In this third and final podcast from the InsTech London evening event “MGA, the new frontier for Insurtech” recorded live at The Steelyard London on 2nd April, join Matthew Grant on the sofa as he talks to the founders and partners of three more new insurtech MGA solutions.
First up, Matthew talks to Robert Jeffery founder and COO of Evari about the benefits that come from having both an MGA and providing a service to build technology platforms. (0:48) https://evari.insure/
Next is Tobias Taupitz CEO and founder and of bicycle insurer Laka and Arslan Hannani from their capacity providers Zurich. They talk about the company's new approach to insurance, their success to date and what Zurich has learnt from working with them. (6.56) https://www.laka.co.uk/
Finally, co-founders of Envelop Jonathan Spry and Paul Guthrie tell us about why decided to work together and how they are getting on with their cyber reinsurance MGA, backed by MSAmlin. (18:00) https://www.enveloprisk.com/
More details from the event, including photographs and articles related to the topic of MGAs can be founded at the InsTech London webpage.
Finally, if you are interested in what an MGA is and why so many insurtechs are adopting this model, you may find this recent article by Matthew Grant of interest - MGAs: the fast track to insurance innovation?
Transcript for this podcast
00:01 Matthew Grant: Hello and welcome to the InsTech London podcast. This is Matthew Grant, one of the partners at InsTech London. And in this episode we're bringing you the final set of discussions from our recent MGA event: ‘MGA is the next frontier for InsurTech’, that we held in London on the 2nd of April. You'll hear in this episode from Evari, Laka, Zurich and Envelop. And if you're interested in actually learning a little bit more about MGAs and in particular what they are, you can find an article I wrote under my LinkedIn profile, Matthew Grant, which gives a bit more background to the slightly mysterious world of InsurTech. If you want to find out more about what we did for the event, you can find us at www.instech.london.
00:48 Robert Jeffery: So Evari, what we've done is we've built technology which enables us to offer dynamic insurance policies. And when I say dynamic, there are policies which are flexible and can change as the insured's circumstances fluctuate or their needs change. We are an MGA, we're a Lloyd's coverholder and we distribute insurance products to small businesses directly in Australia, but as you noted, we actually now are working with other carriers and MGAs and using the Evari platform to help power their business too.
01:22 Matthew Grant: Okay, and what brought you over here from Australia to the UK?
01:28 RJ: Not the weather, [chuckle] that's for sure. The market's obviously so much better here in terms of size, and we're also tackling other markets as well. So we're now working with MGAs in the US and people in Europe to offer our technology, and we're looking at what MGA presence makes sense for us from an Evari perspective in the UK too.
01:51 MG: And so where is the main focus? Are you using the MGA to basically understand and prove the business model but do you see the real growth coming from being the platform to offer to other MGAs, or are you still driving hard on the MGA route?
02:06 RJ: Our business in Australia really serves as a test bed for us to be testing technology and really learning about how dynamic policy should work. The user behaviour in terms of how people interact with those policies is quite important if we want to be able to offer products for that. I suppose our core IP really is in the technology, so we have a solution that we can offer out of the box to run your MGA, which we're doing. It's also modular so we're working with carriers. They can get the advantage of the dynamic real-time pricing engine, the data connections, the underwriting rules, so all of those kinds of features that come through that.
02:43 MG: There are quite a few organisations, we've heard from a few tonight who are offering platforms. How do you distinguish yourself then, when you're building this platform for organisations, from the choices they may have?
02:56 RJ: Yes, good question. So there's a few things that our clients tell us and that we've learned. So one is the richness of the features in terms of the ability to do these dynamic policies, which it can move in real-time, based on data and other impulses, to what changes - so real-time policies and convenient customer features. The other thing I think is really our knowledge of the insurance industry and our ability to help influence the product design. They're some of the things that really make a difference. And then thirdly, the speed to market. So at the moment, we're in the order of a few weeks to be able to launch products, we're trying to get that down to days.
03:38 MG: Down to?
03:38 RJ: Down to days.
03:39 MG: Down to days. Okay.
03:39 RJ: So we're not there at the moment. We're in the ‘weeks' range, but later in the year, we expect to be down to days to launch products. And that's important, not just from when you're launching your first product, but actually, you really want to be able to iterate on that product on an ongoing basis, so you need the ability to go in and change rates and portfolio adjustments and things over time. So that's a huge emphasis for our team at the moment.
04:03 MG: So if you can launch products in days, are you finding partners to work with that also have the ability to make decisions fast enough or are you kind of racing ahead of them and now you’re finding that the ‘choke point’ is decision-making in your clients or in your carriers?
04:18 RJ: Yes, so I guess we're trying to raise the bar from a technology perspective in terms of what's available, but what we find is, in practice you can only configure a product once you know what that looks like. So there's a cycle to work through with people in terms of what a product should be designed like, what the user experience will be, and the technology - we're just trying to make it easy. So it's just a tool that you use when you go through that process.
04:43 MG: Okay, so easier, faster, dynamic pricing route to market is kind of the way you'd position yourselves?
04:48 RJ: Yes, we really see that if you raise the bar in terms of the standard features you should get out of the box, it should be able to switch things on and off, or connect it to data, have digital policies that change in line with the insurance, all of those things, and it should be easy to configure, so you can put your effort in terms of designing a good product, which meets the insured's needs.
05:07 MG: Great and actually on the underwriting side of the MGA business, how is that doing in terms of the access to the market and growth of the business?
05:15 RJ: Yes, it’s going well. Our technology is channel agnostic but for our MGA business, we're focusing on direct and the secondary focus for us is on working with partners, so affinity brands who can use our APIs to offer products at point of sale. And that's really our focus on how we get that big volume distribution, using the innovative products that we've built there.
05:42 MG: Okay. And how long have you been offering that for?
05:46 RJ: We're about 18 months into our journey in terms of offering policies.
05:50 MG: Great, okay, Robert, thanks very much. We've got time for a couple of questions.
05:56 Audience question: If you're asked why someone would choose you over someone who is an MGA as well as technology provider, what's your response?
06:09 RJ: I think it comes down to our ability to understand the business. If you're trying to set up an MGA or are an existing MGA looking to transform your business, we've walked that path already, so we have some suggestions, some solutions in terms of technology or processes that we can offer. And I think that sort of quasi-consulting role that we play in partnership with clients can be really powerful. And for us, the platform has to function, it has to be performing, have the right features. We have the ability to do that because we're learning every day through our MGA business as well.
06:45 MG: Yes, great models. ‘You're in dog food’ I think is what they call it in Silicon Valley. Okay Robert, thank you very much.
06:51 RJ: Thank you.
06:55 MG: Okay. Tobi and Arslan, if you'd like to join me up here on the sofas, that'd be great. Tobi, a lot of people know about Laka, I don't know if everyone really still really understands what you do. And I think the real test is someone can go home tonight and say, "Wow, I saw that guy Tobi and he told me about what they're doing, it's great and this is what they did. So just tell me, what do you do?
07:19 Tobias Taupitz: What we do essentially is we have introduced a cost-plus business model in the insurance world, and by that we mean no more premiums upfront. We say it's kind of unfair that you have to pay premiums first for service down the line. We say let's flip it around, we pay out money first and have a happy customer and only the end of the month we send the customer bill telling them, "Please, please, please, pay me back my £3 I spent on all of you." And basically introducing credit risk where there was underwriting risk before. So cost plus business model, the American Express of insurance, so to speak.
07:51 MG: So just talk me through it, so I take out a policy with you for my bicycle, my bicycle gets stolen, I put a claim into Laka, what happens then?
08:02 TT: So, we are proud that we've build our whole tech platform in-house, so we do the on-boarding administration and claims handling all with us. When there's a claim we would look at that and pay out immediately. We have a bit of money sitting around from our partner Zurich, so we have a claims float at hand that we can spend. We have claims authority for that as well, we are structured as an MGA however we are just using it as a vehicle and then we can pay out first and at the end of the month, we get the money back from the customers on behalf, so to speak, of Zurich and we split the fees.
08:33 MG: So you're kind of creating peer pressure for people not to abuse the system. And I think part of your distribution model is through bicycle clubs and cycle clubs, so you kind of create that group as well, is that right?
08:44 TT: So I probably should have said that our first proof of concept has been cover for high-end bikes, really the choosy ones of £1000 onwards. The average policy value is £5500 by now. So those people are spending a lot of money on that, behaving very well, taking good care. It's an emotional asset for them. And with that, we have very good touch points and communication with the customer. So that's the first starting point, we're moving on to other products as soon as later this year.
09:09 MG: Arslan, just tell me what was it about what Tobi is doing that Zurich decided to back his business?
09:17 Arslan Hannani: I think it was trying to approach insurance in a whole different way, so it wasn't necessarily about the tech or customer acquisition, it was about underwriting in a completely different way. So at the end of the day, the result, it's a digital mutual insurance company and what we get to learn is, well, how do we incentivise those people to behave as best as they can when they're in the risk pool? How do we manage the customer journey of doing something like this? So for us, it wasn't about, "This is going to make us billions," or you know what I mean. This is something that I think should be the way most insurance should be operated in the future. So, for us it's about going on that journey with them.
09:54 MG: And what kind of insights have you found? Any surprises in terms of people's behaviour and how they claim under this new model.
10:01 TT: So I think the most interesting thing is that people want to be part of a group, especially in a formal group like cycling for instance. One person sent us a message saying, "When I get my bill from Laka at the end of the month, I feel that I have helped people like myself getting back on the saddle, literally." So there's this breaking down of the anonymity in the risk pool, and holding people accountable. The more claims you have, the more you pay, which is kept, by the way, to a stop-loss also provided by Zurich. So the downside is limited, but we clearly say, if you behave better, you save money. It's up to you. If you claim less, it's not me making more profit, it's you saving money. And I think people like that message a lot.
10:39 MG: Are either of you seeing other examples out there of other people doing this and/or do you have ideas of where you could use this beyond just cyclists?
10:54 Arslan Hannini: Oh, absolutely. We're under considerable pressure from Tobi to expand the capacity and look at different lines of businesses that we're looking at now. I think also for someone like us, this has a lot of use cases internationally, so if we think outside the UK market, there are some really interesting places that Laka could go and prove this model and grow even more.
11:14 MG: But in cycling or into different lines of business?
11:17 AH: I think it's the digital mutual insurance bit, that's the really interesting bit. I think that the bicycle is the first use case. It's not going to be the last one. I think Tobi will say that before I do. But yes, it's one of many to come.
11:31 MG: Okay, and the technology behind this. What do you see is your strategic advantage in the technology that you use to be able to make you successful?
11:42 TT: I think the beauty is that we don't have to integrate with Zurich. So, it all sits with us. We have full control, we even have claims authorities, we can provide an end-to-end service to the customer. We effectively have built our own policy administration system, we have built our own claims handling platform, and we send a monthly risk bordereau to Zurich and tell them what happened in hindsight, so to speak. So I think full control is really important, customers expect that. I think putting a pretty face on the distribution strategy is a good starting point, but we really want to change the root cause of insurance. Why don’t people trust or like insurance? We believe it's about incentivisation. Making money when you act against your customer is not good - I know it's a strong thing to say in this room - but re-aligning interest in the insurance world is something at the core of the proposition.
12:25 MG: And on that distribution, people are quite conservative about how they buy, they don't want to spend too much time on insurance. How do you sell it and how do you explain to people that it's different? Or do you think it’s just insurance and actually they don't really know what goes on, other than the fact they pay slightly less. How do you shift that behaviour in their buying habits?
12:46 TT: The price is clearly important, right? So we need to be at least market rate and ideally better and on average customers have saved roughly 50% compared to market last year, which is a good starting point. But how do we sell this? So clearly we need the attention of the customer, it takes a minute, the first six, nine months were dreadful for us trading because we had to find the right words. It's a bit like peer-to-peer lending in the early days, a new proposition in the existing field.
13:09 TT: How do you overcome the sentiment that it's just another "Me too", right? To tell them we are really different, it takes a second for them to understand, but once they are on board, they're really bought in. And for me Laka customers are the cyclists, and they would eventually buy more than one product from us. So upselling, cross-selling would be really important for us in 2019, in 2020 and forward.
13:30 MG: Fantastic. Okay, so just a couple of final questions for you, then we'll open it up for the audience. So, Arslan, what has been your experience of working with an organisation like this, very different to Zurich? If it's not challenging your organisation, then it's probably not doing some things right, you know? What have you learned from that whole experience?
13:51 AH: I think in one word, it's probably uncomfortable right? Because we get put in an uncomfortable space which is good for the organisation. But in the lead up to this, I was thinking about what's the advice I would give to other insurers that do have these kinds of relationships. We met before Laka was even called Laka and I think we made the decision to back them on day one and the rest of the time was me trying to sell it within Zurich to say “let's do this”.
14:20 AH: Not just me but a few of us, but I think the key thing is as long as we're strategic. There's a lot of times where we cannot do what they want or they cannot provide the KPIs that we need. So I think we need to be strategic in the long term. The alignment needs to come there, not in the next three months or in the next six months. And what are the sales or what are the loss ratios? It's about “Right, Tobi, where are you going in the next five years?” and can we learn and can we help move our organisation towards something like that while they do that, and learn from them and that's the ultimate goal.
14:53 MG: Yes, and I think what you said before is key, you need to get revenue out of it at some point. But for you it's learning about how do you work with a different distribution model, a different way of paying claims. What's the advice for you, for companies out there that have got a different and new model and have to work with a Zurich or somebody else? If you could sum it up in one thing that you did that you felt made you successful, what would that be?
15:19 TT: I think that's just one word, persistency. I think I had 28 people saying, "Fantastic idea, I'll be your number two but let someone else figure out the paperwork." Arslan was a number one basically, and he wanted to see it through. So I'm not saying you were the 29th person on my list but it took a while, right? A lot of people gave me a yes, maybe and other lukewarm responses. It's just about hitting the right person at the right level at the right organisation and just staying in there.
15:49 MG: Good, and Arslan's comment about he knew you when you had your previous name. Was that his polite way of saying he convinced you to change it to a different name?
15:56 AH: Not at all.
15:57 TT: That is our secret.
16:00 AH: That is our secret, fair enough.
16:00 MG: Right, okay, well we've got time for some questions from the audience.
16:04 Audience question: Hey Tobi, my question is for you. So just going back to when you were starting out and going out to look for some capacity, just looking back, were there any key learnings? Things that you wish you had done better or things you wish someone had told you before?
16:25 TT: Probably, I should have tried to speak the language of the traditional insurance world better. I think that could have helped a little bit. Getting to the point of calling the instruments what we actually needed was a dreadful experience. Claims float, and stop loss arrangement, I think I just called them different things. And I guess having a translator, an intermediary almost - a couple of helpful hands are super helpful for that.
16:53 Audience question: Hi guys, quick question around use cases. Obviously you're looking at things like cycling and other use cases internationally and globally, but are you looking to go into the SME space at all and look at commercial insurances for this type of product or scope?
17:10 TT: Possibly, yes, but I think we deliberately chose a very small use case to convince the regulator, the insurance partners, the customers. The risk is, if we have five bikes or more stolen it might be 20K but the risk is significantly higher with SME insurance. So I guess we've built a lot of goodwill over the last 18 months or so. So there might be use cases beyond that and even bigger tickets. Time will tell, we need to find the right distribution angle. We really like cycling because we can speed and meet the customer in their natural habitat and acquisition costs are much lower than trying to sell dental insurance on Google. So we'll see where it leads us. We have a couple of ideas and I think in Q2, Q3 we're going to do something together again.
17:51 MG: Excellent. Okay, Tobi, Arslan, thank you very much for joining us, I know you've got a lot going on. Appreciate you being here.
18:00 MG: Okay, last up before the shout-outs are Jonathan and Paul from Envelop. Let's talk a bit about what the business is anyway. So talk to us a bit about what you're offering now, with...
18:11 Jonathan Spry: At the moment we're underwriting cyber re-insurance. And I think we've become one of the largest players in that market. Partnering with MSAmlin. We launched that just at the end of the last year, and already I think we're approaching something like 14 million in premiums. We've exceeded our expectations. We are backed by AI and an augmented intelligence approach, so we are a little bit differentiated from some of our competitors, and Paul here has been closely involved in developing that. And at the moment we're just concentrating on cyber and we will begin to diffuse some of our re-insurance capability into primary insurance markets as well.
18:54 MG: Okay, just to make sure I understand, you're a re-insurance MGA, you're backed by MSAmlin. Essentially you're an alternative route to the market for MSAmlin as opposed to their cyber team providing re-insurance into insurers directly, is that right?
19:09 Paul Guthrie: Yes, I guess when we started the company, we started with a very, very simple premise, which is that the first thing that you need to do before you start insuring or re-insuring anything is take the time to make sure that you really understand the risk. And we were really attracted to cyber. I come from an aerospace and defence background, I was doing a lot of work with US military intelligence NASA, and we saw the opportunity, that there were a number of emerging risks out there. Descartes (Underwriting) was speaking to one in weather. And I love the way that he put it where you have to model it like a physical system, you can't just build a statistical model.
19:44 PG: And that was the background of our team around using machine learning and using complex data sets and different types of simulation modelling to get your head around what is this system and how can we start to predict it? And so, the reason that we didn't like saying that we were an MGA upfront or that we were intending to be an MGA is that that was a little bit more incidental to what our core strategy was. And our core strategy was to understand the risk and then find the best way to use that understanding to enable the market. And when you do that, you can find business models that really grow and that really scale. We spent a couple of years building that model, we initially went to market in partnership with MSAmlin underwriting re-insurance.
20:26 MG: So you're underwriting insurance companies? Those are your clients or you're going beyond that?
20:35 JS: For the moment just the reinsurance vertical. And our clients are all insurance companies, which is why you don't see our brand blazing on the side of buses or anything like that.
20:43 MG: Are you finding you can get enough information from your insurance clients, to be able to assess the risk in a credible way, given the site is hard enough to understand and price if you're actually dealing directly with the end enterprise.
20:57 PG: Yes for sure. So we do get data from our clients that are insurers, but that is not a huge component of the data that we're analysing. So one of the paradoxes about cyber risk is that you hear all the time, "How do you analyse it because there's not enough data to analyse it and that's true given a certain assumption about what type of data you would be looking for if you were going to build a statistical or an actuarial model. And the reason that it's a paradox is that I can't think of any other area in the world where there's actually more data available if you want to study it and try to understand the dynamics because it's inherently digital, and so you can get behavioural information on any number of pieces of the ecosystem, and you can begin to understand how those speak together. Certainly, a lot of that data is difficult to get. Certainly, some of it's expensive, there are certainly some holes that you know that you can't get, but that's the core of the system. The data that we get from insurers, we assume that we get no information about those companies in themselves. We're able to collect and analyse a lot of that data. Autonomously, we do it all the time and we're able to, as we get a bordereau and just effectively re-underwrite every single company in that bordereau and then stack them together and see what the correlations look like.
22:16 MG: Great, and Paul, your background and some advice for people with a technology orientation who want to get into insurance or a new business area.
22:25 PG: It's much more important to get the market fit right than it is to get the technology right. And so, back when I was doing aerospace and defence stuff, it seemed like most of my job was just listening to engineers and listening to their absolutely genius idea that no one would ever give them the least amount of money to build, and that no one would ever buy if it existed. And so, if you are a technology person, you have to overcome your ability to look really deeply and see what's possible, and you need to consider what the user needs are. It's kind of a nice sign of the maturity of this industry that almost everyone up here has said that almost exact thing, that it's about product market fit, it's about listening to the users. But yes and from the technology person perspective, the other thing is that what users want is often just incredibly disappointing. You may want to build this incredible system that does something so complex that you just figured out how to do and you think maybe you're the first person in the world that's ever figured out how to do it. But what they want is a button that turns their screen from blue to a different colour blue sometimes, right?
23:29 MG: But the customer is always right.
23:30 JS: We could do that.
23:31 PG: Yes.
23:32 MG: Good, so questions for Paul and Jonathan.
23:35 Audience question: How do you stay abreast, in cyber terms, of the threat environment and how does that impact how you think about re-insurance?
23:45 PG: There's certain data that we use that is incredibly valuable, but it's not necessarily time-sensitive in terms of when we get it. And there's other data that is very... Has a very low temporal resolution but we can't necessarily interpret it holistically, so we have different modules in our system. We spend a lot of time figuring out how to wait the historical data that we have appropriately, does... Do you completely throw out five years ago or do you wait at 10%? Do you look back six months, or do you look back a year? Obviously, the answer is you waited somehow. And then we have different predictive methods that we follow to take more cutting-edge data that we get from data feeds and that we get from experts to tune pieces of our model to say, "Oh, this particular threat factor is increasing," or, "This vulnerability is going to have this effect," and that then propagates throughout the model.
24:46 MG: Great, okay, thanks Paul and Jonathan, thank you very much.