We Develop Products in Reverse Order — Eric Sondergeld on Insurance's Backwards Playbook

Paul Tyler (00:01) Hi, this is Paul Tyler, and welcome to another episode of the L&A Hub Podcast. Today's topic is customer experience, research, and building the insurance industry for the future. With me is a great guest and longtime friend, Eric Sondergeld, Managing Director at Greenwald Research. Eric — I got that all correct, right? Company, title?

Eric Sondergeld (00:34) So far, although the goal of today's call is quite lofty, so I'll do my best.

Paul Tyler (00:42) Well, first off, tell people who you are and what you do at Greenwald, and how you got there over the many years of great work you've done in the industry.

Eric Sondergeld (00:55) Sure. I started my career working at a life company as an actuarial student, and eventually spent many years in market research at LIMRA. About four years ago I joined Greenwald, where I conduct primarily custom market research for clients across the industry — life, annuities, distribution, all kinds of topics. It's been a really nice role because it takes advantage of my entire background: insurance, investments, annuities and retirement, distribution.

Paul Tyler (01:34) You did some really good work at LIMRA over the years — I enjoyed the studies. The last time we were talking publicly like this was actually at one of our incubator conferences up in Hartford. We had you in talking about claims and the claim experience. That was — you told me — about seven years ago?

Eric Sondergeld (02:02) Yeah, it was after my mom had passed, in 2019.

Paul Tyler (02:05) Maybe recap that for people, and then let's talk about customer experience. There's usually no better way to do it than to talk about your own. It's not a great topic to drill into, but I think it'll set the stage for a lot of what we discuss over the next thirty or so minutes.

Eric Sondergeld (02:27) Sure. When my mom passed, there was a life insurance policy with a carrier I'm not going to name. There are five of us kids, and I offered to get the process started because I work in the industry and figured it would be easier.

At that session you hosted back in 2019, the topic was claims, and I gave a brief fifteen-minute talk about my experience. At a high level: when I called the company, I got transferred a lot — they apparently had multiple individual life insurance claim departments, and I'd get bounced back and forth. It didn't make any sense. And not once during the process were condolences expressed. You'd think the life insurance company would be the first one to do that.

It was a difficult, time-consuming process. The other thing I'll mention — they sent out a claim package and application, with a whole checklist of documentation I had to send in, including the death certificate. They gave a little six-by-nine envelope, but I had this stack of materials. There was no way to fit everything they asked for into the envelope they provided. So I had to go to the post office, buy a mailer, and send them the package. They did ultimately pay the claim — it took a couple of months, but it was paid.

Paul Tyler (04:16) A couple of months to pay the claim. Last week I had the CTO from Security Benefit on, and we were talking about experiences. His thesis is that the carrier of the future is the one who actually delivers experiences. It doesn't take AI to give you a big enough envelope, or to say "I'm sorry" on a phone. And that was 2019, pre-AI. I remember the room was kind of stunned. You didn't name the company, but it's a big company, been around, good pedigree — and that was the experience they left you with, as a veteran of the industry. Not exactly the review you'd want. So fast-forward to this year — what happened, and how did the experience change?

Eric Sondergeld (05:14) This year my dad passed, just about a month ago. There were actually three life insurance policies. I'll talk about one in particular — the largest of the three, same carrier. I notified them, and they said within two or three days they'd send a packet of materials to each of the five beneficiaries, us five kids. It took about two and a half weeks, but it did arrive late last week. And they solved the envelope problem.

Paul Tyler (06:05) That's good.

Eric Sondergeld (06:06) Well — they solved it by not providing one at all. I assume the thinking was, "We don't know how much they'll be sending, let them figure it out." It was actually a pretty thin packet. A cover letter, and a one-page, two-sided claim form — really cryptic. One field bled into another; it was very easy to miss a field. And no envelope to mail it back. There was an address at the bottom of the form, though they didn't tell you to mail it there; that was the presumption.

Now, to the good: this company, when I called them, did express condolences. I was happy to see they'd progressed, because that's important. I've had to call lots of organizations managing my father's affairs these past several years, and everyone — financial or otherwise — has expressed condolences, except that one life carrier after my mom passed. So the sentiment improved, but the process really hasn't with this particular carrier.

Part of it is that the claims process, for a lot of companies, is simply a function. There are some companies that treat it as why they exist — they celebrate how much they pay in claims every year. But for others it's just another process, another transaction. And I think we're missing an opportunity to do better with the public, because it's the one time we interact with consumers where it matters most.

Paul Tyler (08:18) I couldn't agree more, and I know our organization is very committed to good policyholder experiences. This doesn't require AI, Eric. It just requires focus on who your customer is and which moments matter most. That's a squandered opportunity — for goodwill, for future referrals, for future business. I'd be reluctant to buy a second policy from a company that did that. Most people would. So we look back and say: some change, some improvement. They said "we're sorry" — we'll give them that. The paperwork still existed. But the technology's changing so quickly now. If we focus on how carriers execute and where the opportunities are — you were kind enough to pull me into one of your forums in Chicago a month or two ago, where you led a really good discussion with leading carriers on how to go to market. There have been a lot of product launches — I've been part of them — where you think you've got a winner, and then you look back and you've sold thirty policies after two months. Why does that happen? We think we're a precise business. We've had smart actuaries like you and good marketing people like me. But it still fails. Why?

Eric Sondergeld (10:13) That happens not just with new products — it can happen with any new initiative or strategy. You have something new, you're excited, you put it out in the world, and it doesn't do what you were hoping. There are lots of reasons. Let me give two examples.

One speaks to our industry's superpower — and we do have one, in my opinion. Our superpower is the ability to copy and paste our competitors' products that are performing well, with a minor tweak, so you can say ours is different or better and therefore you should sell ours. That often works, and companies have been rewarded for it, so they keep doing it. But it doesn't always work. This relates to the whole concept of best practices — in my opinion there is no such thing as a best practice, because a best practice for you might be different from what works best for me. When a company comes out with a new product and it performs well, there may be other factors at play: the distribution strategy, how they present the product, how they support it, the materials, how they implement the new-business process. Maybe professionals try the product and getting it through new business is a bear, so they don't go back.

Where companies are more likely to see results fall short of expectations is when they're actually trying to do something new and different — and they don't create any evidence to support the success of that new thing. They go to market thinking that because they have smart actuaries and smart people who've rallied around a great idea, it'll be successful. Then they scratch their heads. The main reason, in that case, is they just haven't built enough evidence that the thing is going to fly.

Paul Tyler (13:05) You've said our industry develops products in reverse order. What do you mean by that?

Eric Sondergeld (13:15) If you think about classic marketing — which our industry doesn't necessarily ascribe to, at least it didn't used to — from a product development perspective, the classic texts tell you to start by understanding the end consumer. You uncover the needs, desires, and preferences of a particular client segment. If you feel you can build something of mutual value that solves that need and also rewards you, that's how you build your product. You start with the customer, build your solution, and then develop the pathways to connect the solution to the people who'll consume it. That process is called distribution.

What our industry does is start with distribution, because our distribution pathway is already in place, so we assume that's how we'll distribute. It's not a bad assumption. But we start there, then ask, "What do we think they're going to sell?" — often based on what they're selling today and how we can tweak an offering. And then we stop. We never get to the consumer piece. So we do it backwards, and only partially.

Paul Tyler (15:10) It's different. People on the technology side, or coming from outside the business, tell me: "Paul, all these products are complicated. We're going to move to simpler products, and the differentiation will be on the service experience side," kind of like what you've described. That's different from how the industry has ever operated. You said it yourself — "I've got the old policy or the new policy; let me explain the new one, it's like yours but better." Does that maxim change as we move into this technology-first arena?

Eric Sondergeld (16:07) A couple of thoughts. There was a period — mostly over now — when insurance companies were hiring marketing expertise from consumer packaged goods. Great idea, but those people came in, realized how hard it would be to actually affect change, and didn't last.

On service and experience — I think that does exist today, though not necessarily by design. As a researcher, we do a ton of interviews with financial professionals, and countless projects have said the same thing over and over. Product is really important, in terms of putting the right solution in front of a client. But if there are two solutions that are otherwise pretty much identical, and one carrier is easier to do business with — guess who's going to get that piece of business? The experience makes a difference. I've heard advisors say, "I will never go back to XYZ company," or "I'm going to give them another year or two to figure things out before I place business there again." It's usually not a death knell, but people will stop submitting business to a carrier — unless the product is head and shoulders above everything else and it's the only and best product for that client. Then they'll suffer through it because they want to do right for the client. So service plays into it. There are companies with a big focus on experience and getting business through correctly. But it's not all companies.

Paul Tyler (18:31) Let's talk about marketing and product development, since you brought it up. I remember bringing people in from consumer packaging businesses. I got a call once — someone said I should talk to an HR person about a job opening. They wanted somebody who knew nothing about life insurance to do the marketing. I said, "Good luck with that." And of course that whole group lasted maybe two years beyond that. So tell me — what's the future?

Eric Sondergeld (19:15) It's unfortunate, because we need new voices and new perspectives. New thinking is always good.

Paul Tyler (19:31) Yes. People tend to follow the herd, and it may not be the best decision. So talk to me — what does future marketing look like in this business?

Eric Sondergeld (19:51) Do you mean what will it look like, or what do I think it should look like?

Paul Tyler (19:55) What it should look like — I think that's more interesting.

Eric Sondergeld (20:02) I'll go back to classic marketing. There are functions within marketing that aren't functions within marketing in this industry. In fact, in this industry it's often been said — I haven't heard it as much lately, but in the past — that marketing was the sales support department. They'd build brochures and materials to support product sales. They weren't stepping back and thinking about marketing from a broader perspective. That broader perspective includes not only understanding consumers, but building and pricing products, coming up with distribution strategy, distribution itself, and market research. In our industry, product marketing and distribution are separate departments. They talk to each other, obviously. But if those functions were more closely tied together organizationally — under common leadership, really under the umbrella of marketing — I think we could get to better results. Marketing people will be happy to hear me say that; the actuaries and distribution people maybe less so.

I have a theory for how this came to be. Insurance products involve risk, so we need specialists to understand those risks and price the products. That's different from saying, "We're going to manufacture a computer mouse that costs ten dollars to produce, so we'll charge twenty and make a profit." It's not that easy with insurance. We build it today, sell it over time, and the risks materialize over a very long period. So we need actuaries to price those, and we compartmentalized them into a department called actuarial. On the distribution side, in the early days a very high percentage of companies owned their own channels — direct-to-consumer, or more likely a career agency — so they needed a whole set of people to build and manage that system, and it got walled off. Marketing was just there to support those people. So the history of how we got here has perpetuated why we're still organized the way we are.

Paul Tyler (23:26) I personally think it's going to change with technology. Marketing will be more analytic, more data-driven — it'll have to be. And I think there's an opportunity to have a much broader impact across the industry than ever before. Instead of being sales support, marketing gets a seat at the table for strategic conversations, if done the way it should be done. The next question is — if you're marketing, you're marketing to whom? There are a lot of good reasons companies say the agent is really the customer, because the agent brings the clients and is the most influential audience. How should we think about it?

Eric Sondergeld (24:32) That's an age-old question — maybe I should write about it. I have a blog series called Life's Persistent Questions, and this is one companies ask themselves over and over, and the answer changes over time. Companies flip between saying the end consumer is the customer, then switching to say the financial professional is. Obviously they're both customers. You could consider the financial professional a partner rather than a customer, but now we're mincing words. They're both important. In many cases there's far more interaction between the carrier and the person selling the product than with the ultimate client — for life insurance and annuities, there aren't that many touch points with the end customer.

And there's a third customer we completely don't think about, for the most part — the beneficiary. I've written about this: one idea companies might consider is designing with the ultimate user of the product in mind, which is the beneficiary, and designing around that. They're not part of the sale, even though the buyer is thinking about what they believe the beneficiary's needs are. But the beneficiary isn't brought into the conversation. They're essentially given a lump sum when that person passes. And if it was a premature death and the benefit is there as intended, is the money going to be used in the best way possible — maybe the way the person who bought the policy wanted?

Paul Tyler (26:48) True. When I was at MetLife we had an enormous program called Delivering the Promise. When a death claim check went out, the agent had the opportunity to bring it literally to the person. It made an impact when they said "I'm sorry" — having somebody show up at your house with the check in a time of need was incredibly powerful. Agents fought for those opportunities, because very few people experience the benefits of life insurance, and when they do, they're a customer for life. Sometimes what's new is old — we just need to dust it off and create the same experience in a modern way. Companies like Bestow have made a whole business out of doing this.

Eric Sondergeld (27:45) Can I react to that real quick? The very first life insurance policy I bought was an Allstate policy — a combination UL and term. I bought it from an Allstate agent who had space inside a Sears store — back when Allstate and Coldwell Banker had space within Sears, if you're as old as I am. I'd bought my first auto policy from him when I bought my first car, and a couple of years later my first life policy. He didn't do much life insurance — I was probably one of his first sales.

A couple of years after that, he told me he'd had to deliver his first life insurance claim check. It was a life-changing event for him — he saw the impact it had. He went on to become a life leader for that organization. He became a life insurance zealot because of the power of the benefit, and that visceral experience of delivering it. I'm getting goosebumps just thinking about someone going and doing that. Probably sounds nerdy to someone outside this industry, but it's very powerful.

Paul Tyler (29:11) It really is. I think more companies need to think harder about that — to look at it as a key moment in people's lives, not just an expense that gets paid. Talk to me a little about your product marketing lab. But before you do — let me test something I'm seeing: the rapid acceleration of product launches. RILA innovation is exploding. MYGA launches have exploded. IUL seems to have exploded. Some of that's the macro environment — rising interest rates, which can be good for these products. What's the bigger challenge today: creating a product that actually sells, or creating multiple products quickly that can sell? Or both?

Eric Sondergeld (30:27) I'm not sure I know the answer, but it's probably hard to develop multiple products quickly, because we have regulation and a state filing process — you're still going to have a period of time to get things through. Granted, with technology, developing and implementing products may get easier and quicker. But if companies just start developing lots and lots of products, I don't think that's healthy in the long run. I don't think distribution firms or advisors want an especially crowded market just because we can develop more products. There are a lot of products already out there. Companies are better off developing fewer, better products that can sell and get attention.

Paul Tyler (31:28) What's the role of concept testing? You mentioned you have a panel of advisors you pull from. Early in my career I remember doing conjoint work with Blue Cross out of Illinois on group health plans. But the product development I've seen at other carriers is more like: what did this big distributor ask us for, and we'll deliver.

Eric Sondergeld (32:01) We do a lot of product concept testing. We use conjoint in some cases — typically early on, when you're looking for the right mix of things. Most of our product concept testing involves advisor-sold products with some complexity and moving parts: FIAs, RILAs, VUL, IUL, VAs, structured annuities. Probably not so much MYGAs — MYGAs are pretty straightforward, kind of like term. They're spreadsheeted based on rate and rating, and there's not a lot of innovation, because you don't want a product so different they can't spreadsheet it.

But when you have something new and different you want to bring to market — going back to my earlier comment, if you're simply copy-pasting something that's been selling well with a tiny tweak, you don't need to test that; you probably have a good chance until someone one-ups you in the game of leapfrog companies play. But if you have new features or designs you want to test, you have little to go on to know whether it'll hit the market with a thud or take off. That's where product concept testing can be really valuable. And yes, we typically focus primarily — often exclusively — on the seller, not the end consumer, because end consumers don't really understand annuities or life insurance, particularly the more complicated products. But advisors know consumer needs and what they can sell. We conduct interviews with advisors from our proprietary panel — usually it's more confirmation than "carrier, you're way off base." Sometimes they have three designs and are trying to decide which direction to go. That's a great opportunity to test with advisors and see which would do best in the marketplace.

Paul Tyler (34:55) AI has changed a lot. How does it change the world of research? How does it open doors for the product research you just described?

Eric Sondergeld (35:12) That question is very much in flux — probably everywhere, not just in market research. Just like everywhere else, AI can help streamline certain functions. But in terms of actually conducting research, there are a couple of ways vendors are looking to bring AI in. One is synthetic respondents — using AI to create a pool of theoretical consumers or professionals that can answer surveys as if you were surveying real people. The other is AI moderation, where a computer asks a real human questions live. One benefit there is scheduling — if you're in a rush to get many interviews done, it becomes much easier.

I have opinions on both fronts, and there are places for both. But we have to be careful in thinking AI can do everything. It probably can do just about everything — but is it the right tool in every case? Just because it can do something doesn't mean you should. An example: we do a lot of interviewing with financial professionals. My long-term industry experience has made me, I think, a very effective interviewer — I understand their practices, the products, consumers, and research, so I know how to have an effective interview and go down certain pathways where it makes sense to probe. You can program AI to do some of that, but are you going to elicit the best responses? And with synthetic respondents, do we really know those "people" will give the breadth of answers real humans would? I don't have confidence in that. Part of me wonders why we even ask the question — we're in a business that provides services to humans, by humans. I think we need to make sure we're not circumventing their perspectives because we found a shortcut. But we'll see. It's still early days.

Paul Tyler (38:30) I actually tested a bunch of these AI models and their ability to sell insurance and rate themselves as sellers. I've seen some promising studies using AI to model buyer preferences for commonly purchased, smaller-ticket items. For insurance, the data wasn't there. I think we'll only know over time how well we can create synthetic customers or synthetic agents.

Eric Sondergeld (39:10) It's early. Let me ask you two questions, Paul — hopefully I remember the second. First: have you ever called an 800 number with an AI agent trying to get you to the right place, and it can't? It's pretty common. It can do certain things, but for others — even relatively common ones — you actually need to speak with someone. And the second: you personally do a lot of AI work. How often do you have to redirect AI because it went in a direction that wasn't correct?

Paul Tyler (40:02) It's a lot less than it was. Here's my benchmark, Eric: do I have to steer it more times than I steer people? I'm not sure. I see the good, I see the bad, I see the AI slop that comes out. The tool is only as good as the person wielding it, a lot of the time.

Eric Sondergeld (40:29) And granted, my experience with generalized AI is that it's generalized — I'm not talking about agents trained and built for a specific purpose. If I ask it about an industry concept related to a research project, it just gets it wrong. It wasn't built to answer that particular question.

Paul Tyler (40:59) It wasn't built for it. I look at a lot of these models and think it's like taking a slice out of Swiss cheese — some parts are solid, and some are just air. It doesn't know, so it guesses. I think this will change over time. It should let us narrow the options — say, the three we test with live people. Hopefully we get to three from ninety using AI, and then bring in real people to look at them. Hopefully we end up with better products, better experiences, better call-center experiences, better death-claim experiences. I know we're way over time, but my last question: if you could change one thing about how our industry develops products tomorrow morning, what would it be?

Eric Sondergeld (42:24) My main answer goes back to this: I'm sure everyone listening can think of multiple examples where they went to market with something they were excited about, with high expectations that weren't realized. They can avoid that most of the time by gathering evidence — to support, to refute, or to suggest a slightly different path. Market research is one way to do that. If more companies did it, I think we'd be in a better place.

And you mentioned the end consumer — I'd love to see more companies think about the end consumer more than they do today. The vast majority of life and annuity sales are advisor-sold, and the number of consumers who actually have an advisor is pretty small — around thirty-plus percent. So about seventy percent of the market doesn't have access to advice. If they're going to get coverage, they have to find a way to do that directly or through some other avenue, and it's harder to make that happen. There's a large and growing life insurance coverage gap, and a big part of it is people who don't have access to advice in the traditional way. If more companies thought about that opportunity — because it's huge — it would require a different mindset, some work, some investment to learn about consumers, and going back to what I said about marketing: understanding those clients so we can build solutions that work for them. It's very likely going to require thinking differently about distribution. None of this is easy. But there's a big opportunity out there worth tapping.

Paul Tyler (44:48) Eric, thanks so much for having me on — and for having me out a couple of months ago to talk at your forum. What's the best way for people to reach out to you and learn more about Greenwald Research and what you could do for them?

Eric Sondergeld (45:07) Two easy ways. One, look me up on LinkedIn and send me a note. Or go to our website and complete the form — "Hey, I've got a research question, can we talk?" — or send an email. It's my full name, Eric Sondergeld, at greenwaldresearch.com.

Paul Tyler (45:33) We'll put it in the notes. Eric, thanks so much for your time, and I want to thank our listeners for taking time out of their day. Be sure to give us feedback and comments, and most importantly, join us next week for another great episode of the L&A Hub Podcast. Thanks, Eric.

Eric Sondergeld (45:53) Thank you, Paul. This was fun.