There’s so much choice available that customers can pick and choose who they buy from and where, when, and how it happens. They want to discover, research, evaluate, and purchase on their preferred channel. Give them that option, and they’re more likely to choose you. That’s the whole point behind the multi-channel approach.
Big data analytics is not only providing Aon with the tools to improve existing insurance policies and risk management models, it’s also paving the way to product and customer innovation.
The risk management and insurance provider has invested about $350m globally in recent years to build out its data capabilities, culminating in two analytics and innovation centres of excellence in Dublin, Ireland and most recently, Singapore.
Steven Mildenhall is the global CEO of analytics at AON and in charge of the analytics and innovation centre in Singapore, which stretches across all business units and engages in projects for the retail arm, as well as reinsurance, health and HR consulting areas. He also heads up AON Benfield analytics, a global group providing data-driven services to brokers and clients around reinsurance trends, understanding and investigating ways to mitigate risk, business process improvement, and future trends and opportunities for the insurance sector.
During a recent visit to Australia, Mildenhall told CMO that everything Aon does today is data and analytics driven. Big data provides the glue between the supply side of the insurance industry, where there is a desire to expand the types of insurance products offered, and the demand side, which is seeking better ways to manage risks and generate topline growth, he said.
“On the one hand, big data is generating new risks, and on the other hand, the fact you have the analytics allows you to think about quantifying risks and designing products to help mitigate them,” Mildenhall said.
“In the existing lines of business, such as motor insurance, big data allows us to understand things better, and therefore manage risk better. But there are also new risks coming out because of connected devices – social media, cyber liability, brand reputation - and these are new opportunities for the insurance industry.”
According to Aon’s most recent emerging risks research, cyber is significantly expanding area of focus. The sharing economy, and new risks being raised through services such as ride sharing and Airbnb, is another emerging area for insurance. Brand reputation is also a huge area of focus, and was the number one area of risk cited in Aon’s Global Risk Management Survey 2015.
“It’s a difficult insurance problem because the risk is squishy and hard to understand,” Mildenhall said. “There are some interesting examples with cover – for example, if you get a certain number of negative comments on social sites, then you’re paid a fixed sum as compensation.”
The result of this digital disruption presents a vast opportunity for using data to innovate existing and future products, Mildenhall said.
“A great example is in property risk, and using data to drive catastrophe models, which assess a portfolio of risk for things like typhoons or earthquakes,” he explained. “Over the last 20 years, those models have been developed from slightly sophisticated spreadsheets into massive systems that underpin the whole business globally. They created the market by being the common currency of risk everyone agrees on and uses.
“As we look forward to emerging risks like cyber liability, in order for that market to develop effectively, we need to follow the pattern of catastrophe models that everyone can agree on. Once you get to that, you can measure risk consistently, come up with transactions, use it for risk mitigation and so forth.
“That enables the market. But it’s the technology and data sitting behind that enabler.”
What’s transformative about big data is that it open insights into end consumer behaviour, Mildenhall continued.
“In the past, from an insurance point of view, we had to rely on not observing the thing you actually wanted to observe – is someone careful, prudent, do they have good foresight and so on. You just observed if they did or didn’t by whether they claimed,” he said. “With all the data now available, you can directly observe many more of these behavioural characteristics that will be good drivers to understand risk.”
An example of this is credit, Mildenhall said. “The reason that works so well for insurance is because it’s a lot of the same behavioural characteristics we are looking for,” he said. “If someone is impulsive, for example, it can be picked up directly from credit and applied to insurance.”
The more you understand a person’s discretionary behaviour, the greater the chance to design something to suit their needs, Mildenhall said. A recent and successful case of emerging insurance opportunities from the UK is pet insurance.
“People are very attached to their pets and want to do the best things for them, so the product has both an economic and emotional attraction to people and that’s made it successful,” he said.
“Over the next 5-10 years, as we collect more data about different aspects of how people make decisions in their everyday lives, we’ll be able to glean much more insight into personal risk attitudes that will help drive better defined insurance products and potentially, coverage in new areas. That’s the promise it holds for the future.”
Making data analytics projects successful
As a data analytics expert, Mildenhall is full of advice on how to improve the practices and skillsets within organisations to cope with the big data explosion. One of the challenges with too much data is you can quickly spiral into “analysis paralysis”, he said.
“We’ve found that you need a balance between the ability to do a certain amount of adhoc querying… but also guide people by distilling things down to two or three key metrics they can use for decision making,” he said.
As a result, much of work Aon’s data and analytics teams undertakes is coming up with ways to clean data and make data consumable, so people can query it and use it to drive different decisions, Mildenhall said.
“A number of applications here are around dashboarding, benchmarking, and the ability to pull out and have data presented to the decision maker the moment they’re making a decision in way that’s relevant,” he said.
When looking at an analytics project, Mildenhall also noted that 80 per cent of the work is around the data sets, and 20 per cent is analytics.
“The mechanics of ensuring your data is all consistent, and that it can be merged effectively, is not a sexy thing, but that’s the largest impediment,” he said. “Getting data into one place and normalised so you can link to other data sources is vital. Once you’ve done that, the analysis part can be done quickly.”
Like many in the marketing space, Mildenhall recognised the challenges of finding the data and analytics skillsets required for this new line of big data work. He noted the Singapore Government is taking a proactive stance on this front, and that Aon is also actively striving to develop skillsets internally.
“There needs to be a change throughout the system so that we generate people who are data savvy, who can understand looking at both structured and unstructured data, who understand the techniques for data cleansing and manipulation as well as statistical modelling techniques,” he said.
“We’re emphasising the need for training, internal development and that again is a reason why we wanted a critical mass of people in one place [through the analytics and innovation centre in Singapore]. There are different communities in our centre focused on things like data dashboarding, data acquisition, statistical modelling, so we can build our team as quickly as we can while making sure we share ideas across the groups as effectively as possible.”
Mildenhall highlighted the variety of skills needed around data today, including those to assemble the data, asking questions of it, and model building. Likewise, getting the relationship with lines of business right often comes down to keeping things simple, he said.
“Ultimately, if you’re not communicating the insights out of your data effectively and in a consumable way to your end users, then all the work you’ve done is basically useless. It’s not going to impact decisions,” he said. “It’s important to step back a bit from the details.
“You might lose a bit of subtlety, but the gain you get from it being used in the business more than outweighs that.
“Often, business teams operate by rules of thumb. Look carefully at these, as there’s generally an important grain of truth there. And if you can talk in the same language they’re talking to you in, you can pick up ideas and show them how you can help. That’s always much better received.”