The pace of innovation is increasingly dictated by who can make the best use of available data, prompting fears that tech giants like Google are looking to eat the lunch of data-driven industries such as insurance.
In some respects these fears are well founded, but one asset insurance companies have that is actually difficult for companies like Google to replicate is loss data.
Adam Kornick, global analytics director at
Aviva, said: “While Google may have hundreds of millions of handsets contributing data -- they know who people are and where they’re going -- they don’t know the insurance algorithm. They don’t know if someone has made a claim or if someone got sick,” he said. “They can try and estimate but they won’t know nearly as well as an insurer.”
Claims analytics is much more than just fraud, Kornick said. “When it comes to customers, we want to know them like family, we want to be able to predict what they want before they ask for it.”
Kornick referenced the predictive capabilities of web giants like Amazon, Netflix and Google, “which are taking all this information and building a recommendation engine, offering you things based on what they think you are likely to prefer, based on your history and other people who have similar characteristics.”
But Kornick notes these systems are probabilistic and not perfect. “They can’t tell from different users on multi user accounts, or if you’re shopping for someone else. Say if you’re shopping for kids, then you would start seeing more ads for toys. What we’re trying to deliver doesn’t have to be perfect, just less bad,” he said.
“We take all the estimates and combine them to guess what the user is most likely to prefer, and combine that with the thing that is our most appropriate offer. Any place where there is some uncertainty and we can make a prediction to make our proposition better, that’s what we’re all about,” he said.
There is also the opportunity to use data analytics to give value back to customers. Aviva’s Home Advisor product lets users compare postcodes with their existing home to see differences in home insurance premiums. With the amount of data insurers hold, the insight is already created, it’s just a matter of making it accessible.
“Predictive analytics -- places where we can offer cross products, or multiple offerings from the same portfolio -- it sounds simple but it requires a vast amount of data. Something like 600 million observations in fact. Just a few years ago we couldn’t have crunched that dataset,” Kornick said. But now it’s possible, and general insurance, for now, is where the biggest gains can be made. “IF you look at life insurance in general, and investments and savings, these sectors haven’t invested as heavily as we have in general insurance in data analytics,” Kornick said.
