TORONTO – Businesses have a lot of data and are putting big data analytics systems in place to help analyze it, but many are stumped by one key issue: what questions should they be asking of that data?
Attendees at a recent CanadianCIO Executive Roundtable event “Unlocking the Hidden Value of Big Data” sponsored by HP Canada agreed that they have plenty of data, but they’re challenged to ask the right questions of their analytics platforms to derive the business-driving value they know that data contains.
There are a lot of untold stories in our data said Dave Schubmehl, research director for content analytics, recovery and cognitive systems research with International Data Corp. (IDC), who presented the results of research of 2,000 global knowledge workers about what they do to unlock value in their data. He said the state of information access hasn’t dramatically changed over the last decade, despite the big data hype.
“There’s a tremendous opportunity to find the information hidden in your company, whether it’s in a content management system or your email servers. If you acquired a company, it may be in another system you don’t know is there,” said Schubmehl. “All this is potential information but no one knows how to make use of it, or make weak-signal correlations between different aspects of that data.”
IDC used the survey data to create a Knowledge Quotient of the attributes of high-performing organizations around data analytics, consisting of four areas. They need Process to access, analyze and share all relevant data from across the organization and from partners. They need Technology to access, analyze and share data. They need Socialization to share and reuse information. And they need a Culture of data sharing and support from management. Just 10 per cent of those surveyed did well in all four, said Schubmehl.
You can have all the technology in the world, but you still need to know what questions to ask said Jeff Healey, director of product marketing with HP’s big data software unit.
“We work closely with clients to learn their stories, and we remove the technology barrier so you can ask any question you want of the data,” said Healey. “We give you an open platform with scale and performance, so any question you want to ask, you can ask.”
The attendees at the roundtable, which included CIOs and technology leaders from several major Canadian organizations, agreed that what questions to ask of their data is a key stumbling block for many of them.
One company in the real estate space began building its data warehouse four years ago and has lots of information on the buildings and assets it manages for clients, but it knows it could be making better use of all that information.
“I don’t feel we’re really getting the value out of the information assets that we should,” said the CIO. “We’re making a renewed effort this year to re-engage with clients, talk to them about their business problems and see if we have the data behind it. From my perspective, most of the value is still hidden.”
The IT leader at a Canadian educational institution added getting to the story in the data is their challenge. Take graduation rates and retention, a key concern for any organization of higher learning. Those metrics come after the fact, once the student is gone. It doesn’t speak to why they left, but the why may be hidden in data they already have – if they can ask the right questions. And if they know the why, they may be able to prevent future premature departures.
“What are some of the right questions? We can find the answers if we know what questions to ask from our business colleagues,” he said.
Data management was also a challenge for several organizations. What data should be kept, what data should be discarded, how long should it be kept, what are the legal and privacy implications, and how should data be tokenized so trends can be extracted while personal information is safeguarded? In short, how do we manage the data lifecycle?
A tech leader at one company said they suffer from a wealth of data measuring in the petabytes; they’re focused on getting data scientists involved early to help determine what has value and what sources they should be pulling from. He said it’s essential line of business leaders are involved and their burning questions are identified.
IDC’s Schubmehl said you need to turn your spreadsheet jockeys into data scientists and set up your internal information to allow them to do Google-like searches and analysis. He points to HP Autonomy, a solution for the large scale analysis of unstructured big data, as a tool that allows such searches and helps the relevant information bubble to the top. HP Vertica is another solution for analytic database management that allows faster queries against structured data and real-time analytics, and can be combined with Autonomy under HP’s Project Haven.
“You can’t boil the ocean, so pick a manageable app to start with where you’ll get value,” said Schubmehl. “Organizations that measure the return on investment of those projects are more successful getting further projects funded. Pick a project you know will have lasting value you can measure that can be a showcase for where you go next.”
ITWC CIO Jim Love, who chaired the roundtable, added we need to develop a culture of evidence-based decision-making, not decision-based evidence-making.
“Until we get to the former we’ll never be there,” said Love. “Start with small teams to get results you can show off – not all organizations can afford data scientists.”
IDC has a big data maturity model that can help companies looking to get started. It will take time to develop the right processes internally, said Schubmehl.
“The idea is to establish the right set of processes to help you define, develop and implement your big data processes,” he said. “Develop a roadmap of where you want to go, then see where big data projects fit in. Think about what you have in the way of information inside of the organization.”