Eavesdrop at an industry dinner and you’ll hear talk of little else but Big Data and how it’s going to transform the business. But just what are they going to do with all that information?
© Getty Images
New York City has a director of analytics whose job is to search for useful insights in public sector data. The first holder, Mike Flowers, made his name in Iraq using Big Data to identify terrorist hot spots. In the Big Apple he found a correlation between obscure things such as brickwork maintenance and rat sightings with illegal tenancies. The hit rate for inspectors armed with this knowledge rose from 13% to 70%.
It’s this kind of result that’s got the insurance sector interested in how it might be able to use Big Data to improve the way in which it acquires customers, prices policies and handles claims.
At a basic level, Big Data is the use of mathematical modelling and algorithms to find hidden insights in data. The more data the research scientists can access, the higher the likelihood they’ll find something interesting. And in the insurance industry, Big Data has found a business with a huge amount of data.
It starts even before you’ve signed up with an insurer. The question each insurer must answer is how much to pay to acquire a new customer. Previously, the assumption was that you took the margin and advertised up to that point but no further. Why pay £50 for a customer worth £30? Answer: the customer may stay longer. They may tell their friends about your service. They may purchase affiliated insurance products.
Martina King, chief executive of adaptive behavioral analytics firm Featurespace and former MD of Yahoo Europe, explains: “Acquisition comes at a cost, so smart choices must be made: which customers are worth acquiring, and at what price? The typical break-even point for a customer is three years, so insurers use data to predict which people are likely to purchase additional products, who will renew, who may cancel, and so on.
“Insurers with a wider product offering also have the ability to make better decisions by aggregating all data: fraud, claims, underwriting, and renewals, etc. Doing so across a customer portfolio (for example, using pet insurance information to make better life insurance decisions) and using real-time individualised analysis – like Adaptive Behavioral Analytics – means insurers can get ahead of their competitors once more.”
Claims handling is a nightmare. Insurers need to catch fraudulent claims early, but also accelerate genuine claims to avoid annoying loyal customers. Big Data helps them to do that. Davy Nys, vice president of Pentaho, a Big Data consultancy advising the insurance industry, says: “Big Data will let you spot trends. Suppose a kid breaks a vase. The parent will check online what their policy says. They will call a contact centre to make a claim. Big Data can watch what is going on and let you know early on whether there is anything for you to worry about.”
Of particular interest for insurers is what’s referred to in Big Data circles as ‘unstructured data’, which is finding ways of using data that isn’t in a computer-friendly format. Pictures, video, recorded speech and poorly organised digital data are all unstructured.
The insurance industry is riddled with unstructured data, such as written claims forms. Nys says this is a fast-growing area: “Sometimes you can use the metadata, for example the geolocation data, or keywords. Speech to text is another possibility.” The industry is reluctant to let any source of data go unexamined.
And what about the future? Now the insurance industry has a taste for Big Data we are likely to see an insatiable appetite for new streams of data. You have a mobile phone switched on when you drive your car? Perhaps your insurance firm will use it to track your movements. Nys hints that accelerometers in smartphones could be used to establish whether you have been driving safely.
With £1.8 trillion in investments, employing 320,000 people in the UK (according to the ABI), generating £184 billion in premium income annually, the insurance industry has both the incentives and resources to be at the cutting edge of Big Data research.
Anyone hoping to make it in the business needs to become very familiar with this booming science.