Organizations of all sizes are realizing they are already sitting on the answers to some of their biggest issues. The problem? Much of the data that could enable smarter business decisions is not structured in a way they can store – or make sense of – in traditional database systems. The emergence of “big data” offers great promise, so long as organizations are willing to rethink traditional business analytics.
“The traditional relational database is still very good at giving organizations the crucial data they need to make smarter decisions,” says Tom Deutsch, Program Director for Big Data at IBM Software Group. “But it requires companies to ‘know what they want to know’ and then structure their database accordingly. Often, they can’t do that. They just know they have a problem – as well as high volumes of structured and unstructured data that they cannot easily analyze.”
“Big data” promises something different: The ability to provide an answer to a question, even if you are not quite sure how to phrase it upfront.” It is about applying the latest and fastest data-mining and machine-learning tools to get a type of solution organizations wouldn’t have been able to achieve before, because they wouldn’t have known what questions to ask of the data.”
And the really crucial advance, he says, is that it can be predictive. “This isn’t just about looking in the rear view mirror of what happened last month or quarter. Big data is about understanding what is going on now, so that future events can be anticipated and planned for.”
Understanding and predicting customer behavior
For example, he says, one large financial institution is already using IBM’s expertise to help it gain intelligence from email and phone conversations with customers, to encourage brand loyalty and prevent defections.
“This client has more than 30 million customer interactions it needs to understand,” explains Deutsch. “So we’ve helped them to prioritize where they need to look most. For example, customers generally don’t just change their banking and investment accounts overnight: It’s a hassle, and it turns out people usually only do it after they have already informed the bank that the relationship has gone wrong. Our technology can help pick out those conversations and forewarn our client when people are most likely to consider churning.”
Interestingly, they also now have the ability to prioritize whom they want to keep. “So they can let unprofitable customers go, and expend maximum effort to improve the situation for a profitable customer they want to retain.”
CRM, risk and healthcare as growth areas for big data
There will be many key areas where the tools will allow organizations to make better, proactive, evidence-based decisions, but Deutsch believes financial risk analysis and healthcare rank alongside customer service and network management as some of the most fertile areas for early adoption.
“I think these are still early days for managing systemic risk in banking, because we are just now starting to utilize the available information – these companies are sitting on millions of loan records,” he says. “But our tools will soon be able to help identify sectors, company types and geographic areas where risk is heightened.”
In healthcare, he points to IBM’s work at the Universityof Toronto’s SickKids Hospital for babies in intensive care. “We looked for all the vital signs that indicated a baby was going to be sick,” he explains. “And we demonstrated how this could help medical staff anticipate the challenges the infants would face, to avoid their conditions – or at least give treatment at the very earliest stage. That has to be better than the current situation across the healthcare sector, where you wait for people to fall ill and then treat them.”
It is that ability to make discoveries and to enable organizations to navigate the future better – rather than just unravel the past – that ensures ‘big data’ will become an ever hotter topic in IT over the coming months and years.