#117 Successful Data & Analytics in the Insurance Industry

Hi, I'm Maggie, VP of Curriculum at datacamp, and I know firsthand how overwhelming it can be to keep up with the rapid changes in data and AI. But staying ahead doesn't have to be hard. That's why our curriculum team has curated the best learning content to help you and your team stay ahead, covering everything from data and AI literacy for busy executives to cutting edge generative AI for developers, and everything in between. Whether you're upskilling yourself or your team, we make learning practical, engaging and flexible. It just works. Power your data and AI transformation journey. On datacamp.com. You're listening to Data Framed, a podcast by Data Camp. In this show, you'll hear all the latest trends and insights in data science. Whether you're just getting started in your data career or you're a data leader looking to scale data driven decisions in your organization. Join us for in depth discussions with data and analytics leaders at the forefront of the data revolution. Let's dive right in.

Welcome to Data Framed. This is Richie. Today we're talking about data in insurance. This is a really fascinating industry because it's drowning in data and there's so much opportunity for frankly, making a lot of money by using this data better than your competitors. At the same time, the industry has tricky challenges from legacy processes and technologies as well as regulatory constraints. Telling us how to thread the path to data success is Rob Reynolds, a VP and Chief Data and Analytics Officer at WR Barclay. Rob is an industry veteran and has a lot of experience managing various data teams, from business intelligence to data science and to data governance. I'm keen to hear his insights on data success in insurance. Hi Rob, thank you for taking time to join us today. Just to begin with, I'd really like to know a bit about what does WR Berkeley do?

Sure, Richie, and thanks for having me. So WR Berkeley is a multinational insurance company. We do PNC insurance. What that really means is just property casualty. We serve customers that are mostly commercial and we're in lots of markets across the globe, some more than others. Like we have a presence in Europe certainly, as well as South America and of course in the United States and Canada. So Berkeley is simply a large insurance company.

That sounds fairly straightforward. And then, so your role is Chief Data and Analytics Officer. So what exactly does that make you responsible for?

It's an interesting organization. As Chief Data analytics, an analytics officer, I think my responsibilities are diverse. You can, you can read into them based on the fact any organization has some purview over analytics, data engineering, and those types of things like Any other organization, we are highly dependent on it and partnering with it, as well as different business stakeholders that we serve. But at Berkeley, it's an interesting org in that we have a corporate organization, and underneath that corporate organization are a number. And when I say a number, right now there's 56 operating units underneath the Berkeley umbrella. And those operating units operate autonomously. So in many respects, their data analytics and it can be pretty well disaggregated. So while my role looks and feels a lot like it would at almost any other insurance company, there's an added wrinkle there of trying to match our capabilities and the things that we're trying to do as an enterprise to the needs of very specific operating units, and many of them to boot. So, long story short, it's the challenge of data and analytics at any kind of large entity with the what I would say that the wrinkle of trying to be very localized and offering very specific capability to those that are underneath that corporate umbrella.

Okay, so I'm always fascinated with how large organizations structure their analytics teams. And so you said you've got 56 autonomous teams and then you have a separate central data team, Is that correct?

Yeah, so we do. And that's, I think, part of the balancing act for us. Right. With 56 separate units, I don't think there's a scenario where we can effectively deliver centralized data and analytics. Right. So a lot of organizations, they make a choice to centralize certain functions because obviously that offers some scale and leverage. But within this organization, it's a little different. And I think the balancing act that we're trying to strike is to centralize the things that make sense. To centralize meaning things that really can offer us some value. Once they are common, they offer us an opportunity to scale things like that. But at the same time, in keeping with our corporate culture of entrepreneurship and really serving the specific markets our operating units are intending to serve, we need to keep that which is valuable as close to the customer as we can. Right. So we serve markets like operating units are in specific areas of the insurance industry, be it geographic or lines of business or what have you. And they are there because they believe there's a specific opportunity for them, and there's specific value to being close to those customers. So it would be really foolish, I think, for someone like myself to try and ignore that or work against that. I mean, we've been a successful organization to date, and we want to keep that going, and I think that's part of our secret sauce. So really, the Tricky part is how do we figure out how to centralize what we can and then decentralize that which we should in order to meet those needs, right, those needs of those customers. And so we have a central team for sure. But that central team works as much as we can with decentralized teams. And those teams are at an operating unit. Those teams might be at what we call the segment level, which is a. You can think of them as an aggregation or grouping of operating units that are common in one way, shape or form or another. And so we partner at different levels of the organization as it makes sense to do so, to try and bring some consistency and commonality where it again makes sense to do that. And in other instances, instances we also may provide very specific kinds of support to an ou, an operating unit that has a need for something and they can't simply do it themselves. So we wear a lot of different hats. In many respects we're consultants and in other respects we might be more kind of governance oriented. And it just depends on the context that we're talking about.

So it sounds like a really great thing to have all these analysts like pretty close to the customer. Probably something you worry about is how do you make sure that all these sort of separate autonomous teams are not duplicating work and how do they share best practices and communicate with each other?

Yeah, absolutely. I think that's the tricky part, right? We want to allow innovation to occur anywhere and certainly we want it to occur at the operating unit level. So to put it more tactically, if there's a team of analysts or data scientists doing work at a specific operating unit, we want them to be able to do what they want and how they want to. But we also have to find out what are those kind of trigger points or what are those kinds of barriers that if you cross them, you really need to think differently. So that's what we're trying to, for the organization is we're trying to define some ground rules in specific domains of data and analytics where you can say, look, here's where it's safe to play go crazy once you step across this border, take a beat to figure out whether or not we need to engage someone else to ensure that there's either consistency or commonality. And at the same time, if you need to phone a friend, how can you do that? Right? So we're literally, I've been here for about two years and just eclipsed the two year anniversary. And we started out just trying to get some things done and take some very Specific targeted action in some areas of IT and data specifically. And now we're starting to think about how do we refine our strategy to do what I just said. Right. So that that would be through some standards that we would actually generate and distribute and communicate. And that would also be through things like centers of excellence. And that's a common term thrown around. I think of them more as, you know, you can call them either centers of excellence or some sort of almost like a meetup group. We want it to be a little more informal. Right. Like, I know that sounds a little silly, but there is that element of we want to ensure that there's autonomy and there's a way for people to share ideas and we want them to almost be these groups that stand alone on their own without requiring corporate intervention or some sort of corporate mandate. So it's hard to figure out how we're going to do that. We're actually just starting that. We've done that in a couple of areas, most notably in data architecture. We started there. But it takes time, right? It takes time for people to hear about it and then see value in it and keep coming back. And then it takes a little bit of time for them to add their own to the equation. But like my team members have already seen where people are reaching out to them a little more proactively if they do need to phone a friend. So they're starting to see, oh, all right. And if we're gonna do something in the cloud, how might I do that? Well, I don't know. I know this guy over in Rob's team might know something about it and they make that phone call. We'd like to make it a little more fluid and a little bit less friction in that process. But that's essentially where we are now and where we'd like to go tomorrow.

I really love the idea of having just sort of internal meetup groups where people can just sort of chat and share knowledge like, and have that sort of social aspect. Just taking a step back for a moment, I'm curious as to whether you have any kind of overarching goals within your part of the organization. So do you have like a mission statement or is there some sort of high level strategy that you're trying to work towards?

Yeah, so at a high level, what we're trying to achieve is we want to create some enterprise scale for the organization. Heretofore that's been a challenge because we have lots of, like I said, we have each of our ous and some aggregations. Therein of data, of analytic talent and the like. And while that's great, it does still pose challenges if you're trying to see things across the enterprise. Right. And in some cases, an insurance company. Right. You do need to look across the enterprise for certain things. So while we may organize internally by ou, that's not to say that legislative bodies don't expect us to report certain things across those boundaries. Right. They don't really care about how we kind of run ourselves. They say, look, if you've got an insurance company, from like a legal standpoint, you need to report certain things in that way. And so that makes you cross boundaries. Right. And so that's me coming in here day one. That was the kind of most tangible challenge that we saw because it creates a lot of manual effort not having your data together, not having it governed in one kind of singular or even three ways. Right. Like, it's very diverse, very difficult to traverse. And so we said, at a minimum, we need to bring some harmony to our data infrastructure. And so we began there. We began with a cloud journey as a result, because very clearly, as we are bringing data together, we would need to make use of tools and techniques that could really be facilitated by the cloud. And so that's where we started. And that continues to be one of our chief goals. Because obviously, bringing some enterprise scale to what we're doing, be it reporting or analysis or statutory reporting and the like, I mean, all of it is made more easy when your data infrastructure supports that kind of activity. But at the same time, I guess so if you think of that almost in terms of the data side of the equation, from the analytics side of the equation, it's really about trying to advance our data journey and our analytics kind of sophistication. And that's different in our world than it might be in others, because it's really not incumbent upon us to carry that water. We see ourselves more as a facilitator for other teams that are already doing that work. So think, for example, an operating unit already has a team in place that's done really great work around some analytic solutions that deliver meaningful value. We want to try and figure out, all right, what can we take from what they've done and share it with someone else who might be just starting out in that part of their journey. If you think in terms of Gavin Port's model of moving left to right from less sophisticated to more sophisticated analytics, we're trying to make the organization, in its own way, move from left to right on that chart. And it's going to be a little different for everyone. Right. So an operating unit at one stage of that journey needs different things than another operating unit at a more advanced stage in that journey. And so what we're trying to do is find ways to facilitate that. So I'd say those are probably the two big, high level strategic things we're trying to accomplish that are really like 50,000 foot kinds of strategic objectives.

So you mentioned Davenport's model about sort of increasing analytics complexity. I guess. Can you just talk a bit more about what that is?

Yeah. So Thomas Davenport, depending on who you talk to, is a pretty powerful force around the analytics industry. Right. And published a book a long time ago, competing on Analytics. And he created this model that really shows how an organization could mature their organization through differing levels of sophistication in the analytics space. So you start with. And I'm going to probably get this wrong, but it's. You start with something that is telling you where you've been, right? Like basic benchmarks, the like. Now you move from there to then understand trends, so where might you be going? And then you move further to the right and you're trying to predict where you're going to be going. And then finally you move into kind of the last, most sophisticated space. And you'd say, I want to try and either optimize or prescribe where I want to go. So I'm actually trying to drive towards specific outcomes and I feel like I can do things in order to achieve those outcomes. So very, very different. Right. If you think about it, than just trying to figure out, hey, how are we doing today? Which is where the model starts. Right. It starts with gee whiz, how many widgets did we sell last week? That kind of thing to how can I ensure that Jo buys another widget from me next week? Right. Like very different problems to solve.

That's absolutely fascinating. Actually, just this week I was just writing a blog post on the different levels of analytics from sort of descriptive analytics and diagnostic and predictive and prescriptive. And I had no idea that this is the Davenport model. So that's. So I've learned something there.

Hopefully he pays me like a dollar for attributing it to him. But that's how I learned. Right. I read that book and I saw it as first talked about, at least in my world, it was first talked about in that book for me. And so yeah, I attribute that to him.

Yeah, that's really great. So I'd like to get into some of the sort of details around like insurance industry specific problems. So I've spoken to quite a few people in the insurance industry recently and the one complaint they have is like, oh, well, we got this great data set, but it's stuck on a mainframe somewhere or like it's siloed in one particular team or like it's on paper records from like 1920 or something like that. I'm wondering, so what extent sort of data siloing and having different formats of data affected you and how do you go about dealing with it?

Well, first of all, I'm shocked that there's only one thing that people are complaining about when it comes to insurance. But yeah, I mean, I've been in insurance for a while and like a lot of industries that have not necessarily grown up in the insurance age, right? Or pardon me, in the digital age, we started off on paper. I mean, geez, like if you go way back in time, insurance really is an extremely old industry. You can go and see some of that history at Lloyd's of London and I have, and it's just super interesting. But it's very manual, right? Very, very manual. So we have a long history and B, a very manual history. And so consequently our processes have evolved just how, in a very similar way as technology and work has evolved, right? So if in the early 1900s we were working on policies those were in many respects handwritten or something like that, and then we moved to typing them up and then we moved to computers and so on and so forth, right? So given the fact that we still have a history, right? You look at some life insurance companies, they still have active policies where the policies were taken out 80 years ago, something like that. And the big challenge for the industry is how do you take the data that gets locked into old technology and digitize it, for lack of a better word. We used to talk about like electrifying it and maybe that meant microfiche at one point or time in our history, right? So it's funny, the evolution continues and it's still a very real challenge for our industry. And frankly, I'll say it's actually a challenge we're not always making easier for ourselves. Because quite honestly, you would think that today, knowing that challenge, we would stop doing things from a systems development standpoint that would further exacerbate the issue. But lo and behold, we have systems that again, oftentimes don't really embrace digital technology. Right? So for example, I want to do business with a broker and I want to make it easy for them. Well, the easiest way for me to maybe make it make a broker interaction simple for them is just using email and PDFs, and lo and behold, boom, there you go. You've just got data that's coming to me via email and it's attached in a PDF that may or may not be handwritten like there you. And that's a real thing that we're experiencing now. So you would love to think that we make the use of OCR and NLP and tools like that very frequently, but in reality we're still early stages relative to other industries. I mean, I think the industry's made a lot of progress, but we constantly are trying to figure out how to not only unlock data that's been locked away by old technology and older processes, but how do we do it in a way that stays consistent with our existing processes or can work within those existing processes and is something that remains accessible to all of our partners, whoever they may be.

I have to say, it amazes me the number of times I have to sign something on a paper form, even in 2022.

I mean, think of it this way, Richie. Right. Some industries are attached to other industries, Right. So if you're talking about insurance and you're talking about real estate, there's governmental elements to that. Right. So if you're filing something in real estate with a county, you know what, the county probably doesn't have the same amount of sophistication as say, Apple. Right. So, you know, consequently it's like, yeah, they've got a filing cabinet. You know, someone puts that document that you just signed and sent through the mail in that filing cabinet and it stays there until such time as someone else needs it. And that's the weird world we're living in right now, where you have very different types of modes of operation depending on which industry we're talking about.

Yeah. So there's the potential to do things really efficiently using sort of the latest technology, but it's just sort of not evenly distributed, I guess, from one industry to the next. So you mentioned sort of government and interacting with that. So I think regulatory issues seem to be a very important thing within insurance. So maybe you can just explain what the most important sort of regulations you have to comply with related to data are.

That's hard to answer. I'm not going to lie. I think we are in many respects bound by the same regulatory requirements from a privacy and our responsibility to our customers and our partners. We are bound by the same things that banking is and the like. So for example, we take where does Our sensitive and private information land, where does it get stored, how does it get transmitted? And we have to be keyed into HIPAA because of health care information. We have to, to the extent we take transactions via some form of electronic payment, we sometimes get exposed to PCI compliance. There's a herd of different types of country specific types of requirements that come out of Europe as well as the United States. There are state specific rules that in many respects we have to pay attention to. And a lot of companies, other companies may not have to because they're either not domiciled there or they're not filed quite the same way. So we have to pay attention to, I would say, the regulations and the patchwork of regulations that really exist, regardless of how big that patch is. Right. So insurance is largely state driven. So you have. That's probably the biggest difference from a compliance standpoint. It's that we have to worry not only about federal, we have to worry about states, we have to worry about countries. I mean, heck, I wouldn't be surprised if someone said there was a zip code requirement somewhere. Right? I mean, we get down, obviously I'm joking, but we get down to a very fine grain when it comes to regulation and we have to live by all of it. And so I wouldn't say there's anything actually that unique because conceptually we have to adhere to all the same things that any other company would in terms of privacy and respect to our customers and our partners. The variability that we experience is much higher, I think, than a lot of companies just because of how we operate and how we're regulated on a state by state basis.

I'm curious as to how all these sort of regulations affect your organization, like who has to deal with them, because I know some insurance companies have this sort of separate data governance team and in some cases it's going to be all the analysts and some of the times it might affect other people as well. So how are different teams that you work with impacted by regulations?

Certainly, yes. So our team is by no means an expert in the regulatory side of things. We recognize that those responsibilities exist and we owe it to our corporate partners who are experts in that space to work closely with them. So we have a compliance team, we have a legal team and an infosec team that we work really closely with as we sort through how we move data, how we put solutions together, making sure that we live in accordance with their expectations and are consistent with their expectations as it relates to the various requirements that might exist at different levels of the organization. So we really rely upon them to tell us, all right, what needs to be true here in order to live within this regulation. It's interesting because a lot of the regulation that comes out isn't really that specific. So it allows some gray area, right, where you'd say, all right, I'm making this up, so don't hold me to it. But there might be some sort of a requirement out there from a state perspective that says any data at rest needs to utilize this kind of consistency with this other standard. And consistency with that standard runs when you get into the nitty gritty runs, afoul of maybe a regulation that comes out from Europe. Right. So something like that. You'd say you can't win for losing. Right. Because it is a patchwork. And so we certainly spend a lot of time trying to figure out how do you satisfy the letter of it, but also the intent behind it. Because, you know, obviously the legislation doesn't move as fast as technology. And so you're constantly trying to balance how do we do right by our customers and our partners, how do we do right by the regulators and making sure that their intent is reflected in what it is we're doing. And if you talk to anyone in this industry, it's a constant struggle because everyone's trying to get their arms around it. And it's just one of those things that you can't hope to be expert in it. I think you can only hope to be very good at adapting and working together to try and figure out what that solution really needs to look like when something new evolves.

So have you ever had a case where maybe one of your analysts said, okay, I want to run this analysis, and then you've discovered, okay, we can't do this. This is bumping up against some regulation. It's going to be illegal to run the analysis.

It's not quite like that. I'd say what you could. So while my time in my current role, I've not experienced this, but in my past I have. Where I'd say the things that are a lot more risky in my mind are where we're actually. It's not about the data. Like, everyone has sensitivity around personally identifiable information and HIPAA and all that sort of good stuff. I think it's like when you're building models and you're using certain data in order to answer a question or drive a specific outcome. I do think there's an element of you have to be very thoughtful and push back on each other, even within the team, to say Is the way we're doing this correct? Not only from a kind of technical standpoint, but is it right? Is there some unintended consequence here that we are stumbling into that might not be okay and that we would not be happy with? Right. Just inadvertently. And so that's actually something that I think our industry has to do a better job of in general, is kind of emphasizing the need for thoughtful criticism of our approaches and methods when we apply modeling to different business problems. Because quite honestly, I mean, you look at, I think the things that we hear about now about Facebook and other social media platforms and how they apply analytics and on and how they use that information, it's definitely making a lot of people scratch their head. And I think the same goes for our industry in insurance, and I just think just in general in data and analytics, we need to do a better job of challenging each other when we're thinking of these things. Because quite honestly, I've had a lot of experience with very astute data scientists that can build wonderful models and they really understand technology and they're super gifted with the mathematics, and it's really a great outcome, except for the fact that the consequence of that outcome is not what we ever had intended. And we didn't take the time to take a step back and say, huh, what could happen here if this really was successful beyond the variable I'm trying to influence, you know what I mean? I don't know if that makes sense to you, but that's maybe.

Can you give a concrete example of how using some weird variable is going to have an unintended consequence?

Sure. So let's pretend that there's a retailer that is just doing marketing analysis, right? And they feel like based on certain types of analysis and purchase behavior, they can anticipate the needs of a customer in advance so that upcoming purchases, they can take advantage of mailings and to drive those upcoming purchases for that customer to occur at their store. And so you'd say, all right, on the surface, that's great. If I'm buying orange juice, I am really likely to buy milk next week. So I get a coupon for milk and I go to the same store and buy the milk there. That sounds, on the surface, really great. If I start buying a book on pregnancy and then I start getting discounts and coupons for diapers and the like in a couple of months and over the next couple of months, is that really what we need to be doing? Like, maybe that purchase is for a person in a household and we're communicating something to other members of that household that are private. And if that's the case, like, we're going to drive certain types of purchase behavior in the future. Maybe. But what did we inadvertently do in that case? You know, right. Wrong or indifferent. Right. Maybe we were right. Maybe that person was having a child and they could benefit from those coupons, but they didn't want that information be divulged to others in their household. Or that's not accurate. They are buying that for a friend. And now that information has been divulged in a way that creates a different challenge. So as an example, that's kind of a generic example of a scenario where you could see, like, okay, I get it, you move the needle for sales. That's a pretty powerful tool. But using it in every category of product that we sell may not be a good idea. And someone has to challenge us to think through that a bit.

It does. It seems like getting privacy right is very difficult because information sort of leaks very easily. It's a little bit like when you click on those Amazon affiliate links somehow, and then for some product you didn't really want, and then suddenly you start seeing adverts straight everywhere around the Internet. All right, so may, let's pivot a little bit because I'd like to talk a little bit about hiring and careers as well. So your background, you started as an analyst and you sort of worked your way up to chief data and analytics officer. So I'm curious as to how your roles changed and what you worry about or what you do has been different at the different sort of levels of your career.

Sure. So I would start by saying my career was non. Probably not very traditional for a lot of the people that you might talk to or might be listening to this. So I started my career honestly as an accountant. I didn't care for it. It was a necessary evil because I needed a job, and I knew I could pretty much get one coming out of school. And so it certainly served a purpose. And I learned a lot from that discipline, but I wasn't really excited by it. And so I moved into technology from there, and from there I went into information security and learned about applications, and I found my way over to databases, and I started actually with the security side of data. And that was just interesting to me, and it threw a twist of faith. I did work at a retailer and we got involved on some projects that were really around, like, demand forecasting and the like. And I got pretty deep into the demand forecasting space. And that's where I got my first kind of real case of data and analytics and the power of modeling and the like. And from there it was kind of like game on, right? I really enjoyed the analytic and the data side. My career did also move around a bit still. So I went into insurance and I did special projects and I led some data projects and the like. So it wasn't like I've been in the true data and analytics field my entire career. But I'm thankful because it gave me a little bit of diversity there, gave me an appreciation for kind of problem solving and managing others and managing projects. It gave me exposure to a lot of different business processes. So some domain experience that was valuable in trying to apply data and analytics to business problems. And so yeah, I mean, it was how I got to the point of leading organizations, frankly came from the fact that I was in the right place at the right time. I'm not going to lie. I would love to say that it was all part of my grand scheme of things, but it wasn't. I had an opportunity to take a big leap and lead a data project that was going to be impactful. And so I did that and was successful there and kind of took another leap to start something new at the organization that I was working at the time. And they said, well, we've got all this data, we need to figure out how to use it. Would you be up for figuring something out? And I was like, yep. So I guess the answer is really my advice to others is be willing to take opportunities as they come along. Be open minded in what they may present. While they may not look and feel like exactly what you want, they can oftentimes lead to exactly what you need and get you to ultimately where you want to be.

That is really interesting and I think it seems quite a common story that just getting some kind of a career break. It's very difficult to predict. It's just sometimes it happens and you just got to make sure that you take that opportunity. So just think about when you are hiring people. I'd like to know a bit about what you look for. So maybe you can start with what sort of technical skills do you think are important when you're hiring?

Yeah, the technical skills are always important. So you have to be kind of well versed in data tools and languages that enable you to work with data. So be it Python, SQL, some form of analytics tool set is always welcome. Anything from depending on where you are in your career, anything from something like Excel, which is very straightforward and you might roll your eyes at that. But for a lot of people and a lot of organizations, Excel is still a darn powerful tool and relevant to things that are a little more kind of high powered, like Tableau or Alteryx or something like that. Just from a kind of a data movement, kind of a light analysis, visualizations, that kind of thing. And being able to code, if you're a data scientist and you're going to be able to leverage technology in a more real way utilizing Python or other web tools, that's certainly an element of it for sure. So I would say those are the kinds of technical skills we look for, that we test for and we challenge. Once people can demonstrate that, yeah, they have those skill sets. I would say since technology moves so rapidly, I would suggest that being able to also demonstrate on top of those technical skills that you happen to have at the time, being able to demonstrate that you can learn and adapt to new technologies and new needs, I think is really important. So that's one of the things that we look at part and parcel with the actual technology, because that's what's going to keep that technologist really productive over the course of time as the needs of your organization or the technologies that you employ change.

Absolutely. Are there any other sort of. I think a lot of people call them the soft skills or power skills, like these sort of non technical skills that you think important for analyst roles.

Yeah, for sure. So on the technical side, I think there's. Let's use this as an example. Right. So a lot of people focus on things like Python and the like and they really enjoy going deep into a space, but they live in that space too frequently and fail to consider things that can help them along the way because it's not maybe what they have historically done. And that willingness to change and accept other ways of doing things in light of the objective of trying to get value quickly, assuming that's your objective, limits people from saying, oh well, I'm going to stick build this model in an app and build it using Python and Django and whatnot, when I could actually maybe utilize a cloud service that has that kind of packaged and gets me 80% of the value in the time. Right. So I would say the soft skills there are willingness to change, not only adaptability, but like willingness to accept a different way of doing things, even if it runs afoul of what your interests and passions may be. I would also say that focusing on a business problem and being creative in terms of how to solve it using data and analytics is crucial. Right. Because I think it's easy to check some boxes on technical skills and yeah, you're going to be valuable to someone, certainly. But if I don't know how to apply those skills the right way to the problem at hand, I mean, that value is going to be diminished quite a bit. And I. It could be everything from. It takes a little longer to build the ultimate solution, right? Because you didn't quite understand what the need was to. You're totally off the mark and you gotta throw away what you were working on because you didn't take the time to dig into the problem and ask something like the five whys so that you could really assess what it was you were trying to do. Now, again, organizations are going to be very different, right? You know, some organizations where it's very prescriptive. It's like, hey, just tweak another 2% out of this model. And we're comfortable with you taking the next six months to do that. That's very different than, hey, you need to go talk to this person and help them figure out how to drive more value out of their organization using some data. Right? And that might start with a report and evolve to something much more sophisticated over time. So I know there are a lot of different use cases. I know there are a lot of different types of experiences and organizations leverage skills. I would encourage people to constantly be willing and able to think critically, think creatively, take a step back and challenge themselves to make sure that whatever they're doing is meeting that need.

So, I mean, you talked about making sure that you understand the business problem there, which seems like a very important thing, like for any kind of analyst or data scientist role for people who are just trying to break into the insurance industry. What's the sort of most important bits of like, industry knowledge or domain knowledge that you think, like, everyone should know? Like, what would you be looking for if you interviewed someone?

Oh, wow, that's a challenge, I'd say. You know, it's so funny in the insurance industry. I think there's a bias towards talking to people in our industry that are like, actually not part of our industry because we want to see how other people do things. So there is that. Right? So I think if you're joining the insurance industry, coming from another industry, I think that's really valuable. I think you can use a lot of what you've learned to enhance other people's kind of thinking in whatever organization within insurance you're joining. Granted, you do have to understand some of the nuances around the data. And I guess I would say insurance is one of the few, if maybe the only product that gets sold where we actually don't know the ultimate cost. And that has pretty substantial data ramifications, analytic ramifications. So this idea of the cost of a policy evolving over time is really unique. And I would say probably if there was one thing for folks in our industry to think about, it would be that because it does influence how you analyze something. It influences how you look at data and how you store data and how you architect data. Because you have to accommodate the fact that something's going to be slowly changing and evolving over the course of time. And that has to weigh into your thinking when you're making decisions about your business.

Okay, so understanding cost of a policy is going to evolve over time. That's right. That's the thing you need to remember if you want to become an insurance analyst. All right, so one last question on hiring. What's the one thing that someone ought to do to impress you in an interview?

I mean, I think show initiative. Show interest in what they're doing and how they're doing it. Demonstrate that they're willing to take advantage of opportunities and work hard to get there. I think there's nothing more disappointing than seeing someone who thinks they figured it all out. And I should be just excited as heck to even talk to them. Like, that's just a less than ideal foot to start off on. So I think if you're excited about what you're doing and you have passion about it, that's awesome. And I think you should show that you're willing to take the initiative to help add value and drive whatever organization you're joining and to the direction that it needs to go to finish up.

I'd like to ask you a little bit about the future of analytics in insurance. So how do you see the insurance data landscape changing?

I mean, I guess it goes back a little bit to what we were talking about around data being locked away and kind of old processes and the like. Like I said before, there are a lot of processes that continue to change and evolve, but just not by enough. And that's, I think, largely driven by the fact that we think in still linear terms and our systems work still in sort of linear terms. They don't. Even new systems act that way. I think the biggest shift will come once we have someone in our industry willing to remake the process soup to nuts. Right. And approach it as you would a digital organization. I think there are a lot of companies trying to do that. I think they find that they can get a lot of value out of not overhauling the entire industry and they just try and overhaul a little piece of a process and there's a lot of benefit in that. So they stop there. I think there's eventually going to be though, a sea change in how we think about how we use data as part of our process and we're going to remake those processes and that'll bring some boundaries down in the organizations that are currently big players in the insurance industry. And once that happens, I think you're going to see really material change where you can use data that's enriched in an automated sort of way and make decisions in a very different, using a very different paradigm. I think there's a lot of capital flowing into FinTech and InsureTech. So I think there is no shortage of people trying to figure this out and trying to find a way to crack that nut and get value out of it. I think it will take some time though, because I think we are still very, we have a long memory, we're very historically driven and we have a lot of legacy still in the industry, be it processes or legacy systems. And right, wrong or indifferent, it just, it influences how we think. And I think it will take a bit of time before we really kind of shrug that off and embrace what I would say is a more fully digital kind of way of doing things.

And are there any data skills or more generally skills that you think are becoming more in demand or other ones that are becoming obsolete?

Obsolete? I don't know. I mean, certainly there are tools and things out there that are falling out of favor, just like technology tends to, right? Some tools rise, some tools fall. So I'll steer clear from that. The skills and things that would make someone successful and durably successful throughout their career are really in the how do you work with others, how do you influence change and how do you help people be successful to use data? Because I do think that it's still a big challenge for a lot of people. Data and analytics is not intuitive to a lot of people in the workforce right now. And I think you're seeing that in a number of different ways in society. I mean, you're seeing the way we use or misuse statistics and analysis and et cetera. And I think that if you can really have solid communication skills, if you really have some willingness to, I don't know, maybe it's empathy to put yourself in someone else's shoes and try and help them figure out how data can help their day to day lives at Work and so on. That's a huge skill. I mean, that's super powerful. It has a compounding effect. And so I think if you focus on soft skills like that that are really kind of centered around driving changes, that will put you in good stead in a number of different ways. And you can use those skills at any level in your career. Right. If you're just starting out and you're helping a peer figure out how to use data differently to make a decision about how they sell a widget, or you might be working 20, 30 years later with a CEO trying to figure out how to enter a given marketplace, the same skills in those cases will apply. And so if you build those over time, you're going to be in good stead.

Okay, just imagine you've got a time travel machine. You can go back to the start of your data career. What's the one thing you would tell yourself? What's one piece of advice?

Well, I would tell myself what to buy in terms of the stock market. But if I'm going to be limited to career advice, I'd say I think I really early in my career and frankly into my mid career, I struggled a lot like I had. I was a guy that had a game plan and I tried to make that plan work. And quite honestly, the world and life doesn't work out that way. And I think just now that I'm an old man, I figured out that you can't plan for everything. So what I would tell myself is to probably chill out, try and be open to opportunities, stay open to opportunities to learn and opportunities to work in different areas and not put too much pressure on any of those, any of those kinds of decisions. Like, some things are good, some things are bad, some things work out, some things don't, and enjoy the ride a little bit. I think internally I put some pressure on myself. I'm a driven person and I think it took away from some of the fun. And I think at the end of the day, you have to figure out a way to have fun with what you're doing and that's what's going to keep you going. And that and building some relationships as well. Because if you're having some fun with some other people on a team, it's usually a good day. And so, yeah, that's what I would tell myself. Chill out, to be open to that and recognize it and invest a little bit more in the fun side of things and the relationship side of things. As long as you keep your eyes open and your head on the swivel, open to opportunities. Things will work out.

I love that. Just sort of accept that careers and data are weird. Have fun. Roll with it. That seems like great advice. So on that note, thank you very much for chatting. It's been a real pleasure talking to you.

Awesome. Thanks so much for your time, Regine. I hope this was well received.

You've been listening to Data Framed, a podcast by Data Camp. Keep connected with us by subscribing to the show in your favorite podcast player. Please give us a rating, leave a comment, and share episodes you love. That helps us keep delivering insights into all things data. Thanks for listening and you thank until next time.