News Stay informed about the latest enterprise technology news and product updates.

Getting started: Five key big data analytics decisions to consider

LAS VEGAS — Analytics and big data will soon transition from a supporting role to the center stage. Together, according to more than one speaker at the Gartner Business Intelligence and Analytics Summit this week, they’ll become the heartbeat of the enterprise, underpinning just about every process, customer interaction and business activity.

Thinking in such broad terms “opens up your mind as to where analytics can be used,” John Hagerty, program director for big data and analytics category marketing at IBM, said during his presentation. But making the actual leap is easier said than done. To help CIOs and IT organizations make the transition, he presented five key pragmatic decisions to make on big data and analytics.

1. Build a culture that infuses analytics into everything you do
Not everything associated with big data requires plunking down a fistful of dollars. Establishing a successful big data program also requires a change in mindset, which costs the business practically nothing. “This is about people,” Hagerty said. And it’s about curiosity by supporting people to figure out what works, what doesn’t and why.

It also means making big data and analytics a core competency — a part of what everyone does, embedded into the way the company operates. That shift in mindset will help the organization “move from the select few to the empowered many” and put “statistics and predictions and prescriptions into the hands of individuals and processes,” he said.

2. Find the right use cases
For the folks building an architecture, looking for use cases before investing in technology might not come naturally. In many cases, Hagerty said, they’re looking “for all of the pieces that may be used in order to solve any potential problem.” But going that route leaves the company open to poor or, even worse, useless investments. So start with the problem first and “look at it from the outside in,” Hagerty said.

It’s also not a bad idea to brainstorm ideas by looking at what other businesses are doing. As a little food for thought, Hagerty provided six types of use cases:

  • Attract, grow and retain customers.
  • Optimize operations and reduce fraud.
  • Manage risk, especially for financial institutions and insurance companies.
  • Transform financial and management processes in HR, balance sheets, profit and loss, and so on.
  • Build a nimble architecture to support the business.
  • Figure out ways to create new business models.

3. Invest in the technology to support your use cases.
You knew investment had to be on the list somewhere. But part of the goal here involves becoming fluent in all forms of data and analytics. Those two terms overlap to a degree but shouldn’t be seen as the same thing.

From a data perspective, ask yourself if you’re getting all of the data you need. Answering that question means considering the world beyond the data warehouse. Think, for example, about the streaming data and dirty data in which data scientists like to poke around, and the loads of third-party or external data sources. From an analytics perspective, consider the full spectrum — data discovery, text analytics, predictive and, yes, even prescriptive analytics.

4. Be proactive about privacy, security and governance
“If you don’t take a proactive approach, you’re going to get bit in the butt at some point,” Hagerty said. He should know: He’s had his financial information exposed due to company breaches. That’s why he recommends securing the data used for big data analytics as you would “your most trusted, internal financial information,” he said.

Building in security, privacy and governance strategies will help to establish a high level of trust in the data for internal customers, but it will also create company value in the eyes of your external customers, Hagerty said.

5. Understand the levers you can pull to differentiate your programs
Start with use cases but think big. Once a big data analytics program is operational, it will need to be cared for, attended to and, hopefully, stretched beyond its original intent. Based on research published by IBM’s Institute for Business Value, here are three tips to help you establish a successful program:

  • Be aware and understand what’s valuable, what your measurement practices are and what your platform is to support analytics going forward.
  • Drive better performance from these programs by making sure you’ve got the right data and establishing trust.
  • Amplify the program by incorporating such areas as sponsorship, funding and new or additional expertise.

BONUS: Understand your deployment options. Consider the following: “You now have choices: Do you want it to run on premises, in the cloud, as a service or take a combination of that approach?” Hagerty said.

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.