Banks need to start asking ‘right questions’ on unstructured data
7 December 2011 | 0
Banks that are struggling under the weight of large unstructured data stores need to start asking better questions in order to get value from the information.
That was the verdict of senior executives at some of the world’s largest financial IT suppliers, speaking at the Bloomberg Enterprise Technology Summit in London.
Bindia Hallauer, Microsoft’s chief technology officer for the finance sector, said companies were missing the point if they only thought about how to improve the technology for managing the data.
"What if you take the traditional approach, slice and dice the data and normalise it? How does that do anything to deal with more human questions, such as reactions in the market, or natural disasters?" she said.
"We must rethink the questions we ask, and take into account these real factors. Then you stand more chance of monetising the information and understanding the real issues your customers are having."
Simon Aspen-Taylor, financial services consultant at IBM, said retail banks were "sometimes much better" at managing and monetising unstructured data than investment banks-including how they understand client sentiment.
"Most investment banks are excellent at handling structured data," he said. "But they have so much more to do to work out how deal with unstructured information, such as the feeling in different parts of the market."
SAP’s chief technology officer for its Sybase business intelligence division, Irfan Khan, told delegates that armed with a more modern approach, they could then adapt the technology.
"There’s a real problem with the breadth and depth of unstructured data. It’s not new but it’s getting worse," he said.
"Companies are starting to make a lot of investment in bespoke search engines when they know the information they really want," he added. "A lot of the problem is that the software isn’t keeping up with the real world issues or even the power of the hardware."
All three vendors have made significant moves in the big data analysis space-with SAP pushing its in-memory database as the best route to fast intelligence, and Microsoft and IBM now working closely with the Apache Hadoop software framework for managing distributed applications.
As their clients struggle to improve their data management, there was a new specific role emerging in the enterprise to deal with the problem, the vendors noted.
"The CDO, or chief data officer, is starting to emerge," said Khan. "This person is not focused on the technology alone, but the management of information."
IDG News Service