Q&A with Michael Xiao, FSA,CERA,MAAA

What made you pursue a career in actuarial science?

I had a nontraditional start to my career. I was actually a film major in college, and right after I graduated I went to China to shoot a movie. Before I left, my mom suggested I look into the actuarial field as an alternative to a film career in case it didn’t work out. I did some research and found that being an actuary was a combination of business, computer skills and math, which fit my interests really well. Though the experience filming in China was interesting, the movie wasn’t very good, so I decided to look into actuarial jobs and started working at my first employer, Cigna, shortly thereafter.

You started in a more traditional actuarial role; what piqued your interest in predictive analytics?

I was in the rotational program at Cigna and worked in traditional roles in underwriting, pricing and forecasting. When I moved to Aetna, I got involved in the implementation of the Affordable Care Act. That’s what really sparked my interest in predictive analytics. If you think about traditional actuarial modeling, many of those models are based on aggregated data rolled up into a higher-level model. I noticed that a lot of the granular data and lower-level modeling wasn’t being utilized, and I was interested in how to use that level of data. Aetna formed a data science organization in 2014, and that’s when I put what I’d learned into use and took on a leadership role within several predictive analytics projects.

Was there a point in your career that steered you toward what you do now at Blue Cross Blue Shield?

There wasn’t one single point in my career, but there was a culmination of experiences. I had been interested in computers since I was a teenager, and I took a lot of computer science classes in college. The actuarial roles I took on were really exciting and involved building models. But when I was working on the ACA, I started to do research on individual underwriting because it was becoming less common. It got me thinking: “What’s a better way to potentially underwrite individuals?” That’s how I moved into predictive analytics full-time – modeling specifically for individuals.

What has been the biggest challenge as an actuary transitioning into predictive analytics?

The biggest challenge has been explaining the value of actuaries in this space – what it is that actuaries provide that a traditional data scientist with a Ph.D. in statistics or mathematics doesn’t bring to the table. We have a lot to offer. Organizations that have incorporated actuaries into their data science roles see the value immediately – we have a deep understanding of the data, and we know the business as well. Actuaries have the prerequisite talent to do all of the things these jobs require, but the key is whether or not they have invested time in learning the skills that have emerged in the last 5-10 years.

What are the specific strengths that actuaries bring to predictive analytics?

There are three areas where actuaries are very strong. The first is self learning. As actuaries, we’ve gone through all the required exams, and generally speaking, we’ve had to do a lot of self-directed learning, buckling down and taking the time to learn something we haven’t studied before. That’s really valuable in predictive analytics because it’s still a new field and there aren’t as many educational resources available as there are with other professions.

The second area is that we have a deep understanding of the data and the context of that data. Sometimes when you bring someone in from another background, they don’t have the insurance knowledge to allow them to catch inaccuracies immediately.

The third strength actuaries bring to the table is that we have a deep background in business. We might be thought of as the math nerds, but we have to understand how our businesses work, whether that’s healthcare, life insurance or another industry. We talk to our customers, understand their pain points and distill easy-to-understand conclusions from complicated models — we tell them both how we got to the answer and the business reasons for it, not just what the answer is.

What’s the most rewarding part of your work?

I get to do new things almost every day. I don’t just update a model – I get to think of new ways to solve existing problems and come up with ways to solve new problems because it’s a constantly changing field. If you like working outside of a routine, this is a great field to enter.

Where do you see the actuarial profession moving in terms of predictive analytics?

I see a lot of great investments in the future of predictive analytics, especially within organizations like the SOA and the different certification programs for predictive modeling. These programs are designed to set actuaries up for success. Beyond that, in the next 5-10 years, we as an industry need to make more investments in machine learning and big data.

What advice would you give to actuaries who want to take a nontraditional path in their careers?

You have to shift your mindset to take a nontraditional path and move into predictive analytics. You’ve passed all your exams and you’re comfortable with traditional roles, so to step outside that you need to take a risk. And once you’ve taken that leap, it’s critical to understand how to sell yourself as an actuary. When you’re discussing skills and careers with someone highly technical, like a hiring manager on a data team, you need to be able to explain what you bring to the table that they don’t already have on the team — business skills and a deep understanding of the data. That combination is the foundation of success in predictive analytics.