Gamelight’s Günay Aliyeva On The AI Algorithm That Revolutionises UA

[Music] hello world it's pg. Biz podcast I am your host Peggy and Sals Brian Brian is actually braving the snow but he will be around for the next episode for sure and in the meantime well we're going to look at a topic that you couldn't start January 2024 without it is the word right AI everywhere I recall it must have been 2017 I'll give myself a little pat on the back that I did write in Forbes at the time that you couldn't do UA without Ai and fast forward guess what you cannot but you can do it with the help of companies like the one I'm going to be talking with because it's not just about AI to optimize it's about AI to as my guest puts it revolutionize what she has is a revolutionary AI

algorithm and we're going to chat about that and what you can do to take the heavy lifting out of UA with the help of AI so let me welcome gai ala co-founder of gam light a company that is doing just what I've been talking about right we're going to chat about AI but also we're going to talk about you because I am very intrigued your career has spanned three startups that you've founded and scaled successfully including a mobile publisher cat bite love cats we'll get to that though um so you've been there at both publisher and Company helping Publishers yeah welcome first of all to pg. Biz aleva great to have you hi piggy great to be here we're excited to discuss Ai mobile gaming mobile advertising with you and uh love to hear about the AI algorithms I'm a little bit of a geek but first first first first you know I have to congratulate you you're very active in Social and you've been talking about

your listing your latest listing in the singular data you account for get this nearly 12% of all ad spend for match three games globally um that's quite accomplishment so first of all you know congrats for you um what' you do have a have a party there at the uh at the company or how did you mark that how important was that um it's very important because we've been building the component scaling since over a year now we launched just a bit year ago and initially were thinking are we doing the right thing then we started getting the feedback from advertisers that they very happy with the quality but then uh for a year we wouldn't have something like for example singular reports or ABS performance indexes which just post for a year and then right one after another there was abside performance index and we were top three there right next to Facebook and uh Google and then few months after we saw on singular we are covering 12% of Match 3 gaming at spent so it's very exciting to see and we are

very excited to see what's coming next and you've only been around for a year so that is a story in itself probably for another show just like how do you do that that is that is Stellar such a rise but you're also in strategy you're in midcore you're across them all and I think the thread here that ties them all together is the focus on inout purchase that's what you can optimize exactly exactly so we are very uh good driving gross and retention engagement for games that Iz primarily Vis in purchases those would be midcore strategy Social Casino match three other casual titles so I would say yes that's what also ABS index showed and what singular report showed to so I have to ask I mean we'll get deeper into the algo in a minute right but that is intriguing in itself because there's something therefore about the dynamic of an in-a purchase and the ability of an AI enabled marketer through game light there's there's a match there no no pun intended for match three I just thought of that but there

is a match there about what is required and what can be done un unpack that for me a little bit why is it or is it the case that AI is effective or even more effective when you're optimizing for inap purchase as perhaps some other activity in the funnel so games that monetize with in a purchase they usually would have more events within the game right they would have different events different tutorials levels Villages and so on and uh usually it it is very helpful for AI because it needs maximum data as much data as possible to be input there um this is lacking a bit more if it's a hyper casual title for example because users stick around shorter there's less time for AI to learn there are less events for them to learn because it's more or less around watching more ads uh those also work quite well on game light depending on the genre of the games for example if it's a puzzle title with admon uh does would work well too but primarily I would say it's in a purchase because there is way more things to learn for

the algorithm and optimize towards that goes in with what you have said about your AI algorithm it looks at user Behavior you're going to see more user Behavior over time if you have more events now an algorithm is also programmed by people and uh you have had a talk recently lifting the lid a little bit on the secret or special feature in your algorithm we know it learns from player Behavior this is real Behavior real Activity real events what else do you have baked in there the best thing about algorithm is that it's self-learning so if for example there was a time period where we had a lot of data input and for some reason it didn't reach the goal it will optimize the way it's approaching the campaign optimization and it will improve the way it's recommending the games to users how the user behavior Can impact the RAS within the game and Will self-learn from its mistakes and improve it and will be true for all campaigns for all games

from then on it's about optimizing that experience getting the most you know increasing in that purchases obviously generating revenues but the prize I'm hearing more and more in the industry is when AI can sort of like look at the audience and identify the new segments the the players that you might be able to make a bit more loyal and a bit more active if you only really tried so tell me a little bit about how you extract new value out of exist in audiences or even pinpoint new audiences entirely you know there are a lot of games companies they say our Target audien is X and they completely ignore the Y and then AI is that little nudge that also helps you find those new audiences what are some examples that you can share yeah uh that's a example that you gave is very good because that's basically core how AI optimized too so how what we have seen until now is companies optimized towards manual set groups such as if it's a the strategy title they will

Target male 30 plus maybe higher OS uh but it doesn't mean that every let's say 35y old male with OS 13 plus will love their game right they will not that's an average numbers they've got from a lot of data that they've analyzed and there are also female which might be 18 years old which might absolutely love their games and this is very difficult for humans to analyze because that's very granular data to look into and you can't then simply Target one single user and then the other single user is in different age gender or S groups uh this is not happening that's what what UA managers are doing is they are setting those group based on absolute averages and hoping that now the results will reach the goals that the kpi goals that they set what AI does it doesn't do any of those groups it doesn't create any of those average segments for uh for uh campaigns what it is doing is analyzing patterns so it's trying to find patterns between different users which could be the app usage data how they uh uh behaved within the game outside the game

within our app before with other games and so on and tries to find patterns that would be the best for each game this pattern might be for 18y old female or 40y old male they might have similar patterns which means for this certain game this would be the best user no matter what is the average data says what's really cool about this is it really challenges what we think we know and and I like that I like being surprised what was the surprise in your case you know where it's like it discovered an audience like you're kidding me you know it's like just the wild thing that you have discovered and the incremental value that it created the wildest thing weon is probably the RP that we deliver to partners because we have done UA ourselves too right uh I've also been in the industry since more than 10 years and I've done exactly UA for mobile apps and games and we usually learn that you can optimize you can increase your outputs or us by 10 20% 30% if it's a good Channel um and those are the numbers that I was used to as well and then once our AI algorthm and we built it and then it optimized

first results and we started seeing twice higher output then three times higher output then four times higher output and then you realize how much uh revenues and value you can create with your game and with the limited user base and how it could have been possible before to is because it's still the same users what we do is we simply make sure that we find the right user and don't show the game to users that wouldn't love it so you get your core audience to your game and you get massive arpus thanks to the thanks to the AI algorithm and it's the same audience I mean it's the same players but you're forx and your arpu exactly so what what actually we have seen also this feedback that we get quite often from Partners um because we work with very large games too they usually have very saturated markets a lot of people play their games and they have very limited audiences already right and it's difficult sometimes first acquire new users within limited audience and then generate higher revenues with them too and what happens with us with the same limited audience they cannot generate at four times

higher revenues this massively improves the whole revenues of the company of the game the valuation because uh now instead of making way less revenues with the same limited audience you can make way more with simply using AI so you focus on arpo absolutely get that very important metric but help me understand if I'm one of these games that wants to be like you know gaming is a destination in fact now they're calling it destination gaming pretty cool stuff right where you're looking at different you're trying to be more complex it's really about a journey arpu absolutely absolutely important but there are other metrics that exist alongside that the core metrics that we uh analyze and optimize towards are R uh because this is what we hear most from UA managers too because our pool is good but as long as you don't meet your's goals it's kind of useless too because you need to get the same money back at least over whatever time that they have set um second super important metrics is retention because because we also see from UA managers how important this is

what they want to do is engage their users create a larger user base create Community make the game more popular in this case we set the goals that they have in mind so if they have certain things that are not necessarily Revenue related but some different engagement metrics we can definitely optimize towards those two what is super important is that those metrics should be natural for the game so if they say that uh we want to reach retention 70% within day 90 but this is not happening usually AI probably can't do too much you know it can still be better than other channels but it can't just do the magic and change the game because games have different um steps within them and AI can't change those things but whatever AI can do by finding the right users then those will be optimized what can be exciting about AI is it's sort of like that that thing that'll tap you on the shoulder and say look I'm looking at all the data that you're not seeing because it's way too complex and I'm watching it all the time when when you're asleep and this is my suggestion

you know I'd like to understand since you called it yourself a revolutionary algorithm you know to what extent is working with game light yes it does what I tell it to do and it does it well but it can also make that helpful suggestion to say hey have you thought of this so how it works we can't come up with new things right AI can't come up with new events so it's very limited to what our partners are sharing with us so if they are Shar ing more events with us then it can definitely optimize and pinpoint what's the best thing uh but if we don't see the events from Partners if you don't have those certain prospects for certain events AI can't come up with this because it doesn't know what is happening within the game right so it's uh very important with uh AI if you work with AI algorithms to make sure that it has a lot of data as many events as possible for as long time as possible because then you can maximize this better and if you limit it and treat it like Noni uh channels if if you simply Set uh certain Revenue numbers this pass

the postp for purchases and give few goals then it can't do much because you are right now treating it as if it was a manual targeting Network or network with no algorithms at all you rais an interesting point there gai because it is as smart as I make it which then if you zoom out to the larger picture it means that I need to give AI the data it needs to perform and that's a level of trust some people are a little cautious about that or some people just aren't informed about what data and and how and how it needs to be prepared or packaged this is a lot for a marketer to get their head around especially if they're just starting out or maybe an indie who's just embracing AI because hearing so much about it I mean the the 4X numbers would would get anybody interested in the gains that you can have what do you say to someone who's CAU cous about giving up control to the

algorithm and trusting it with this data first of all usually marketers when they start launching campaigns with us they see immediately how much higher uh Ross or po retention engagement numbers they can get we have multiple case studies where our Ur managers mention not our U managers but U managers of game Publishers that work with us they mention how they get their results immediately from day one there is no learning periods long learning periods also what AI does is because before the users even see the game we already analyze a lot of data so it's not after they download it but also before before the download happened before recommendation happened to them we already know that this user might be a good fit for this game so we can deliver better results immediately from day zero day one we are reaching the partner goals uh now this makes it very easy for UA managers usually to optimize because there is very little um in input that they can uh they do because they can simply set Ross goals they can set retention engagement goals and share the

data with us and reach their goals without too much optimization going on and when it comes to trust I think it's more of a mindset probably because we have this issue with any new technology that comes around right uh people might not trust it immediately but once they start seeing first results from themselves or maybe from others they open up more because they see that AI algorithms right now are the next generation of UA and sooner or late more companies will switch to those because it can bring much better results and there's no reason to refuse using those I heard you say that you deliver results right off the bat which is which is astounding because what I'm always hearing when I'm talking with marketers is they have this this this stress right because when you first do it it dips a little bit and then it's the stress test it's like how do I explain to my boss that I just optimize that's why it's in Decline but it's going to pick up and everything's going to be just fine yes we don't have this isue at all what what happens with us is that Partners launch

their games and within one day two days they just double caps then they double caps again they increase but it just goes extremely higher scale within very short time and usually U man just that manage our Channel manage gameel become star employees so to say because we grow to be very big for the game publisher sometimes number one network or number two number three uh within a very short time so and it's also new network because there were not too many new channels within the last two years that you couldn't work right it's very established companies it's meta Google iron Source unity appin and it's basically covering the top channels for everyone and now some uh you manager on boarded us testing something new and suddenly it became higher than those channels that they used to and it's working immediately really well so usually we don't see this concern from U managers it's more like they were excited to manage the channel because more they do more it grows and we reach their goals and they can simply launch more CH more countries or more games is

say oh it is interesting watching your Evolution I'm going to continue to watch that because of course when something has that impact uh it's going to be one to watch your company has evolved because you have a team I spoke with them they are so pumped about you it's like oh we you know and and this is without you being on the call so they're not they don't have to be like this and I've spoken with them is like oh you know she's she's um she's a multi-talent I know now that you speak several languages I didn't need to know that it's like she speaks five languages and she's a cat person how incredible is that combination uh tell me about that combination because absolutely you're more than a title you're more than a co-founder all of this comes together what's that all about for cat person uh I think we are in general I think everyone within the team are uh animal lovers so we have cat people we have dog people we have people with parrots so everyone loves do for example in my family we have cats three cats and fish and parrots and I here in Germany have a

cat my partner has cats uh so everyone within the team are animal lover so to say and I think we we're making sure that when we build the team when we were getting the people on board first they must be uh obviously experts in what they do right we we uh hired people that are very good in what they're doing if it's isev if it's account management if it's UA Tech we guys the best people within from the industry with our team and second we want to make sure that people get along with each other right because we spend so much time working with each other and we should make sure that we enjoy working with those people it just naturally happened that we have so many animal hours and our team chat is filled with gifts of cats and dogs it's just completely normal just to send cat gifts to each other or dog gifts to each other and um our previous company was also cat bite also because we loved cat boss me and my co-founder are cat people so um it kind of just naturally happened that we cat is a thing in our company I'm a cat person as well so I

just I just naturally gravitate to that and uh you know it's a different type of person but you're also a Founder um and female I won't go there because I believe personally that if we keep on talking about what's it like being a woman in Tech you sort of feel like something extraordinary and it should be the norm right we shouldn't be doing this I'm not g to say how's it what's it like being a female founder because you're doing it and that's what matters when I want to understand is how you're doing it consistently because you've scaled three startups founded and scaled three startups right that means trial and error that means lessons and learnings um share with me what keeps you grounded what keeps you focus what makes you um the successful woman founder that you are I think the first of all um when we think of funding we didn't have investment so we bootstrapped uh the companies all my three companies are bootstrapped um

which meant that we had to be performance- driven we had to be profitable to grow we couldn't simply waste investment that we would get and then go bankrupt and let go parts of the team and so on which this wouldn't happen so what we did is that we made sure that we focused on what we are doing I and my partner we both are very handson we know how to grow we know how to uh build the teams and how to do exactly if it's account management or if it's user acquisition if it's design Tech Business Development uh we can do all of this because initially when we didn't have team we two of us were managing all of this together so we had to build a lot of skills before we manag to build a team and hand over those tasks to them um and I think this is what also Keeps Us grounded because we kind of build it with our own hands without investment without huge teams from the get-go and we had to learn everything we are doing and build over it it's quite a path it's the path L traveled the other one is is a lot easier which is uh you know you don't do

everything on your own but that means you get to understand everything and you can think it through and make those decisions you know when you see it you fix it there's nowhere else to go right so that's the advantage and I want to leave our audience with something from you that will help them grasp that do that action on that be inspired by that so give us as a last note something that you would advise people to do that maybe they're not thinking about that will help them do exactly this I mean this is challenging times the games industry is being challenged but you have to up your game what would your advice be I think biggest advice I can give which I see a lot of people are making that try to launch their companies or want to build a game is try they should try to make baby steps so not from the scratch build plan the

whole app all monetization everything before they even launch because this is a direct way to fail uh they have no idea of how users will react they have no idea of if it will work or not instead they should do as fast as possible the first MVP and then build over it we we were doing updates every day every single day when we launch our first app and we were building over it because then you see the feedback from users and you say okay this worked this didn't work maybe I have a wrong idea right because usually Founders they have higher ego they always believe that they know exactly what they need to do and what should work because they have so much expertise and then they build it and it doesn't work because it's not about what is their opinion it's about people's opinion if the users love it and you should do the first MVP you should update the app often to see how users react and then build over it I actually saw also similar to this with the creatives what we're making the mistakes initially is that that we would launch a creative uh for for ad networks

and we would do it beautiful we would do it so nice and it was so pleasing and we were like absolutely in love with them and then they wouldn't work and then we would do another one which was even more beautiful and then it also wouldn't work and then we changed attitude we we're not looking at creatives only numbers if it works if we see that it's better we make more variations of this creative and if that works more varation it doesn't matter if you love it personally or not it doesn't matter if you like it if it just visually pleasing or not as as long as it works that's good and I think U just have this approach more but Founders usually lack this kind of approach to look more at numbers to make decision uh instead of purely based on their opinion so baby steps analyze get first MVP out as soon as possible and build over it that is very helpful advice and that that's the balance right you're passionate about what you build you're passionate about the game you're passionate about your studio but you also have to be aware of just reality it is what it

is and the numbers will tell you that now you can balance that against your hunch because it is about team you said so yourself you've got a great team you need to balance it and that's I guess where the ability comes in that you have gai because it is balancing that it's saying this is working this isn't working but this is my decision based on all of these inputs and I want to zoom out again to what we all always do at the end of the show we ask our guests every single one of them every week what's your favorite game you're currently playing and what's your favorite game of all time okay um currently I'm playing on desktop tropical 6 since a while now actually and project highrise for mobile it is uh match Masters by candore and dice streams by Superplay U huge fan of both games um of full time is more diffult to say because it depends on the time I

would go from uh match three titles to Hidden Object to simulation game so it depends on the time um but yeah I would say for desktop it's project highres for mobile could be genies and gems or also match Masters so you're just as equally into Match 3 as your company is excelling in it that could be a reflection of that interest I mean M three games are amazing I think whomever I introduce them to if if they were like no no no I don't play this kind of games and they would play for first two minutes and they hooked they're like oh I'm on level 200 I'm on level 300 so I think they are great games it's addictive core mechanics I find that too it's like I'm just GNA do a little bit and it's like no I'm actually dipping in quite often and I'm getting quite into this probably because it doesn't seem like that it just sort of builds right and uh and speaking of Builder you are quite a builder yourself in your company um I'm really interested in learning more about game light I will be watching it and uh

also watching you you are out there speaking loads and doing stuff gai where could someone catch up with you in person maybe even get a couple tips in person about yeah founding funding scaling the whole bit what would be the best way yeah um I'm based in Hamburg so if anyone is in hurg drop me a line on LinkedIn I would gladly meet I also often go to events that I was in Hamburg Berlin uh sometimes with in Europe too usually it's a bit more difficult with time but if I if there is any event in Berlin or Hamburg make sure to check out because I might be there well that is very cool because I'm based near cologne didn't know if you knew that and I've always wanted to just sort of like do a little trip up there you know do the Ino games and all the cool Studios up there I will stop by and see you as well and until then I have to thank you for a very energetic and inspiring talk thanks so much for sharing and hey uh my my my greetings to your three

cats thank you very much Pig and it was really nice being with you here have a great day this show is all about how to do your job better how to make an amazing game how to Market it and you have a say so if you have a story or know someone we need to shine a light on then we would love to hear from you we want to hear from you we want to reflect the reality of the mobile games Market in all its wonderful complexity and strangeness so if you have any suggestions for us if you have any feedback for us you can always get in touch you can email us at podcast pocketgamer.biz you can find us on Twitter at PG Biz and you can reach out to us through the pocketgamer.biz website if you're interested in listening to all of our podcasts you can find them at pocketgamer.biz SLP podcast and we would love to hear your thoughts on future shows and we've got you covered on all the major platforms so

subscribe to the audio podcast as Brian said look for us on YouTube if you want to read it hey you can do that too because we have a companion post for you as well on the pocketgamer.biz website tune in again for the next edition of the pocketgamer.biz podcast and we look forward to speaking to you in the near future until then I'm Brian Bago I'm Peggy and salts and that's a wrap until next week oh