Big data has certainly been making waves in the insurance industry over the past few years. But as it turns out, big data is actually a concept that has existed since the first actuarial tables were made hundreds of years ago. The question is - how do the insurance players of today utilize these information mines?
Indio co-founder and CTO Adam Bratt (pictured) said carriers seem to look at big data as a new concept as they struggle with information overload. The trouble boils down to figuring out what these data possess - how can these data sets be used to the advantage of their organizations?
“On the agency side, big data is on everyone’s two- to three-year roadmap but I’ve seen very few agencies that are really using it to make proactive decisions in their business. Agencies without staff that has experience doing data analysis are sitting on a lot of lost insights,” he said.
With regards to the benefits of big data, there are two perspectives to consider. From a carrier perspective, big data enables insurers to provide the best coverage possible at the best price. Meanwhile, big data helps agencies figure out tasks needed to be automated to optimize their operations and increase their margins.
“It’s also super useful for aligning staff towards overall goals and giving them an indication of how they’re doing day to day quantitatively in reaching those goals,” he said.
Breach of consumer privacy
One of the most common issues surrounding the use of big data is consumer privacy. Bratt said while regulation is not needed on the existing data sets inside carriers, it becomes a different story when insurers start to procure data the way credit bureaus do.
“If you’re buying data sets about someone’s credit card debt and then using that to determine their insurance premiums I think you run into some privacy and regulatory risks pretty fast,” he said.
To appease consumers’ apprehension towards big data, insurance groups have the responsibility to equip their clients with a deeper understanding of the data sets being used today.
“The majority of big data analysis that is being done on the underwriting/rating side is using data sets that are fully anonymized and aggregated in a way that an individual’s data isn’t at risk of a privacy breach,” Bratt said.
He added: “There have been more and more conversations in Silicon Valley in the past few years about the ethics of data aggregation as it applies to someone like Facebook monetizing user data, but I don’t see a lot of those discussions being as applicable to the way that insurers are siloing data today.”
However, another problem lies with the algorithmic bias big data can have. For Bratt, it becomes a concern when the algorithm removes all emotional decision-making in the process.
“While that sounds good in theory, it removes a lot of the human component of insurance. Algorithms don’t see gray areas — for someone dealing with a substantial loss, trying to automate all human interaction out of the claims process seems like a really bad idea,” he said, “Although humans are getting more and more comfortable with technology, no-one wants to talk to an emotionless computer when their house burns down.”
Bratt said insurers who are seeking margin optimization and automation above all else will not survive in the long-run — only those who perfectly balance the human touch and the use of smart technology will.
In the future, he believes big data will just be called data. Major carriers will start using data to fuel machine learning and artificial intelligence on their rating engines.
“You’ll also see a lot more software out there that allows people to make proactive decisions on their data without having to be a full-on data scientist themselves. For any agency with over 100 employees you’ll be seeing at least 1-2 data analysts on the IT team,” he said.