Steve DeCarlo built AmWINS into the largest wholesaler in the industry. He is an industry leader in data analytics, having started years before any of his competitors. Steve retired in May and sat down with founding editor Rick Pullen to talk about his career. This is an excerpt on his views on data. —Editor
Where did your love of data get its start?
I started my career as an accountant, so I was always dealing with data. I purchased my first personal computer in the ’80s and used technology as it grew and advanced. We were using Lotus123, not Excel. When we were starting Royal Specialty Underwriting, I was able to buy an IBM AS/400 in the summer of ’88. I hired a programmer because I wanted to take what we were doing in Excel spreadsheets and program that data. So he created a way for us to build a database, and I bought green screens for everybody. We put this system on everybody’s desktop, and all of a sudden, we’re designing software.
Just one programmer doing all this?
One guy—Kelly Walls—the guy who sat next to my office. We would sit there and try to figure out how to do binders and quotes, and then we started thinking about documents and policy issuance and coding and accounting and reinsurance. We designed and built. It helped our firm understand its portfolio.
When did it become important that you could see and use the data you’d collected and then maintain it?
I’ve learned building and maintenance are two different words. Building—whoa! Maintenance—hard. But we have to do maintenance. So we were in the wild period. We had nothing to start with, no legacy systems.
No legacy system is an advantage.
A huge advantage. I’m an accountant brain, but I’m in the insurance game.
How long did this take you?
We worked on it every day from ’88 on. We began attracting capital because our data was so good. They’d say, “I’ll give you reinsurance because I can see my portfolio.” I was a young accountant working with spreadsheets, so for me it was just a tool. But it was analytics, it was data, and it’s one of the things that made our business so good.
How much do you know about actual software?
Not much. I know insurance. It’s just data. To me, it’s linear.
Describe a couple of “aha” moments using your data.
The first time we blew people’s minds was with a limits and attachment point profile. It was so basic, but back then, nobody had it. Nobody had the concepts to be able to do it. Ultimately, you could negotiate better reinsurance terms because you had better transparency in your data. That’s the war that’s going on now. So that was educational. That was foundational.
You never shared this outside the company?
No. We didn’t sell the software to anyone. When dot-coms were coming out in ’97, ’98, people were telling me, “Steve, you need to go do this.” And I told them, “I’m an insurance guy. I know what I’m good at.”
Back in the ’80s and ’90s of course, we were using old languages. So we became a dot-net shop and basically committed to Microsoft. That was an early lesson. We made lots of mistakes. There was no cloud, so you had to buy servers and put them everywhere. You had to have server farms. At AmWINS, when we started partnering with firms, they came with technology that needed to be updated. First you had to control their hardware, then you had to convince them to migrate away from their legacy systems. In the early years that was hard, because we were just getting started and building. Ultimately, we didn’t give them a choice.
We said, “You’ve got to throw away your legacy system.”
Today, if you sell an account in Seattle, it dings in my office [in Charlotte], and people are amazed at that. But if you walk into a 7-Eleven and you buy a bag of potato chips, it tells somebody that you bought that bag of potato chips. And guess what? That night they’re going to deliver another bag of potato chips to that store. It’s called a cash register, and it’s not that big of a deal. But in insurance, the reaction is, “Oh my God, you can see what you are binding?” People think it’s a wow, and it’s not a wow.
So what does that say about our industry?
Lots. But the industry continues to have a lot of legacy systems. I’m now trying to figure out how we’re going to use Microsoft’s Cortana audible digital assistance. This is the way we’re going to work. We’re not going to type on a computer, anymore. Right now, I ask questions. “How much is this?” “How much is that?” “Where do I get this?” And our data team gets me what I want. In the future, I’ll ask for information using these new technologies, and our databases should be able to provide it to me verbally.
I built a “valet” phone app because when our brokers go to lunch with a client and they hand the keys to the valet, they say, “I wish I knew how much volume I did with this client.” That’s the only time they thinks about it—instead of running a report, like we used to.
Now, it’s on their phone. They type the client’s name in and up comes their file. They doesn’t have to run a report. In the future, brokers are going to say to their phone, “I’m having lunch with Joe Smith. Give me Joe Smith’s financial biography.” It’s all going to come out of databases. Nobody is going to print off a report anymore.
So you are in the process of tying your database to the digital assistant right now?
Exactly. Just asking it questions. Just like you ask it what the weather is. “Give me AIG’s premium.” That kind of thing.
The thing about data is that it’s just data. I’m a little frustrated that I won’t see that come to fruition as part of my career, but the young people are working on that. AmWINS’ biggest advantage now is the amount of effort we’ve already put into this easy stuff, which leads to the next level of implementation.
Which is more sales.
More sales, exactly. We’re in the small-business arena now. We’ve invested an amazing amount of money in technology and small business. Everybody talks about small business. It’s easy to talk about it. It’s easy to do it the way you have always done it. You’re just not going to do it in big volumes, and then you’re not going to have the analytics.
How do you make small business profitable?
With better process, data and analytics. Today, we partner with an insurance company that built a small team to implement a small-business strategy. They actually trusted our data. So now, instead of them typing it in and me typing it in, they just analyze me. They trust me. That’s a step forward that most people—that the industry—has got to get to.
Who owns the data?
The insured owns the data. Clearly the wholesale brokers, the retail broker and insurance company all use this data to make decisions. What we don’t do is give carrier data to a competitor. So if I work with insurance Company A on a data structure, I can’t give that to insurance Company B. We receive one million submissions a year. That data is valuable.
What’s the value proposition?
Knowledge. The advantage is analytics—the ability to see how to correlate different subjects. I do business with 20,000 retailers and 500 insurance carriers, so I have a view on a lot of business.
What do you share, and what don’t you share?
I don’t share competitive data. I may say I have knowledge that the Florida marketplace pricing is X because I have a broad perspective, and people are interested in that knowledge. They may say, “Well, show me that knowledge.” I will, as long as I hide the source. You have to be careful with data.
Basically, you’re tech savvy. Your ability to collect data gives you business opportunities other people don’t have.
They have them. They don’t see them. Here’s a great example: an insurer gave us the ability to underwrite on their behalf. They started losing business. They couldn’t figure out why. So they asked, “What are you doing? You’re not giving us business.” We told them their price was no longer competitive. Then, we explained that three new carriers had entered the market and they were pricing their products differently. We told them, “You’ve been around 50 years, and you can’t rest on what you think you know.” So they changed their pricing, and now they compete in the marketplace.
Where are your competitors compared to you?
Well, I try not to know what the other competitors are doing, because I have enough to worry about on my own.
But you still hear. C’mon, you’re out there. You know, a lot of them still use packages, and packages are limited by when the guy can get to it. A lot of them don’t see value in data.
Still. I was at an industry event recently, and they brought in Billy Beane, the Oakland A’s GM who went into analytics about baseball. Michael Lewis wrote a book, Moneyball, and they even made a movie about him. The establishment called him a fool. I sat there thinking, “That’s what they said about me: ‘He’s just an accountant.’” But Billy Beane analyzed the statistics of baseball to look for the hidden gems. I thought to myself, “That’s what we did.”
Everybody in baseball is now following Beane’s example. How about you?
They’re trying to, but they have legacy issues and money issues. It’s expensive. You know, we’ve been doing this a long time. Now, everyone is trying to play catch-up. Catch-up is hard.
Everybody is afraid to share data.
Today, people are in a war to get data. If every insurance company had the exact same data, then it’s the company’s beliefs that change the price, not the data. I think we should share data. I think we should build one database that all insurance companies can share. We’d all have the same information. We’d all pay for it. If we did that, then how we price insurance using the same data would be based on our beliefs, our capital structure and how we buy reinsurance protection.
Do you think that’ll happen someday?
I’ve talked about it enough. AIG believes that they can do small business better than everybody else. Chubb believes they can do small business better than everybody else. I believe both of them. But their confidence is based on their beliefs around risk selection, their capital structure, and their claims-handling ability—actually paying losses.
How difficult is cyber security for a company like yours?
I’ve seen millions of dollars being wired incorrectly because of fraud, so the need for robust cyber-security systems is real. The perception is that all brokers are equal. We’re not. I’m not your mother’s wholesale broker.
You don’t see any drawbacks to sharing data? I mean, as long as you restrict its identity?
Yeah, you have to be careful with that. But, you know, I’m trusted with the insured’s data. They give me their financial information. I have their loss information. I’ve got to be good at what I do. I’ve got to protect it.