Supercell's CEO, Ilkka Paananen, on How AI Is Changing Game Development Forever | ElevenLabs Summit

[music] Okay. Okay. [music] [music] Ela, it's so Great to have you here today. And to set a bit of context, Ilka founded Supercell in 2010 with a mission to create games that people play for years and remember forever. Since then, the team has launched so many iconic titles such as Clash of Clans, Heyday, and Boom Beach. But I want to start off with something that you've said. You've said that it's never been so hard to launch successful new games. And in 2024, you even mentioned that about 60% of mobile play time went to titles that were at least six years old. Can you share with us a little bit on how the cell approach at Supercell helps you make say competitive in this market and maybe you can share a little bit about the founding story of the cell approach to team building? Uh, sure. So, first of all, thanks for for having me. I've been a huge fan of 11 labs and and Marty as a founder for a long time. So it's a big big honor to be here. Um so the story uh behind the cell approach is actually relatively simple. So when we founded Supercell like we there were six of us co-founders uh and each of us had been in the games business by that time uh for roughly a decade or so. And then we had like uh and we had been lucky enough to be part of some quite successful companies and and in these companies like this thing would like repeat over and the same thing would repeat over and over again which was that the more successful the company became like somehow like the the creators the creative people these kind of game designers people who you know program the games and people who do art somehow these creative uh game teams they kind of lost control and and you know like business people like I guess like myself like somehow like would come in control and would start to dictate to those teams what to do and um and then we like we had this and and and it's sort of like a it it results in a very kind of a process driven like optimizationdriven culture and maybe it's it's really hard to see like how outlier hits would come from that type of an environment. And then we had this idea that what would happen if you would give the complete control of the game to the game teams themselves and and what would would happen if you could flip the organizational organizational chart upside down. So actually like who actually ultimately owns the vision for the game and owns the entire game is actually the game team and not to say the leadership team of the of the company. and and then somebody said that hey uh this is a pretty radical approach and we probably shouldn't even call these game teams like teams let's call them cells and then you know somebody asked the follow on question which was that well what should we then call the entire company and that's then how the name supercell came about amazing and building on the philosophy of giving more agency to your teams um you have also said actually that you are the least powerful CEO in the world. How do you actually see what great leadership looks like today? And what is your top three responsibilities being the least powerful CEO in the world? Well, at least in our culture like the most important element is trust. So, so basically my and our goal is to hire the best possible people in the world. Then they out of those people they try to form the best possible teams. Um and and then they also try to create kind of culture where those teams can have the biggest possible impact. So that's sort of a I see culture as this kind of fuel that you pour on creativity and then there's a massive fire uh in the best case. Um so uh in a positive way uh and uh so that's how I I I see kind of my and our leadership team's responsibility. So we they are there we are not there to build games or decide what type of games to build or any of that stuff. We are there to kind of you know pro uh create sort of conditions where those games can be be uh be born. So that that is really leading idea and and then like the logic is that you know if you kind of imagine a company or a company like Supercell where these individual cells have like full autonomy uh they are completely independent with responsibility of course like there's like at least in theory there should be like two really great things that follow from that logic. The first one is that uh because these cells can decide on their own. So they don't need to go to the upper level of amate management to get approvals for anything. You know they can move really really fast. Uh and and the second uh point is that because these cells can make decisions on their own that means that the decisions are made as close to the user or to or to the player and as close to the product or to the game as possible. Uh and in theory that result in much higher quality decisions. So that those are kind of the basic principles and values that they've built the company around. And when these teams are building these companies um or games, uh how does the development process look like? And is there a big variance between cells once they got in the green light? Well, uh how we like to think about it is that you know each cell has to figure out their own path to what a great game looks and feels like. And and you know one of our biggest advantages is that you know we sort of tolerate like differences between the cells and uh and and you know one of our like biggest sources of strength I feel is the diversity of between ways of working of of these sort of cells. So basically you could like in especially on the new game side where we could try to innovate on new games you could think Supercell as a say almost like a almost like a venture capitalist that kind of like you know makes bets and investments to these new game teams and in a venture capitalist portfolio I mean it's not that one portfolio company is is similar to the next one. They're all kind of different and that sort of is the whole point. Amazing. And some of you in here might actually be aware of the Supercell Champagne approach, but for those of you who are unaware, Ela, could you tell us what is the champagne approach to failure and mistakes at Supercell and why is it so important to your success? Well, it's so basically we the goal here in our culture is that you know we have this one of our core values is ambition meaning that we always try to like shoot for the moon. like whatever we do, we always try to shoot for the moon. Uh and we believe that at least in our business like uh the only way to make a outlier hits is to build something that doesn't exist yet. And and then if you're trying to build something that doesn't exist yet, then you know, by definition, you need you need to take a lot of risk and by definition a lot of things will fail. and and we actually see failure as a kind of a necessary part of a creative process. And most of the stuff that you're going to do, it's going to fail and and it should fail. And and and we often times say that uh if there would be a year and every single game project that we do succeeds, then we should be really worried for two reasons. We first like maybe our bar isn't high enough. Actually, we are not trying to shoot for the moon. uh and and uh and and and maybe we aren't taking enough risks. So um so therefore we try to create this culture that really like encourages risk takingaking and kind of celebrate failures and and we are not trying to like pretend that fail failing is fun. Of course it's not. But what we think is worth celebrating are those learnings that come from the failures. So uh so to cut the long story short, we have this tradition which started 15 years ago that uh if we kind of kill a game project, so there's kind of this failed game. What happens is is that the game team, the people who built the game, they get up on the stage and uh it used to be a very small stage when we were like 20 people. Now it's much bigger because we are 900 people, but it's still the same thing. They get up on the stage, they tell to everybody like you know what they learned through through that uh failure and and then we as a company we you know offer a champagne and we kind of toast to the learnings from that failure. And what are the explicit kill or scale criterias in this process? Uh well actually even that is like up to the teams. So they can again as I said they completely own it and they are in control. So when we going to kill a game project, an extremely important part of that thinking is that it's not me or anybody from a leadership team. It's not anyone from the outside of the game team who can make that call. It is the team and the team and ultimately the lead of the team who can make that call. Uh [snorts] which I believe makes it a lot easier actually to make those decisions. Um uh and uh you what and basically like how a decision is made like and kind of we basically obviously we ultimately look at like maybe two things like first is of course the obvious one the data we look at the game metrics and mostly we look at retention and engagement metrics like you know things like day one retention day 30 retention and and and even longer term retention metrics. Those are the most important things. Of course we do look at monetization but that for us that's the secondary metric. We believe that if you are able to create something that is you know you know provides is fun for people to play and that means that you know they come and return uh to the game like you know day after day you know over time then they can figure out like how to monetize that experience. Uh but it but you can't do it the other way around. I mean if you don't have a great product that people love and they are proud of playing and talking about then then you have a problem. And if that's the case then the game should be killed. And for those of you who are not aware, Supercell absolutely crushed it last year. If you see the numbers from last year, it's absolutely fantastic. So big congratulations on that. And a big a big reason for that also been your huge success in creating communities. How so once these titles, you know, have been launched and and and created, how do you go about the community building um and getting that distribution? Well, um, first of all, like it all starts from like building a great product. It's it's sort of, uh, obvious, but but that that we always try to like focus focus on that. So that's the ultimately like I shouldn't say that's the only thing that counts, but that of course is is the most uh important thing. But once you have a a a a great product that actually is worth talking about, then actually what you know the great thing about today's like app store ecosystem is that people actually do talk about these games if they're kind of proud of of of playing them. And and I and I guess like one like something that makes super quite different among mobile games companies that we are way less like reliant on user acquisition and and the reason why we are way less reliant on it is is because we believe that we you know build these great games and people talk about it and that's how we become like this viral phenomenas um and and um and and you know our approach to community is we try to be as open as transparent we you know try to you know when Of course, every once in a while we screw up. We make mistakes. And if we do that, then we try to, you know, admit them and and try to correct them. We try to listening to to the community around us, which by the way sometimes is extremely hard because there's imagine a game and it has say say a 100 million people play it every single month month like they play one single game. Well, there's a lot of different opinions among those 100 million passionate gamers and and you know, there are like this kind of hardcore, you know, fans and gamers who like spend like lots of time playing the game and they know every single detail of the game even better than we do. And then there's more of these casual players who maybe have like [snorts] five to 10 minutes per day to play it and they don't maybe take it as seriously and they are not as competitive. But then how do you balance between all these different uh uh you know requirements and and wants uh it's uh it's it's really really um uh sort of a difficult process I think. Yeah, I can imagine. And going into a little bit on AI, we have to dive into AI. We're after all at an AI summit at a high level. How is AI influencing the development process? Well, um sort of um there's two ways to think about it and and [snorts] I think both are true. The obvious one is that it just makes everything faster. We can make better quality products faster than ever before. And I see this like accelerating a lot and and the way we going to think about AI is that AI uh gives our like creative developers like superpowers that they didn't have before this age. And we very much like see still the human creator, the creativity in the front and center. But then it just again as I said AI just enables them to do more and better for for the players. But that sort of is the that's the obvious stuff. Uh the other side of of same coin which I'm at least as as excited about if if not more is like what type of new game play might AI enable our developers to to kind of develop for the players? like what type of new games could we invent today that were not possible even like three four years ago and and you know in a in a funny way um this sort of a time uh reminds me actually of a year 2012 I remember it was a game developers conference in San Francisco and and and Supercell had developed like bunch of games and two of those games were heyday and Clash of Clans and I had those games on my I think on my iPad and and and I went to San Francisco and I I had like and I wanted to like show those games around those prototypes to like some very senior and experienced and very successful people in the industry and and you know typical sort of a case would be that I I take my iPad I you know launch Clash of Clans on it. I hand over the iPad to the other person and the other person takes it and plays for a little while maybe 20 30 40 seconds hands over the iPad pack back back and asks uh so I I don't understand like why would anybody play this game and I [snorts] just remember like how crushed I was because I really looked up to this these these people and then since those days like Clash alone has generated way way more than 10 billion dollars in in its lifetime you know more than a billion installs and all that and still an extremely successful game. And I've been like reflecting on on those moments and and like why didn't these extremely successful people in games like why didn't they see it coming? And I think the answer was that it was such a different thing. I mean these these I mean they they couldn't first of all why would you give out these games for free? Doesn't make any sense. And and and then the UI is is different and it's a different type of a game. and and they even like questioned they were that's a real game. I like I can't remember how many times I heard this term that you guys aren't developing real games because real games are played on PC or consoles and and etc. So that was 2012 and of course because of this type of attitude it enabled you know companies like King and us and and many others to sort of like emerge and and become quite quite big. So that was 2012. Well, when fast forward to to today, my biggest fear is that so today like Supercell is one of these like one of these like bigger successful company like 290 million people play our games every month and and and so forth. And my biggest fear is that you know now GD GDC game developers conference is again approaching. We go there in in March and then this like uh uh maybe somewhat younger person comes in with some kind of a device whatever that device is hey we want to take a look at the game that I did and it's some kind of like AI native game it's some kind of transformative like experience something that they've never seen before and then we would ask that hey I don't understand like uh why would anybody play this game it doesn't even feel like a real game [snorts] and and you know that's something that um um I don't know if it keeps me up at night But we're definitely like you know we we talk about it a lot at the company and and therefore you know we've like for example founded what we call AI labs which we have in Tokyo and San Francisco and Helsink and we are really trying to invite uh also other people to come in and almost we are welcoming them inside to Supercell and and tell them hey disrupt us you know disrupt the way we think about games. No, that's super interesting. And you also mentioned something very interesting a few minutes ago on that your creative teams can get superpowers with AI. What comes to mind when you think about that? Could you give a share a few examples of how where you see your teams are having a bigger impact with AI in the creative process? Uh well few examples come to mind like immediately. The first one is that I I was talking about kind of the the speed of iteration. So if you think about uh the age before AI like it you had an idea uh and then you somehow had to like uh either you create some sort of a prototype you have to somehow visualize your idea for example and hopefully there's something to play and you know and and from that sort of that first idea to like actually having something on device or something it probably would take you at least a few weeks at at at so in in a even best case. But now that same time can be literally like uh hours and and that what that means is that you can like you know and that's what I mean by that boost creativity because you can try out like so many different things so so quickly and and and and what's best is that you can actually try those your ideas out your real players real human beings uh and and and get their feedback and and I think it relates to like you know obviously it relates to visuals uh audio as as you guys well know um and but even like you know the entire playable game prototypes and and I think you know thanks to that I [snorts] and and and also the best thing is that uh like one single person can actually do all of those things. So previously like I guess a minimum size of a team would be that at least you would have an engineer and artist at least two people but oftent times say a designer and uh uh and and and maybe have like client engineer server engineer etc etc and for anything to happen like you know those say three to four people need to agree on the idea and you know I've heard a lot of complaints about for example from say designers or artist that oh I tried to convince that client engineer to do this but you But but she wouldn't just do it and there's a lot of arguments and etc. And all that makes the iterations like slower. But now what if there's just one person? Yeah. And who just you know you know tells the AI like what to build. So but I I feel you know we'll see a lot of great games just because of this uh much much faster iteration cycles. Yeah. And fast forward, let's say like five years from now, what do you think the future of gaming will look like and me maybe even entertainment in general? It's a dangerous question. You know, I get a lot of invites to or at least used to get a lot of invites to these um panel discusses where the topic is future of games and I always like decline them all because it's impossible to say and and my sort of like how I think about it is that you know no one knows. I mean games is it's extremely hard to forecast. I would argue it's impossible and and and the bigger is the hit game, the harder I mean I mean say think about something like say Minecraft which I'm sure is familiar to many in the audience like I mean who could have imagined like what it would become and of course the same is very much true about the Supercell hit games and and and that Clash of Clans example in San Francisco 2012 I mean even the brightest people in our industry people had seen it all uh they couldn't see it coming. Um so uh so I guess a short and honest answer is that I have no idea but I what I've tried to do is is what I can control which is I try try to make sure that we have the very best people the very best teams working in a culture which you know somehow uh helps them to figure that out and and and somebody has some sometimes said that heat games are almost like accidents and and what what kind of a leadership of a game company can do is that you want to like create the conditions which kind of maximizes the probability for those accidents to happen. And um as content is constantly exploding, what do you think is the hardest to maintain? Trust, taste, or attention? Uh what a fantastic question. Um doesn't have an easy answer. Um I uh I mean lots of people probably don't yet think this way but I actually think that probably over the years like um AI can also develop like some sort of taste. Uh I'm I'm quite sure that it'll happen uh given how fast things are like moving forward. Um, so it's not um, so I I I probably would say trust because [snorts] I mean that ultimately like that's probably like the most important thing and it almost it's it's and it's also very binary in nature. You either have it or or you don't. Yeah. Yeah. Totally. And in an AI enabled organization like Supercell, how often do you think that leaders should be revisiting team structure or maybe even decision- making? Well, given like how fast things are moving, how fast things are changing, I would say that that you you should be revisiting uh and questioning and challenging it like uh all the time. Great. And as a very final last question, Ilka, as a very busy CEO, when do you actually find time to play these games yourself? And I have to ask, which one is your favorite at the moment? Uh well um well, I wish I had had had even more more time. Uh but um I mean but obviously I play games and and you know obviously I play uh the supercell games and there's actually a lot to play because we have like only five live games but then there at any given point of time there's 10 plus game like in development and of course those are the ones that I also al always very curious to uh to play. Um which one is is my favorite out of those I I just can't say. It's uh almost the same as asking like uh like like what who's your favorite child? So uh you know I can't answer that but uh but like outside of the super uh games I actually tend to play a lot u what my kids play like and and that changes often and their taste taste quickly and uh what have we played recently? I mean my my son for example played Fortnite a lot u then he and actually now has just switched back to Rocket League so we played that quite a bit. Nice. And have you uh when you're playing Fortnite, have you experiencing the 11 Labs speaking to Darth Vader mode agents? Absolutely. Great. Well, it's been great having you here, Ela. Such a pleasure and thank you so much. Thank you very much. [applause]