Insurance firms unsure how far to push big data mining while remaining ethical

Insurance industry unsure how far to mine data - and the implications of ever-more accurate information

The insurance sector must be mindful of how it uses big data and analytics as more and more information about potential customers becomes available via telematics, wearable devices and fitness monitoring applications and technology.

That was one of the conclusions to come out of Computing and Talend's recent "Big data and analytics challenges in the insurance sector" roundtable.

There was much discussion about the ethical dilemmas that come with the analysis of big data, a practice that could potentially reveal every bit of information about a person to an insurer - even those things they'd prefer organisations not to know.

"Perhaps what's important here is how the data is going to be consumed, and in that part there are questions surrounding analysis of those pieces of data: how the data is integrated, how the data is shared and what is it that you want to build? Do you want to make all the data available?" asked a representative of BNP Paribas PSC, who didn't wish to be named.

He suggested that there are particular questions that need to be asked surrounding the use of information in the health service, for example, although he conceded that insurers would be able to take on much less risk if they were able to use the data.

"In the NHS, do you want to make all the data available? Perhaps if I had had cancer that would change how insurance brokers see what my risk factors are. So given that the business on the insurance side is based on those types of data, I think that it has very big complexities," he said.

"So, with big data, if we're talking about harvesting or accessing data for analysis, to determine new products, new trends or new opportunities, then that represents a different way of addressing that," he added.

Simon Davey, group head of IT for Kelliher Insurance Group, agreed that there are questions to be raised on the issue of using big data in the insurance sector, but conceded that at the moment, there didn't appear to be any organisation or regulator around to advise.

"I think there's an ethical side to big data. To my mind, insurance is about shared risk, but how far do we take this? How far do we adjust risk according to criteria like medical data? And I don't know who answers that question," he said.

The representative from BNP Paribas PSC suggested there are also questions to be asked surrounding 'garbage data', especially if insurance firms are Hoovering up every bit of information they can.

"What about all the data we just class as wasted data? There's a lot of information, everywhere, that we've not catered for at all. There's no one filtering all the garbage that goes in there, or information about things which are outdated," he said.

The answer, he continued, is that the business needs to have a clear intention of how it wants to use information before it starts collecting information and analysing it with big data tools.

"With data you really need a clear business perspective, a clear business strategy about why you want and what you want to filter - otherwise you'll have a lot of work to do, but for no insight," he concluded.