This summer I spoke with Nick Hamwey to discuss his career and personal journey with Telematics. We also discussed some advice for young actuaries and the best technical skills to keep in mind.
Nick has been a Fellow of the Casualty Actuarial Society since 2018. He has spoken at various events across the actuarial and data science field and is a member of the Exam Committee for the CAS. Nick speaks at conferences almost every year and has found that if more than a year passes, he “starts getting that itch,” and the need to get out there again and speak about something he is passionate about. Since Nick has started at Cambridge Telematics, he has only spoken at CAS events but in the past, working at Data Robot, Nick spoke at many virtual conferences, including the Insurtech conference.
Q.) So Nick, how did you get started with actuarial science?
A.) I started out at the Hartford as both a Data Science & Actuary, working with commercial lines, predictive modeling & machine learning. Then I went to Data Robot. And I worked as an actuarial data scientist. That was the title when I was over there. And then for the last two years now, I’ve been working at Cambridge mobile telematics focusing more so on their risk and underwriting based analytics and products that they sell to insurance companies, typically focusing on personal lines and commercial lines.
I didn’t have any internships, so I started with a temp position at the Hartford. They said, alright, we’ll give you a shot and see how it goes. That went well and I got hired into their full-time program. I know that is not the path most people take nowadays, now you must get an internship earlier on. And now more colleges are better prepared at getting people involved earlier on.
Q.) And what would you say made you want to be an Actuary?
A.) I actually didn’t even know what an actuary was until I was a junior in college. So, I didn’t go to college with the intention of becoming an actuary. Back then, I was just trying to figure out what was going to work for me. Freshman year, I found that I was good at statistics, and then junior year, I had this moment of- ‘how the heck am I going to pay back my student loans when I graduate?! I better find a job, or something that I’m going to try to get out of college.’ And then actuarial kind of fit in with statistics at the time, so that’s how I got started. It was really just trying to connect the dots between my statistics major. I originally started as an astronomy major, then I went to business, and then I went to statistics. When I finished statistics, while I was still in college, I took my first actuarial exam and I passed it – I think by like one point or something – I’d barely passed it.
I’d never had to study for an exam as hard as I did for that one, in my entire life. I looked and there was a very low pass rate for that exam. I was super lucky to ever pass it, because if I didn’t pass that exam, I don’t know if I would have continued down the path.
Q.) So in terms of your exam progress, would you say you got your designation pretty quickly? How were the exams for you overall?
A.) I mean, I definitely know people who got it faster than me but I still think I was probably towards the quicker side of it, it took me six years start to finish. I think I only failed one or two exams. But even failing one or two along the way, you can go through phases of suffering and then acceptance. You know like, ‘oh, well, I’m going to go back and take that same exam again next year.’
Q.) So you’d say the exam process taught you a lot? Beyond the actuarial work?
A.) Definitely, the exam process showed me that if you put your mind to something, you can accomplish it, right? We can agree you have to be intelligent to pass those exams. But it’s more, it’s more about how much work you put into them, even more so than your intelligence. I know a lot of brilliant people who tried to just take them without really studying and couldn’t pass them. And I know a lot of people on the other side, not naturally as gifted but definitely hard workers, and they were able to pass them because they put in like hundreds and hundreds of hours to pass those exams. So, I thought that was a good core message that came out of taking those exams, hard work can pay off. And it’s not always about whether you’re born smart or not. It just matters how much time and effort you put into something.
Q.) Agreed, so what made you interested in telematics?
A.) I mean, I’ll be honest, I wasn’t actively seeking to work in telematics, I was reached out to later on. In hindsight, it was actually a really good fit, and I remember hearing about telematics when I first started in insurance, when I was working for Commercial Auto.
That was one of the things I brought up in interviews. They’re like, ‘so what do you know about auto insurance?’ And I would tell them I heard about this thing called telematics, that was back in 2012 and was when telematics wasn’t really a thing yet. I believe only Progressive was doing it at the time. But you know, to everyone it was driving behavior, the way of the future, so, I thought it was a good time to get into telematics and it’s really taking off recently. I think there’s more than nearly 15 to 20 million people or policyholders in the US who now have telematics, which is a huge difference in the last 10 years ago, especially with connected cars coming around.
I find Telematics very interesting, there’s so much innovation that’s going on right now in the sector. At our job I feel like every day, we’re like, what are we going to work on next? And we’ll put our heads together and we’ll come up with a brand new project and a new product. There are so many possibilities with the driving data. It’s cool and it’s breaking out of the box of traditional insurance.
Telematics gives you a different angle into all these different pieces; we’re looking at things like what is your fuel economy score look like? And giving feedback while you’re driving, relative to your fuel consumption to get better risk scores. We’re even looking at other things like, telling you which routes that you take are going to be the safest based upon how many people have gotten into car accidents on different routes. So, it’s kind of cool that we’re doing all these things that I don’t think anyone’s really thought about before.
Q.) That’s really cool. Now, I guess you already touched on this because your first role is more so in Commercial Auto, but are there other things from your past experience that you’ve seen play into your current role? Or how it’s been helpful for you?
A.) Definitely. When I started off, it was traditional actuarial so we were just doing things like rate filing indications, so I never worked on the claim side of the house. I think a huge core element is understanding how pricing works. That’s sort of the starting building block. Later, I took a role at the Hartford in data science where we did predictive modeling using machine learning models.
Using a combination of pricing and predictive modeling is really what I do on a daily basis at Cambridge mobile telematics. We’re building, we build models for everything. For an insurance company to actually use one of the products that we built, it has to conform to actuarial standards and insurance pricing standards.
Just because you can build the best model, doesn’t mean an insurance company is going to accept it. And because we’re selling to insurance companies, if they don’t buy our stuff like, we just wasted our time, right? So, I think those pieces were very core; have a core understanding of how pricing works, and then have an understanding of how data science works as well and try to fuse those two together.
Q.) What are the biggest differences that you’ve noticed between actuarial science and data science? Or do you really see them working together?
A.) I think differentiating them is sort of arbitrary at this point. I think there’s a lot of overlap and to say they’re different is kind of silly. I know at my company we call one person a data scientist and another an actuary, simply because one has actuarial designation and the other doesn’t, but we do, for the most part, work on the same stuff.
Maybe core data scientists work on some of the more complex machine learning models than actuaries do. I’m not really building neural networks all the time. I have built them but we have a team of data scientists who are more focused on the models that don’t need to be as interpretable. I would say they’re more focused on accurate predictions, so I think that’s probably the biggest difference.
I think actuaries need to embrace that there’s not necessarily a choice of whether you go data scientist or actuary. My viewpoint on it is you either take it all in, or you might be left behind eventually. I think there is always going to be traditional actuarial work and just using Excel to do things. But as data becomes bigger and bigger, we can’t just use traditional spreadsheets anymore. We have to figure out how to model everything, because that’s what everyone’s really doing.
Q.) I think that’s a really interesting perspective. What would you say some of the big challenges you face with Telematics?
A.) We definitely have some ‘unique’ challenges that other traditional insurance companies probably don’t normally have. For one, since we’re an insurtech, we don’t actually have traditional claims data. So, we have to foster relationships with our customers to get them to collaborate with us on claims and underwriting related analysis; that can be a pretty big challenge.
That was not a challenge at the Hartford, I had a whole database of millions of claims I could look at whatever I wanted. But I didn’t necessarily. At Cambridge, I have access to the latest and greatest cutting-edge technology. I don’t have all the restraints in terms of the way that I code or what tools I’m using. But it’s mostly because we’re more like a tech company, it’s not like a traditional insurance company that needs to put additional rules and restrictions on how people work.
Q.) Would that be your main advice to traditional insurance companies as they work to keep up with innovation?
A.) Yes, I think traditional insurance companies need to be less restrictive. Because I think it actually diminishes innovation, by only allowing you to use a certain set of tools.
You have to get permission I think that actually holds back insurance companies from being more innovative.
Q.) Are there other challenges you face within Telematics?
A.) Another big one is that data is getting bigger every day. Telematics data is huge. To put it into perspective, we collect data from every single truck trip that people take for those who participate in programs with us. So we are collecting data from anything to one hertz, which is one data point per second up to 100 hertz, which is 100 data points for every second that you drive.
You can just imagine, if I’m collecting at 100 hertz, and you took an hour drive, right? So you’re multiplying 60 times 60 times 100. And that’s how many data points I got from your trip alone.
We are storing all that data, trying to analyze that data and make sense of it. So a problem we seem to be constantly dealing with is how do we efficiently analyze the huge amount of data that we have?
Maybe another data point to give you; I think, last I heard we have something close to like 30 petabytes of data in the Amazon cloud. And add into perspective, I think there’s that 1000 terabytes, and petabytes something like that, which is more data than anyone really needs to have. We have a ton of data sitting up in the cloud. And Amazon loves us for that, we have to pay them to keep it there because it might come into play later on. So as data gets bigger and bigger, we also have to figure out how to efficiently make sense of it.
Q.) So, how have you seen the industry evolve as a whole? More specifically, in terms of software and technical skills?
A.) When I first started off at the Hartford, I think from a modeling perspective and data analysis perspective, everyone was using some of the traditional tools like Excel, SQL, and out of the box, modeling programs such as Emblem. I know I used this other WTW product for a while that was called Igloo, which is specifically for capital modeling. More recently, I’ve seen a shift to using open-source packages like R, which was very popular with actuaries. But in tech, Python and C# are the more popular packages that are used, SQL is still used as well and I don’t think SQL is going away anytime soon. But that’s just how you get your data together. Then you have a choice of what package you are going to use for your modeling and visualizing your data.
That said, I think what I’ve seen in tech is the shift towards Python.
I don’t use R at all anymore. I actually think I might not use R ever again, to be honest with you. My recommendation to younger actuaries is to start getting good with Python early on. My personal opinion is that R is just not designed to be coding language that it can. Python was designed for large scale items, but R was not.
I know in the past, there has been a lot of talk about R being way ahead of Python for bayesian statistical modeling, but I think it’s caught up, if not surpassed R recently.
I think there’s really no reason to use R anymore, although, that is a really controversial statement and there are lots of Actuaries that would argue with me about that. But that’s my opinion.
Q.) I’ve seen a lot more employers asking for Python recently.
A.) Yes, the other thing I would say if actuaries want to stay a step ahead and stay involved, they need to think about software engineering, not becoming a true software engineer, but just some of the software engineering practices.
If you are a notebook coder, which is how a lot of Actuaries are trained, that can only really get you so far. I’ve spent a lot of time recently studying it, just putting a little bit of time into writing even cleaner code and it has paid off a lot in my personal work. You’re just building, when you go to code you’re building something that becomes more and more complex over time. And if the foundation of that code was junk, you just scrapped code together and its going to come back and bite you when you try to make it more modular and more flexible with additions later. So, if you build it off scrapy coding, it will end up slowing you down, but if you put a little bit of effort into it and put some key components from software engineering into how you design that code, it will become more sophisticated, and others will be able to use it too.
Q.) Now, a more open question, what have been some of your favorite moments of your career?
A.) I’ve made a lot of good friends along the way. I met a bunch of people at The Hartford who I still reach out to. I also met a lot of cool people at data robot and got a lot of exposure there. And now I work with a lot of very smart people at Cambridge. The connections have been really cool for me, you should always try to collaborate with people along the way. I think that’s been big. I started off being very much an introvert. I want to work on my own project and I don’t want people’s feedback, because I know it’s right. And I think the more you open up and talk with people and collaborate, the better you’ll get at your work and the more friends you’ll make along the way.
Speaking at conferences has also been interesting as well. To get out there and see what other people are working on. I was not a natural speaker at all, I still don’t think I’m really a natural speaker. But I’ve definitely gotten over the immense fear of public speaking. And I think one thing I learned from speaking at conferences is if you put a lot of time and effort into your presentation, you’ll be surprised who you meet.
I’ve always had at least one or two people that come up to me, either virtually or after an in-person presentation. They’re like, ‘wow, that was very interesting, I’m working on something very similar’, and you kick off a conversation and it’s great for connecting the dots. When you feel like you might be working in a vacuum, you’ll be surprised there’s other people thinking about the same problem you are and trying to solve it with different approaches.
Q.) What is a lesson that you’ve learned early on, and then something that you feel that you’ve learned later in your career?
A.) I think early on, I learned a lesson that if you make a mistake, you should tell everyone about it as soon as possible. Otherwise, you will regret it later. Don’t try to hide your mistakes, especially if you know you made a mistake. I think one thing I learned later is that you should try to reflect on mistakes you’ve made more generally and try to do that regularly. I still don’t do that 100% of the time. We call it retrospectives at Cambridge Mobile Telematics. I’m not in sales, but I help support sales deals so if a deal didn’t go through, or you caught a bug, or you miscommunicated something to a customer or you wasted resources on a project that was just never going to work in the first place, I think it’s always good to do retrospectives of what went wrong so that you can try to not repeat that in the future. You can make yourself a more valuable contributor over time by learning from those experiences. I think keeping a journal is similar and actually reading through it and reflecting.
Q.) Any advice for actuaries interested in insurtech?
A.) Yes, I think just goes back to some of the pieces I mentioned before. You should think about coding in Python and think about adopting, or just looking into, some software engineering best practices. There’s a ton of books out there, one book that I still need to finish reading, but I found it to be very helpful for me writing better code is called “Clean Code in Python”. And there’s another one called “Effective Python”. To be honest, they’re kind of hard to read, since it can be hard to read about coding best practices, but I’ve definitely implemented all those into the way that I work. And I found them to pay off very well.
Another thing is to just put yourself out there and try to present even when you don’t want to, because I found at least from personal experience, the more you just force yourself to do it, the better you get at it, and it becomes more natural as you go. You don’t get all those nervous butterflies in your stomach anymore. I will say, I still do sometimes, but nothing like when I first started. Back at The Hartford, I would dread company presentations and I’d be like, ‘Oh my goodness. Why did my boss nominate me for this?’ And now it’s kind of like, Oh, whatever. I’ll just give a talk to everyone at the company about something I care about.
Q.) Do you have any advice for college students or those just starting out as an actuary?
A.) For starting actuaries, I would make sure you really want to go through all the exams.Do you think you’re going to have the dedication to study for all of them?
Because I found that out the hard way; I passed exam six, and I was told exam seven was going to be easy. And that was my first massive failure. Heartbreaking. So yeah, I don’t know if I could have taken another failure. So just make sure you’re committed going in. And if you are, then you’ll do well.
Q.) How about for actuaries looking for a next step in leadership?
A.) I think that you should always be willing to see what else is out there if you’re feeling a little stagnant in your current role. That’s how I made the leap to data robot from my first role. I don’t think you should go searching every year, but just see what’s out there, especially if you feel like you’re not learning anything, you can always see what else is out there to evaluate your worth.
Interested in sharing your story or learning about a specific topic? Email us at – blog@dwsimpson.com