Long before anyone knew what ‘big data’ was, the insurance sector has been a data-driven industry. While selling no physical product, virtually every decision an insurer makes revolves around processing reams of information to calculate risk, determine pricing models and better understand customers and competitors.
But as Simon Gratton, chief data officer at Zurich Insurance explains, the sheer volume of information insurers have to deal with has exploded. And because commercial insurance carriers don’t have a direct relationship with the customer themselves, giving them a reputation as ‘aloof’, it makes bridging the knowledge gap even more of a challenge.
‘If you think about insurance traditionally, it’s a process that uses lots of internal information,’ says Gratton. ‘But now, when we talk about “big data-enabled insurance”, it’s about gathering more information than ever before, not just from inside the organisation but from outside.'
‘Each key event in the customer cycle, we have to capture that information and make it available immediately. And we can now analyse everything from flood patterns to driving patterns, with sensors on aircraft and cars – big data is about using as much as that information as quickly as possible to make the right decisions about risk.’
The operating model at Zurich was, and is, one of a traditional insurance company in many ways. But in the six months since his role began, Gratton has set out to explore the art of the possible with data science and emerging technologies such as the Internet of Things.
Rather than the quality of information, which in the insurance industry has always been well structured and organised, the challenge to Zurich now is the speed at which it can ingest and deliver it.
It’s very early days, but Gratton is setting strategies in place, and a number of focused pilot schemes, to look at specific problem areas and demonstrate how the process of getting the most out of data might evolve into whole new operating models.
‘If you look at what’s happening over the pond,’ he says, ‘there are small data-driven startups in the insurance space, particularly life insurance, where they are tying things like wearable devices and the collection of fitness-related data into insurance products, but also using it to create health and wellbeing services.
‘They’re using health data not only to be responsive to the market but to actually offer value back to their customers, much like Amazon enabling people to lease computing space on the back of the infrastructure they’ve already built for their other services. That’s one of our aims.’
For Gratton, the more sensors his firm can interact with, the more it can understand customers’ risk profiles, how best to price products in the market and personalised towards customers’ needs.
The insurance products of the near future will be data-centric, offering a different way to directly interact with and build more personal relationships with customers.
According to analyst firm Gartner, by 2017 the average mobile user will provide personalised data streams to 100 apps and services a day. ‘When you think about the amount of information out there,’ says Gratton, ‘there’s an enormous untapped area.’
A virtuous cycle
Major US insurers are now using a dongle device that can be plugged into car dashboards to monitor driving habits such as speed, distance and braking – and the trend is moving to the UK.
Such telematics tools are growing in popularity among drivers who want their insurance quotes to be calculated by their driving ability rather than their age or profile.
For teenagers particularly, it can be very expensive to get insured, so pricing insurance products more accurately according to the behaviour of individual drivers can make them more accessible to a wider audience.
Last month, however, researchers from the University of California, San Diego, demonstrated at the DEFCON hacker conference how the dongle device could be hacked via text message in a matter of minutes – turning on the car’s windshield wipers remotely and putting on the brakes.
‘These telematics tools do bring down the cost of insurance quite substantially for customers,’ says Gratton. ‘But not only does security and hacking have to be taken into consideration, but the appropriate use of data. Knowing where that information is going and that you have control of that supply chain is really important.’
Fortunately, for insurance companies, security and compliance – much like data analysis – tends to come second nature in such a highly regulated industry.
‘Because of the speed at which technology is moving, protection of that data, and governance of it, is definitely a big challenge,’ says Gratton, ‘but I wouldn’t say it’s our biggest one.’
When the role of the chief data officer first emerged in the insurance sector, it revolved primarily around ensuring the governance and compliance of data.
While that is still very much part of it, Zurich is now more focused on delivering innovation at speed, as well as empowering business users to work with data in new ways and take ownership over such projects.
When leading data transformation at Canadian internet service provider TELUS, and later at management consultancy group Capgemini, Gratton built apprentice-based ‘data science factories’ that took on graduates and built relationships with universities and apprenticeship schemes to encourage big data skills, effectively nurturing cultural attitudes to data right from the ground up.
‘At Capgemini, one of the biggest gaps I saw was predictive modelling or statistical modelling through technology like SAS,’ says Gratton. ‘Capgemini already had a scheme set up using traditional technologies, so it was all about cross-training technology we were not used to.’
It took about 18 months to set up, but, through the data science factory at Capgemini, Gratton trained a number of apprentices and put them through projects in the credit card industry.
‘We found from this that if you empower people, they’re very passionate about the capabilities they have, and they can do more under their own steam,’ says Gratton. ‘They look after the data and they govern it better – it’s a self-fulfillment cycle really.’
The new digerati
The term ‘digital’ is increasingly being used as a catch-all for anything to do with technology. But as Gratton argues, it comes down to far more than just taking physical processes and putting them on mobile phones or tablets.
At Zurich, one if his core focuses is changing the speed at which people interact with each other, and how much employees know about customers at each point they interact with them.
‘Whether you’re a call centre operative or broker dealing with a claim, the more information you have about that company or individual the more likely you are to produce a better service,’ he says. ‘For such a traditional industry, it’s quite a change to go through that.’
Meanwhile, Gratton hopes that more traditional and repetitive tasks, such as duplicating information across numerous applications, will soon be replaced by software robots.
While artificial intelligence and cognitive computing technologies such as IBM’s Watson are supporting diagnoses in the medical industry, they could also intelligently collate and process information about insurance – allowing the right human decision to be made at the right point.
So with employees and robots taking data science under their wing, where does it ultimately leave those like Gratton and the role of the CDO in the insurance sector?
‘In ten years’ time the role may still be there,’ he says, ‘but the world is changing. There’s an argument that says the CIO could do what the CDO is doing, and ten years from now they probably could.
‘But the current breed of CIO tends to be focused on the IT function, big bang projects and simplifying technology, whereas the CDO right now is about agile, quick value creation, driving innovation around data through lots of pilots.’
As everyone in a business evolves to become more data oriented and fluent in using data in their everyday processes, Gratton believes the CDO is likely to become part of the fabric of the organisation, and may not be a role in its own right.
‘There’s no reason why it couldn’t go to the COO or somewhere else, but what’s for certain is that whoever is in charge of these functions, they will sit at board level.’