Kambi Capital Markets Day - Sportsbook 2.0

Thank you. My name is Andreas and I'm the CTO of Kambi. So I'm going to talk about the two first challenges, and in order to do that I'll need to spend some time telling you the story of how we came to this capability, like the slide Oliver showed previously. We started with pretty much a grey market. There was dot-com operations pretty much all across the world, in particular within EU boundaries anyway. And regulated regulation started to happen. Italy was first. Back then we were still part of Unibet. We didn't do anything in regards to Italy; we launched but that was entirely outsourced. And then France happened, and there was quite a significant impact on revenues for Unibet at that time because the business conditions were quite detrimental versus what they had been doing under the dot-com regime. So we realized that this is something that we really have to be able to deliver on, and we started to work on this. That made us able to launch a regulated solution for Denmark back in 2010, and we actually managed to switch Italy from an outsourced solution to in-house as well. Back then we were maybe five development teams, and this took a significant part of our development capabilities just to be able to deliver a new regulation technically, process-wise, installing servers, setting up APIs, regulators, etc. But we started to hone these skills and we were able to launch in Spain pretty much as that market opened up. It was also the first time where we really saw that going into a regulated market is an excellent business lever, because our first customers we signed from Paf were actually Spanish operators. I think the first three operators were actually Spanish. Then fast-forwarding a bit, I might say also that regulation from the Northern Territory in Australia was regulated way before Kambi started, so with the UK, but during the years from 2013 to 2017 we launched in quite a few different continents and countries across the globe. And South Africa was probably where we ended up in a situation that made us able to launch in the US, because they have a very difficult regulation in the sense that you have to strike the bets in South Africa. A lot of our business model centers around not being forced to do local deployments. We want to maintain the core and not have a separate install base just for trading in South Africa; they would never make sense from a business point of view. Our entire business is high volume and low margins. We have to really take care of the underlying fundamental capability of our company. But we had honed our skills and we were actually able to create a component specifically for South Africa that made sure that the bets were stricken in South Africa. And that capability was like the perfect solution for New Jersey, Pennsylvania, and onwards. We started development early 2017. In the end of 2017, maybe Q3 Q4, we realized that the Supreme Court hearing in December 2017 was probably going to happen. We always thought that this would someday happen, but it became more or less a very high probability that this will happen. So then we started to talk to regulators, we showed the solution we had in order to scale our business into new states, having the capability of actually striking the bets in state. We had the additional complexity with the Wire Act in the US as well. So we were honing our skills a little bit more, but this has really put us into pole position for launching anywhere. And we did the US pretty much at the same time as we did Sweden and a few other regulations. So the scalability of our business now in terms of our ability to launch into new regulation is almost unlimited actually. I mean, we need to deploy a small level of infrastructure in any new state that has this requirement, but it's kind of a copy-paste operation. We can do this at speed. Obviously, there might be regulations that have more stringent or quite different regulatory requirements, and then we have to do some development. But for the most part, and so far in the US, we've been quite successful with this capability. So we're looking forward to the four tiers of states being regulated in the next two years.

The second challenge was retail. We've been running retail for quite some time before entering into the US market, and our core strategy has always been to build on our core capability in terms of sports betting and the online experience. So we're trying to bring whatever we do online into the retail space as well with as little adjustments as possible. In order to do that, we had to be firm about where our service starts and ends. So we're not doing any hardware, but we're really good at integrating into the hardware that you need in a retail operation. We do APIs for those who already have a massive installed base of over-the-counter or for retail outlets like ATG in Sweden, which is the horse racing monopoly. So that made us able to actually launch six on-property sportsbooks in the time span of just a few months in the US: New Jersey first, and then I think five or six installations in Pennsylvania over the course of two months. And they're all different, but they're yet the same underlying core product. And at the same time, we launched with ATG 2,000 stores, 6,000 interactive terminals. I'll show a short video of that later. But my main message is that retail is not something that's just bolted on to an online service. We're doing retail really well actually. So this is a movie from a launch at Parx Casino outside Philadelphia. One of the things that many US operators and casinos were quite skeptical about was our take on retail. We have self-service kiosks, which was quite a foreign concept for Americans. They were very teller-based: you go through over the counter, you place your bet, you get a paper slip, etc. We were emphasizing the ability to scale the sportsbook business with American operators. You can imagine a place like this being absolutely packed when any Philadelphia team is playing. It's going to be hundreds and hundreds of people more than willing to place a bet on anything during that game. It can be quite crowded, and without having the self-service terminals, there would be massive queues. I don't know if you've been, so that was a demo. I'm not going to show that. And then for ATG, which is the horse racing monopoly in Sweden, we launched two minutes past midnight for all their online and retail service. So they're using this concept of interactive wallpapers where you can browse their horse racing, which is very much their own internal, but at the same time you can use the Kambi sports betting which is powered by Kambi, built on our APIs. So it contains all the elements of our online service but it's brought to the retail channel. In Sweden, you're not allowed to put cash into machines; you have to go to the teller. So you go through the teller, you pay your bet, but you have a QR code on the bet slip so it's automatically scanned and being placed. And we're also trying to bridge the gap between retail and online. Obviously, the US being a market where retail is kind of going first in most states, it's important for us and for our operators to kind of familiarize the audience with the notion of online betting. And again, it's quite a lot of people at the sportsbook at large events. So we've built this solution where you can actually find your selection, you can create the bet slip, and that creates a QR code that you can go to the over-the-counter and just show that. It is scanned and you pay. So there's no selection process taking place at the till; you're speeding up the process of actually placing the bet quite a bit. You can also use this capability obviously for on-prem mobile betting. In the US, mobile betting usually means that you're at the casino, within the casino compound. It's not online in the sense that we think of mobile betting. You can use a solution like this as well. You can have a hostess walk around the casino floor; someone can select their bet, just scan that, hand over the money, and your bet is placed. One of the things that we have done in order to be able to do this is obviously to streamline the launch process. Our ability to launch on six or seven casinos in a short time, one of the things we've spent a lot of time on is really improving the launch process as much as possible. It's quite a large operation including logistics and configuration, etc. We're at the point right now where we could, for an operator that has already an installed base and wants to go into a new casino, just send them a USB stick, and for them to just plug it in, it will automatically configure itself, connect to our services, and be up and running within a very short timeframe. So the latest launch for Parx in Pennsylvania, we were practically not involved with. And that's not to say we're leaving our customers to their own luck in terms of installing this, but we want to make this process as seamless and as efficient as possible. So we're looking forward to more states regulating and more casino properties opening. And now I hand over to Jonas Johnson.

Hi, my name is Jonas Johnson and I'm CEO of Kambi. I'm going to talk you through what we have done to deliver for the new end users in the US, but also what we are continuing to do right now and in the future. Of course, I will not dig into everything, but I will take snippets here and there so you get an understanding of what the challenge has been. At the same time, I also hope we can get an understanding of how strategically we are looking into when we develop the product. I want to reiterate one thing that Andreas said around the self-service terminals. So I was visiting Parx last week in the US, and it was 3 o'clock Wednesday, so it wasn't super busy in the casino, but the self-service terminals were busy and the over-the-counters were totally empty. And the CEO of Parx came to me and smiled, saying 'It's excellent,' because of course that's increased efficiency for them. The more people that go to the self-service terminals, the better. So, localizing the product. First of all, we have been doing this in the past. We've been localizing products for Colombia, we have been localizing for the UK, and so on. But the USA is a bigger challenge. First of all, because even though sports betting has been legal only in Nevada, it really has a different look and feel for a localized sportsbook. So from that aspect, it's a little bit more challenging. And then also the importance for us because the market will grow and be such a big part of our revenues. So what we have done when it comes to the localization before we went live is a lot of things. Obviously, we changed the odds format, we changed how you present the different back office, you have their home team after the home team, and so on. There are changes to the clubs, to terminology, and bet types. But I think one interesting part to look into is what we call teasers. In the US, they have the concept of teasers, where on the main line you can add more points. So if that's built on a table, it's not really built on the actual odds. So for us, either we could have a choice to just copy what was existing in the US market at the time, or try to utilize the handicaps we already had. Because today we are already pricing all the handicaps, so adding four points in basketball or six points in American football is nothing new for us. So because of that, it became more of a front-end challenge, and we didn't need to go into the backend and change the actual calculations. Because of that, we were saving time when we built it, and also we could round it. It's a plus because an average end user will get much better odds with our solution than they get with the table. So these are ways around scalability that we need to think about when we build it: to do things in a clever way. This will not be used in Europe maybe, but we utilize the product we already have to do it in a better way in a new region. During this time, we are doing a lot of UX research both ourselves and together with our partners. So the localization has by no means stopped here. We will continue to build and make our product more and more American for the American users. And I think we will go in now to the next slide and look more to what we are doing now around the offering.

So of course when we went into the US, we also had to increase our offer, especially in the American sports. So this came at the same time when we saw that we really have to update our trading platform. The reason to do that is to get a more holistic trading platform which integrates the core sportsbook, the combination bet builder, and also player specials. I will not go through the player specials in this session, but think about it as a third one. So strategically we chose to focus mainly on the US sports together with soccer. When we look at the core sportsbook, this is really our trading system. It's now an integrated system. This is of course not our first version; it's our fourth version of trading systems. And as of now, we have rolled out baseball and basketball in production. They are not fully finished yet; we will continue to develop on them to add features all the time. But baseball and basketball are out in production. That's why they agree. American football will be out with the new version before the NFL season starts in the fall, and then we will do soccer and ice hockey at the latest stage. So from a quality perspective, Erik talked about that we don't only look at the margin. An important aspect is to evaluate the quality of the product. So after releasing basketball, we have seen up to 50% improvement in the KPIs we are following from a UX perspective. So that's one reason. The other reason is for us to be able to increase the number of events that one trader can handle at the same time. So this will also continue to increase the efficiency in the operation we are doing. So for the next step in the US, it depends a little bit on what feeds we will have, but we are prepared for college basketball to offer up to 5,000 college basketball events next year. This year, the available feeds were around 3,000 to 3,500. But it's also an expectation that the feed providers will cover more events. In college football, we will be able to cover around 800 events next season. On the professional sports, NFL, MLB, NBA, we of course cover all events. But when we develop these new trading tools, it's not only about the tools. A huge aspect of it is the modeling. And same here, it's not the first version or first iteration of the models that we are currently using for the American sports. The last version we went live with a new American football model, and the ambition is to have a further upgraded American football model before this season starts. Baseball we are just now rolling out a new version iteration of the baseball model. All of this we are doing to improve the user experience but also to be able to offer more markets. The last few years, we also started to look more into machine learning models in order to add predictive capacities into the trader models. So for pretty much we have been live with certain leagues for up to two years now, where really the model is pricing the markets and the trader is just overseeing what goes out on the site. We also have a big team of traders to oversee this. I want to cover a little bit about the American office. We're in the process of opening an office in Philadelphia, and the long-term aim is that we should not increase the trading team; that should stay as it is. And long term, being smaller now when we're opening in the US, what we look for a lot is sportsbook control and risk team, to have people in the US to get them closer to our operators because that will support the service level we can give to them. We are also looking for specialist US traders to complement the US trading team that we have in London and other places. I think it can be interesting as well to talk a bit about what we look for when hiring, because that might not be exactly the same as another bookmaker. The profile of the first trader we signed, we have now three or four traders or odds compilers who will start when we open the office. The profile of the first trader we hired: he just graduated from an Ivy League university, he has been playing baseball on the university team, and he has a high analytical capability. And in my opinion, it's fantastic to be able to hire that kind of people into our organization. That's something we have seen in other locations, especially in London in the past, that it is to be in this industry you can attract super skilled people from university who are interested to work with sports, their biggest passion. So it's a good place to be.

So the next part of the trading platform is the event bet builder. There is the possibility to combine different outcomes within an event. We will release that for soccer, and we have these three American sports in train. They will be in the first step a little bit more simplified version, so it will not include player specials and so on, but it will be the possibility to combine the main back office. I will not go so much into detail into the bet builder and exactly how it's built, but I think in the next picture I will point out a couple of interesting places. And here I will reflect on them on the UX research. So now when we talk with end users in the US, they can't understand that this hasn't been possible because that's the behavior that has been in the US, and is partly driven because the offering has been much smaller than it has been in Europe. So then it's fairly easy to do these kind of combinations. While in Europe, if you look, the offering has increased this much, and now we are trying to do this afterwards, and that is not always so easy. So I will go into we can have a look at our bet builder and then afterwards look at the challenges. Yeah, so this was supposed to be an animation, but I think it works very good in this way. So you can populate here with three different back offices: Tottenham to win, total goals scored, and both teams to score, and then place that combination. But one thing that the challenge is here from a technology and a model perspective is the performance needed to ask for this combination. Because a normal bet slip you can save down the probabilities in tables and it's readily available when you want to place the bet. Here, it's so many different combinations and each of them is unique, so you run into a performance problem when you're going to ask for this each and every time. The second one is there has been a possibility for us to outsource this part of the product and use that one in conjunction with our core. However, as we see this is now, it is a different tab up here that is just kind of front-end because that's how they are used to approach presenting this. This is not necessarily how it's going to be in the future. In the future, Erik talked about the core; this will be the core. Because if you looked at the American user that we talked about, they are used to combine everything. So why would that not apply to the European user even though the offering is much bigger? And for our core, we need to be able to handle this and build that internally. Hence that product development decision that we can't outsource; it's something we need to be able to build internally.

So we have talked now quite a bit about what we have been doing and what we are doing. I would also like to just finish off with a video of the product we have today. It's pretty good. So it's a video of NBA playoffs last year, and it's the same as NBA playoffs this year, and it is the instant betting product for basketball. If you follow the video, you will be able to see how you can bet on each kind of possession on next field goal, and you will also have player specials on the most famous players in each team. Now you have the next field goal. LeBron James getting a rest, wants to start 15 points on 7 of 8 from the field. Hard screen set on Shoveling. Kyle Korver loves to wolf backfires of three, knocks it down. Step 30 with this first field goal, it to me that's got you've been trapping him down. Next possession you can add the next field goal and that straight again. This is something that Christian showed in the beginning: the penalties. If we go back in time, I mean tennis, we were the first really to produce the games and the points in play. And for a certain segment of our end users, this is something they really like, but it's also something that pushes us all the time to offer a really good product. I have a quite interesting anecdote. So it was Oliver when he talked earlier, he mentioned that he was travelling around in the US having some presentations at different regulatory events. I don't know which one this was, but I think it was early August last summer. Then the day before Oliver and our legal representative were going to have their presentation, it was a CEO panel in the US. And on that one, it was one CEO from a company who mentioned that we can't offer live betting on basketball because it's too fast. And the day after, Oliver and Tommaso showed this slide. So of course we can offer live betting on basketball. It's just about the setup you have. And we promise you that the players in the US, we love this product as much as they do in Europe. And what we have seen so far is that the proportion of turnover on this market is actually slightly higher in the US than it is in Europe. So it works there as well. So thanks a lot.