Follow JRNY on LinkedInFollow JRNY on MediumFollow JRNY on Twitter

In the second segment of our interview with Glenn Renwick, the Kiwi ex-CEO of Progressive USA, we talk about the evolution of insurance distribution. Check our the video or transcript below.

So getting straight into it … At Progressive, I understand that you started with 100% agent lead sales and then moved the model and brand to direct-to-consumer. What was the catalyst behind that change?

Yes and this all happened pre-internet. So we have to think about 800 numbers and the like. These are these are really important decisions because they had great resistance, but it seemed that in the United States and I must say the very late 70s early 80s there was an evolving evolving 800-number culture for certain products, even Sears Roebuck - not necessarily a storied company today, but had more catalog sales and so on and so forth at the time. Then by sort of the late 80s it was starting to have appeal in other segments of the retail industry, not so much in the financial services industry, but in this particular case it involved me the pronoun is not really not important but I did sort of have a couple of key people. I said, you know, I think this could be a way that we could supplement our distribution and this may be the vector of change that really is going to happen in the marketplace.

We all look awfully bright when it comes out to be the right vector but that doesn't necessarily always happen on every business decision. And in this case it couldn't be more positive for Progressive because it allowed us not only a different distribution channel, but it allowed us to sort of introduce innovations a lot faster because we had one less constituency. Our agents - our agents remain very important and I don't want to sort of suggest that Progressive was trying to in any way to move away from agent distribution - everything that was sort of on the horizon and well you couldn't necessarily see exactly the way the Internet has developed and how users have gravitated to it so quickly.

We certainly could see that that was a potential and there are individuals who want to go see their local agent and get advice and maybe they've got other insurance needs that they want and share with that particular agent. But there was a massive part of the marketplace that just wanted auto insurance. And as much as the industry would love to be able to say well you need professional advice and so on and so forth, and again I am not in any way negative towards agents, we still generate a great deal of our business through our agents and love them; but it wasn't that difficult to create an intelligent interaction with the consumer.

Now our first attempts at this were user interfaces were not nearly what they are today but moving to direct and direct now represents more than half of the premium income of Progressive but the agent part is extremely valuable and certainly for some other products that they're a little bit more suited for. So it's a mixed model.

Yeah however I make it seem like it was okay, but there was great resistance from agents and they felt that we were trying to take their business from them. That really wasn't the case and now I think we've got an extraordinarily healthy model. And if I could just on a little bit...

...because we do the direct to consumer we had the need to develop a consumer brand which may never have happened if we'd been agent distributed so now agents can actually represent a company that for them is the best consumer brand that they can represent. So it actually did come sort of full circle to being positive for our agents as well.

So you've spoken about the agent resistance. Were there any other hurdles when implementing a huge change like that?

Um that is by far the largest, however there are some regulatory issues and they are US regulatory issues so I won't go into them in great detail. First of all, the insurance in the United States is regulated by the states and we have really multiple jurisdictions, 50 jurisdictions, that regulations surprisingly are not the same in individual states. And they weren't written to oppose direct in any way shape or form, but they weren't written with direct to the consumer in mind. 

So the whole notion that insurance policy would have an agent was sort of embedded into the regulatory framework and here we're coming along - as Geico had also done similarly - and selling directly to the consumer.

So that part of the regulatory framework really wasn't there and surprisingly isn't there today but there's no opposition to it and it just sometimes the regulation sort of looked like they were written at a period where this didn't exist and I'm sure to some of your entrepreneurs and people in the insurtech they're facing issues were what they're doing wasn't anticipated by regulations.

But luckily we didn't have anything that was a major hurdle that prevented us doing things but there were oftentimes delays and proofs that we had to produce and things that we had to do a little differently.

And going direct when you started me using the telephone. In what way did the internet change the game?

Ah it was with quantums. Yeah the telephone - least we forget - at that time was still quite an effective device but clearly costly and clearly from a scaling point of view problematic, although we have lots and lots of call centers but we would have had even more had we not had something like the internet.

But more so the internet would have provided the ability to start really getting into much more intelligent user interfaces and we created labs where we would actually watch the retina of people's eyes so that we could actually determine whether what looked really cool graphically was effective for them so we took it to a level of science that I think was extreme, I'm sure others do it, but we all know websites that look good but are frustrating. And then there are others that may or may not look as good but it just works, and it anticipates your flow. We had to become very good at sort of that mindset. And then on top of that, what I'd call intelligent data fill. Now I don't know, I haven't been in market long enough to really know what is available and what's done in NZ so you may be lightyears ahead of where we were or this may not be as relevant.

Look you don't want to spook the consumer out by the moment they type their name you said “oh yeah I know everything about you”. On the other hand, really intelligent well paced data fill so that you can actually “okay I understand really quite well that you have these vehicles” and do it in a way that just doesn't put the consumer “hold up this is getting a little spooky for me”. And then little subtle things that we could sort of laugh off but actually in a great design look good. You put in your vehicle and actually will bring up your vehicle, and in some cases we even got the colour right just because of what was available to us in external databases that we could feed in at the time.

So that type of obsession with getting the interview for the sale right really led us through and we were certainly the market leader in that.

We were the first to ever sell an insurance policy through the internet and a few of us still remember the first lady in Wisconsin I think it was that bought the policy and some of the details about the policy.

Obviously, with privacy we don't give out any more details than that but it's it was one of those fun things where you go “okay we've now got a, back then, a very basic website - will anyone use it?”. And then the first sale, then the second sale, from there it has been quite a quite a wild ride.

How else would you say insurance distribution has changed and where do you see the future of insurance distribution going?

Well I think resting and again moving aside from the agent which is the right choice for some people but on the direct side just really sort of resting control to the consumer is great. This is a little bit of a distribution but more of a product in and of itself - we allow people now to have their insurance premium judged or determined by their usage of the vehicle. So we started off by actually putting a little dongle in the car which would give us telemetry data, actually it was batchit first but then ultimately telemetry and so on and so forth.

Now we can do it all from the phone so we actually can know how you're driving your vehicle which frankly I'd be surprised if people don't sort of say in NZ, with less of a focus on liability, you know why don't you rate me based on who I am versus all the other factors - my age and so on and so forth, where I live.

And this is a great movement away from what I'll call correlation variables to causation so actually we now can say look we're gonna rate you as an individual that's very different than the original notion of insurance which was we'll put you in a large group of people that look similarly from a insurance underwriting perspective to you and your price will be determined by that group dynamic versus your individual dynamic. Technology is allowing us to move towards a true customization or individualization of insurance premiums and then that kind of movement combined with the data collection and data initiation on the internet, really does change everything.

A second one that hasn't really happened yet, it was something I was, let's just say we were planning the seeds for is - should insurance be fixed term? You know today we think very clearly about “I'm buying a six month policy or an annual policy or whatever renewable”. Is that necessary? And with technology well, why can’t it be more of just a continuous variable with relatively flexible start and end dates that are based more on the ease of need where if someone doesn't need their car for three months and it's not being used, well should they pay premium for that - telemetry can certainly help in that respective.

So I think we're moving more to sort of insurance for the one and more based on how that one actually uses their vehicle and whether they're a very fast accelerator or a very fast breaker which we found to be two causal relationships far more important than whether they were going at a high speed or even above the speed limit or something like that.

So data is extraordinarily valuable, extraordinarily rich, you just got a mine up for the right things. 

And with all these changes and distribution and products do you think the place of trust has changed and insurance sales?

Yeah I think that people would certainly have great trust in their agent and I think that model I don't want to date that model because that remains true today, but I think that's why people go to an agent. We are clearly talking about a generational type thing and people are willing to trust websites and perhaps have never had an intermediary in any parts of their life so that trust can be transferred almost to yourself. But you want to make sure that whoever you're dealing with, and this goes for lots and lots of industries, feels trustworthy and I'm not talking about publishing your privacy policy or your data release policy or anything like, that I'm talking about just every single element of it why it feels right as a match for you and that, that's a discipline and a skill that isn't just “Oh get it right one time”; it's it's the small subtle details all the time of that interaction with the consumer. And by doing that, yeah I think that the consumer can start to say “I trust myself” so I'm not sure whether that trust is going on to the insurance company as much as it's going to themselves, but I trust myself in combination with let's say Progressive and I think that's an outcome that with the brand of Progressive and giving a lot of power to the consumer, that trust relationship is now shared versus abdicated.

So still very much important just change slightly?

Yeah in insurance trust is critical, yeah.

And what are your thoughts on automating parts of the insurance customer journey? Are there some parts of the customer journey that are easier to automate maybe than others?

Yeah I'm a huge fan of automation, that was you know a big part of Progressive, I tried to always present Progressive at least to Wall Street as a statistics factory and a technology company - which wasn't really there for the smile value, that was really how we looked at ourselves versus sort of a more classic underwriter. So from a technology point of view, we went to great lengths. You know the most important thing is really how it manifests itself in the product and I talked a little bit about telemetry now and using the car and the driver relationship as a primary rating variable. And I think that'll continue because they know where, they are they know how they're being driven and those things are critically important as we start to develop a price point that can be individualized.

Then you can move on to the claim segment. We took claims - this was a longer story - we took claims extraordinarily seriously in terms of process, innovation and making sure that we redefined our product as a company. First our product, I think this is probably true for most insurers, was cash. Something happens you get cash back. Well we found that wasn't what the consumer wanted. What they wanted was their car back. I'm not talking about injury which is a big part of the insurance premium in the United States but just physical damage. They wanted their car back in the driveway. 

But we often, at least in the United States, exposed them to a supply chain that was very a low-frequency transaction for them, they didn't understand body shops and that up didn't always go well. So we ultimately changed our process to make the consumer ultimately have their car back in the driveway and we took proxy and care of the car during that time. And that was a lot of technology involved in that as well.

Now you have even more tools that will supplement that process and probably build on it with photo inspection. So it's one thing to take a photo of damage. It's another thing to take a photo damage of a car and have it matched with five million other similar quarter panel damages and throughout artificial intelligence actually be able to say this is probably the right price for fixing that particular quarter panel. Now, if you have five you can't do it; if you have five million you can get very close. And with artificial intelligence, combined with photo inspection, and we all know what can be done with some of the Google tools and so on and so forth, today that's a really powerful notion. So your question is really technology, it pervades every pot of the cycle. We talked a lot about the intelligent data fill, we talked a lot about just did claims, I could also talk about how he presented Progressive as the only company I think in the world that actually presents its full financials every month - we can talk about the reason for that later - but again, it was all based on technology.

Here's my “but” - my “but” is that when your customer really wants to talk to you because it's one permutation or combination that you didn't get inside of the primary flows of your software they can get extraordinarily frustrated by not getting the service that they need and occasionally that means talking to a human being. You know the chatbot might not get that one. So it's horses for courses.

My advice would be don't necessarily cut the human to human interaction out - we know it's expensive, we know we want to minimize it - if you really focus on the details, the subtleties and process improvements all the way along your food chain, you can get an awful lot of it out. But don't deprive your customer of the ability to occasionally just have that one, as I said, awkward permutation that didn't fit. Now, if you get enough of those maybe there's a possibility you can also automate that and backwardly filled your product or your service gaps. That would be my only advice, so I think it's it's absolutely an industry that can take and use technology, no questions asked. But it can be very brand damaging if the consumer gets frustrated by not being able to feel like there's a way to get into that system.

So would you say that the future of robo-advice and automation is really about giving the customer the choice about speaking to a human when they want or using automation when they want?

Yeah and it's up to you the provider. If you do it really well you'll change that balance point; if it's not done well people will off for a human way more than you probably want them to. Yeah but one of the things that we always encouraged developers or customer service people think about is lifetime and premium value of the customer. It's very expensive to acquire a customer and I'm coming from my perspective where television advertising, Google advertising, social media advertising etc etc it's very expensive. Keeping a customer can be extraordinarily profitable especially if you can keep them for decades. And where, unlike some of the more state farms of the world they've been around for a long long long time, we don't have too many that are multi-decade customers, but as we sort of continue to advance that timeline we expect that we'll be developing multi-decade customers and their lifetime earned premium or lifetime profitability let's just say that, is extraordinary, and we want to be fair to them, we want to be fair to the company; but you don't want to miss the opportunity to have them say “yeah I'll stick rather than change” just because of that one small interaction. So I'm a big advocate, big two thumbs up for technology but not at the cost of not switching if that's what the customer needs or wants at that time.

Keep your eyes pealed for the next part in this interview series where we speak more with Glenn and about how to innovate in insurance.

Sharing is caring

What's next?

Ready to
start the JRNY?