holiday hangovers. The domain expert is likely someone who is already on your staff.
- Third is a strategic consultant who frames the problem which you are using Big Data to address. While the domain expert might be really good at knowing what the data is saying and the data scientist might be really good at mining the data, you still need someone with the expertise to set priorities and define the pain points to be addressed. The strategic consultant knows how to ask the right questions, identify the root problems, set priorities and build a pipeline of what set of issues to go after in a very coordinated manner. The strategic consultant keeps your organization from “spending lots of money chasing poorly defined problems or opportunities.” This person could be a third-party consultant or a senior person in your business.
With the combination of these skills, your organization can deal with the challenges that Big Data presents without diversion or paralysis. The data scientist provides the technical know-how to extract the right numbers. The domain expert and the consultant let business leaders distinguish the noise from the true insights and determine if a question is worth asking.
While it is true, as Professor Jordan points out, that Big Data tools are still evolving, you can rely on data scientist expertise to deal with that changing technology. Our IT team works with our Greenplum Big Data platform and a variety of tools, including Hadoop, MADlib, R and SAS. Through our Data-Science-as-a-Service offering, our business units can rely on our data scientists to leverage the right tools to provide the statistical and data mining techniques and machine learning capabilities to meet their research needs.
We have found that our business leaders are doing well at absorbing the benefits of “thinking big” in the face of Big Data because they are able to leverage our Competency Center in conjunction with domain experts and strategic consultants to keep things on track. For example, our Sales team was able to use the latest data analytics technology to hone in on maintenance contract renewal opportunities and dramatically increase our renewal rate.
Clearly, the success of Big Data analytics goes back to identifying the right problem to address, extracting the right data, and interpreting the results. (Read: The Big Data Wilderness: Finding Your Way Starts With Asking The Right Questions) The business needs to gain belief in what they are learning as insights from these data tools. That is the journey we are on at this point and we are not rushing it.
How does your business gain belief? Start small and don’t expect to change longstanding business processes overnight. Verify what your Big Data analytics results are showing and take gradual steps to make changes.
As Professor Jordan notes, numbers don’t speak for themselves; but by marshaling the right skills and resources, your company can use the powerful tool of Big Data to gain new and valuable business insights.
To learn more, read Liberate Your Big Data And Close The Loop To Leverage Its True Value
