Todd Curry, CEO of the Aon Centre for Innovation and Analytics, believes the art of data science is clear and effective storytelling.
We spoke to Todd Curry, CEO of the Aon Centre for Innovation and Analytics (ACIA), about how data science is transforming how the company makes decisions.
Curry said that the key is finding the gold in the river of data, searching for clues across millions of data points, and often these can be nuances and insights that are vital to how the insurance giant can grow and identify risks and opportunities.
He said the work of ACIA has been transformative, not only among the team of data scientists and executives but also in terms of the tools and learning that can be shared across Aon.
Insurance giant Aon is using data analytics and innovation to transform its business.
‘The best analysts are those who tell stories with data and, let’s face it, we as human beings are very compelled by stories’
– TODD CURRY, ACIA
Aon is a provider of risk management, insurance and reinsurance brokerage, human resource solutions and outsourcing services. Founded in 1919, the company has grown to employ 66,000 people worldwide.
Aon has invested US$350m globally in boosting its analytics technologies and capabilities.
The ACIA in Dublin employs around 150 people and last year announced 30 new jobs.
Leadership in data science
Curry describes himself as a something of a tinkerer; he was always ripping machines apart to see how they worked but not always putting them back together again. He grew up with the tech industry, recalling VAX mainframes and, in his early career, worked with Boston Consulting where he learnt core lessons around analysing data, solving problems and evolving and growing tech companies.
“Analytics has been at the core of everything. I suppose I never left my geekdom behind. I still program and I have a passion for what we do here.”
As head of a 150-strong organisation responsible for analysing millions of data points every single day, for Curry the road to developing clear and accurate insights from data is often a rocky one.
“Any good analytics leader develops some scar tissue from mistakes they’ve handled in the past, and we’ve tried to address those head on.
“First, we try to time-box our efforts. It’s easy to be in search of an answer and to keep going after it for ever and ever. We like to time-box to take bits of problems and see what we think we can solve. We are constantly evaluating, do you think we can get there; how good is a low-resolution answer, a vague plus or minus 20/30pc answer versus six decimal points? That for us is an important part of being pragmatic with our own data.”
He said the resulting insights deliver their own return on investment and can be a source of enormous pride within the ACIA.
“We try really hard to make sure we are working on the biggest problems. That’s allowed us to focus more on how we work as a team. Earlier this year we made a switch from more waterfall-like working, where you set a big goal and everybody marches up the hill to meet it. Our technology team moved forward with Agile Scrum at the beginning of this year and we felt that not moving the whole centre would create a two-class system and it was important for us all to be on the same framework. So we moved the whole centre here from what was formerly a software development methodology to what is now a business-centric methodology, which is Agile Scrum.
“The result has been people are very engaged, happy and accountable to each other. We are seeing a lot of esprit de corps, which makes it a very fun place to be.”
Cutting through the hype around big data: find the narrative
While many commentators have speculated on the challenge of big data, made ever more complex by rising amounts of data, Curry is pragmatic and believes goals and parameters are key.
“I think a lot of those run-rate estimates on the growth of data are skewed by folks using GoPro cameras to upload their entire lives. I take a dim view of the terabytes and exabytes argument because not every snowboarder’s GoPro experience is going to yield great data to drive the insurance industry.
“I tend to avoid the big scary view, but we look at data in a few ways, one is dealing generally as an industry with highly-structured data; two: our data tends to be entered by human beings and that means they are prone to errors and delays, which are a natural part of human data entry. We look at what we have, how we can improve it and create alliances around data in the industry. The clean-up, as any analyst will tell you, is as much [work] as analysing the data itself.”
Ultimately, the job of good data analysts and scientists is communicating a narrative, being good storytellers.
“The best analysts are those who tell stories with data and, let’s face it, we as human beings are very compelled by stories. I encourage my team to spend as much time [as possible] focusing on the acceptance of their work and how to make it clear and unobjectionable.
“Spend as much time on that as they can on the actual analysis; so that’s a challenge, a constant challenge. We are asking questions constantly of ourselves: is this clear? Is it true? Almost like sales, make sure you anticipate all the objections that are going to come.
“I think the best analyses are the kinds that really can have a good clear hypothesis and bring clear data, and I try to make sure everybody understands that.”
Siliconrepublic.com’s Data Science Week brings you special coverage of this rapidly growing field from 28 September to 2 October 2015. Don’t miss an entry worth your analysis by subscribing to our news alerts or following @siliconrepublic and the hashtag #DataScienceWeek on Twitter.
Get your early bird tickets now!