Making mobile games go BOOM in 2023: balancing pleasure, pain, and generative AI
Why not build mobile games that are only and always pure fun? Because they will completely fail and your users will churn.
Recently I was complaining about a game I play, Alien Shoot: Galaxy Attack. My win/loss ratio hovers around 50% despite every power-up I buy or new, powerful weapon I acquire, XP I accumulate. That’s precisely as designed, I learned.
“That’s the art of building a great economy and game progression systems … it’s a balance between pleasure and frustration,” says Elad Levy, who sold his last game to Playtika, on a recent episode of the Growth Masterminds podcast. “Those are the most successful games. Sometimes you’re super happy. Sometimes you want to throw the phone out the window, but that balance between pleasure and frustration, that is what builds amazing games.”
Levy is pretty much a game maker’s game maker, or a game marketer’s game marketer.
He’s been a head of growth, digital marketing manager, VP product, VP marketing, an operations manager, and spent the better part of a decade working with companies to implement and optimize their marketing technology solutions for analytics, reporting, and automation. And yes, he has sold a game company to Playtika, which is a pretty cool thing to have on your resume.
We chatted about making games explode … making games grow fast.
Marketing comes after the hard part in mobile games
That starts with making a great game, of course: the hardest part of making a game grow. Other challenges include implementing LiveOps and/or CRM, and ensuring your retention and LTV numbers are where they need to be to support growth … all the hard stuff after launch.
In a sense marketing a game comes after this.
But in a very real sense, marketing also precedes this.
What game are you going to build? Who will play it? What sub-genre will you fit it into? Will you build a game in a vertical that you already know, so you understand the KPIs and benchmarks and costs, or will you break new ground? All of those are questions that game marketers can be part of answering.
The other hard part is working with product on the most important task after building a great game: making it monetizable so it can support growth budgets.
“Your next most important task in games … is to maximize your lifetime value out of every player. And the only way to do that is through LiveOps and CRM activity,” says Levy. “Basically: continue and automate content pushing to players, optimizing pricing, upselling, focusing on conversions and continue to maximize lifetime value while working hand in hand with the UA teams to make sure that the channels that they bring tie back to those who bring the maximum value of those players, the best players and not just the one that go in and out or turn very fast.”
That’s critical, says Levy, because the more you make the more you can spend, and a game that generates a known and proven return on ad spend is a license to print money.
Automating the mobile games machine (which can include generative AI)
Building the foundation for growth is one thing. Scaling that into a massive global game is yet another.
“This is where you need to be able to combine the whole thing, the whole nine yards, technology and product and automation,” says Levy. “You need to combine and automate everything in a way that … you can pump more and more and more content.”
It’s a content treadmill that requires more and more people, but there are solutions emerging that can take the edge of that massive personnel expansion. One of Levy’s clients has an automation engine that builds — and tests — new levels in the game automatically.
Others are starting to use generative AI to build bits and pieces of art or story for their games, with varying results so far. Even the most technical game publishers, however, like to reserve a human touch.
“To be honest, it’s really hard to trust a computer to do everything,” Levy says.
One example with deep and engaged fans of successful mobile games: they’ll notice anything this is off-brand or doesn’t quite fit with the story. So ensuring there’s quality control is key.
Much more in the full podcast (of course)
As usual, there’s much more in the full podcast. A few examples …
Brutal honesty on testing:
“Most of the time you do something and you have no idea how it’s going to affect and what it is going to affect exactly.”
Insight on LiveOps and/or in-game CRM:
“The beauty of having the right technology in place is that you can create new journeys for every new user.”
How to get free goodies in mobile games (as a gamer, not a producer):
“Usually in seven days of no login is a funnel we call LTC, likely to churn.”
Why building something in the same genre as your existing game is smart:
“If you specialize in the genre, then you already know the UA numbers … you know all the KPIs. So if you launch a similar one, you know, there’s always a place for a second spot or in the third and the fourth. And, you know, there’s a lot of low hanging fruit and you can make a version that is maybe, maybe you target with the first version, the high paying countries, like the tier one countries, but you can maybe do like a third world country where you focus mostly on ads and not on the in-app purchases, there are so many ways.”
Subscribe to Growth Masterminds on YouTube and/or your favorite podcasting platform …
And here’s an idea … click play on the video while you scroll this post and scan the transcript?
Full transcript: Making mobile games go BOOM
How do you make a game go boom in 2023? Hello and welcome to Growth Masterminds. My name is John Koetsier.
Today’s a bit of a different Growth Masterminds. We’re chatting with one of those people. He’s a growth expert. He’s been there, done that, VP marketing, head of growth, SEO manager, 20 experiences plus on LinkedIn, 20 years in games, sold his last company to Playtika, and spent the last six years helping brands and apps get big fast.
His name is Elad Levy. His name is Elad Levy, and we’re gonna talk about making a game explode. Welcome, Elad.
Hi John, nice meeting you, thank you for having me.
Super pumped to have you awesome. Okay, let’s make some assumptions here. We’re going to talk about making games grow and we’ll talk from, you know, starting also to scaling, but let’s make a couple of assumptions.
First assumption, it’s a good game, right? You know, we built it, it works, it’s a good game, maybe even really good, possibly great, but let’s not assume that because, you know, if it’s astounding and amazing, it probably grows itself maybe, maybe that’s a myth. I don’t know, but let’s assume it’s a really good game. So it’s got potential.
And let’s also assume we have some budget, a decent amount, not we work below billions overnight budget, but we have some budget. Where do we start?
Well you actually jumped through the whole part that is super hard like finding something that actually clicks but
You are allowed to go back in time. You’re the guest. You’re in charge.
I’m no, I mean, it’s okay. It’s okay. Maybe that’s a different podcast, but let’s say that the game clicks, the KPI are there and you will be chased basically by every VC on earth to accept their money because.
Money in hand chasing you down.
Exactly. Exactly. So I guess that your next step would be to swallow market share as fast as possible and you would need to double down on UA big time… like a lot of user acquisition in every possible channel.
And hopefully you had a good foundation before doubling down on UA so the data points and tracking is there so you can measure it properly. If not you can …
Let’s not assume that. Let’s assume that we, you know, we built a game and we think it’s fun and people think it’s fun and it’s great, but you know, we’re kind of new to the growth game.
What do we need to do?
So I will jump forward now and say you have like the executive team and you have the data, the marketing, and you’re starting to push money in UA to scale.
Now your next most important task in games – and I’m specifically focusing on games that let’s call it game as a service or games that live for a very long time – is to maximize your lifetime value out of every player. And the only way to do that is through LiveOps and CRM activity. Other industries actually call it CRM and the games industry, they mix it with LiveOps, but basically continue and automate content pushing to players, optimizing pricing, upselling, focusing on conversions and continue to maximize lifetime value while working hand in hand with the UA teams to make sure that the channels that they bring tie back to those who bring the maximum value of those players, the best players and not just the one that go in and out or turn very fast.
That is like the biggest step because the higher you maximize your LTV, your lifetime value, the more you can bid when you buy players, because you already know that in 18 months that player is going to give me back $200 so you can compete on a completely different level and you can buy really expensive if you want to.
So that really changes the game.
How do you work with LiveOps and product together? Because your game, your app is always changing. You’re adding new content, new seasons, new levels, new, let’s say, ships or weapons, new challenges, all that stuff. And you’ve got your LiveOps as well.
So you’re always fixing the plane or building the plane while you’re flying the plane. How do those two things integrate?
Think about it as a journey. And the beauty of having the right technology in place is that you can create new journeys for every new user. So let’s say that there are new users that go through new journeys, and there are existing users that already go through some sort of a journey.
And there may be players that are likely to churn or already churned, and you want to reactivate them. And that is another treatment.
But if you focus on those that love and engage and play the game a lot, then that basically has, you can treat it as something completely different, it’s a running product, it’s live, and you need to continue feeding them. So if they’re watching their favorite series on Netflix, you just need to continue pumping seasons.
But if it’s new users and that series is not appealing to them, you’ll have to find some like … some else, something else to convert them and to hook them on too, you know, like find another hook, basically. The game will stay the same, but it changes a lot if the whole onboarding experience is different.
It’s really funny to hear you talk about that because I have one game that I play most of the time and I have this continual fantasy, right?
I will get good enough. I will have enough stuff. I will accumulate enough power that I will be amazing and unstoppable. But my win/loss ratio in my battles is always in and around 50%. Sometimes a little higher, sometimes 55%. Sometimes it dips down to 44%. And I think, oh man, they really want me to buy something now, don’t they? And it’s a mountain that you never stop climbing.
That’s the art of building great economy and game progression systems … it’s a balance between pleasure and frustration. Basically, those are the most successful games. Sometimes you’re super happy. Sometimes you want to throw the phone out the window, but that balance between pleasure and frustration, that is what builds amazing games.
Interesting. So if I’ve built a game and it was all pleasure, that wouldn’t work?
No, people will churn and if it’s too hard, people will churn as well. If it’s too easy, they churn. If it’s too hard, they churn. It’s always like that.
I mean, I see my mom play. My mom is 75. She still plays Candy Crush. She’s like in level, I don’t know, over a thousand or something. She’s been playing for years. And she already learned and she’s no, she’s really no… computer or games expert, but she already realized that when she cannot pass a level she lets go of the game and after like a few days … all of a sudden she can pass because you know, this whole thing is money. Everything is manipulated. That’s the beauty of it.
She’s gaming the game. She’s telling the game. Hey, I’m not looking. I’m walkin away.
I’m about to churn, yeah, I’m about to churn. Do something because I’m about to churn.
Give me something … do something … a little tip for gamers here. You want to pass the next level? Stop playing for two or three days.
Well, yeah, exactly. Usually in seven days of no login is a funnel we call LTC, likely to churn.
So you’re doing live ops. You’re building the mountain that never ends and you’re adding seasons and content and other things like that. So that’s all working well for you.
Talk about scaling this to … a significant size. We’re not just talking about a small game. It’s doing well. It’s making you, you know, low seven figures or something like that. Talk about scaling to the next level.
Well, this is where you need to be able to combine the whole thing, the whole nine yards, technology and product and automation anyway. You need to combine and automate everything in a way that … you can pump more and more and more content.
Now to do that is not that easy because even if you automate, you need a lot of people. And this is where the company starts really growing in the amount of people because more art, if it’s a narrative game, oh my God, like the amount of, like that’s the worst.
It’s a content treadmill that never ends because you have to continue writing and writing content with no end. And that also requires rethinking the whole way you manage your R&D teams. And when you launch features, giving them more independence, the whole thing can be launched on its own.
And, you know, doing a lot of automation, but this is really next level stuff. And when companies reach that size, most of the time, the original CEO would probably not be the same CEO because it’s one thing to manage, I know, 50 people or 10 or 20 or 100 and then another to cross a thousand or over 500.
Yeah, there are some people who are interested in building stuff, and there’s some people who are interested in scaling and managing stuff. We talked a little bit as we’ve practiced about AI, generative AI and AI in other areas.
Where can that help people who are working to scale their games?
I would actually go back for a second and there are more ways, I mean, to grow. And some of them will take the same game, re-skin it and try to launch a similar game, because you already have the engine.
So, you know, you can do the same thing with a different theme and then capture more market share. So some of them would try to do that.
Another game in the ‘Scapes series.
Exactly, exactly. Yes. And they’re all great, by the way. They’re all great.
Ton of art, tons of, you know, tons of content, tons of like, there’s a story. The game is not the match tree. The game is the story and fixing things. That’s the game itself. That’s the actual game.
But yeah, which means more, you know, UA and more playable ads and more everything. But Some would probably try to do that, which is logical. It makes sense. Some would be brave and try to do a completely different game. I actually do not recommend that because it’s crazy.
I mean, if you specialize in the genre, then you already know the UA numbers, you know, they are down, you know all the KPIs. So if you launch a similar one, you know, there’s always a place for a second spot or in the third and the fourth. And, you know, there’s a lot of low hanging fruit and you can make a version that is maybe, maybe you target with the first version, the high paying countries, like the tier one countries, but you can maybe do like a third world country where you focus mostly on ads and not on the in-app purchases, there are so many ways.
And if you take content that is really amazing, for example, Angry Birds. They took the same content and squeezed it. And after running it on every possible device, they started making movies. And because you have an IP at that point, you know, it becomes an IP and that IP is valuable.
There are very creative ways to continue growing. And about the AI, yeah, the AI point that you mentioned.
Well, I didn’t really see, I see a lot of experiments, but all of them miss polish. I think that on the content generation, this is where it really is like generative AI, you can say that is where I see it helping the most when you need to create, I don’t know, more text or more art.
I mean, let’s, let’s take hidden object games, for example. That’s one of the biggest nightmares in the games industry because content is so hard to make. If you can automate that, obviously that would be amazing.
So there are a lot of opportunities, you know, in match three, we actually have a client that has like the whole level building is automated. So it builds the level and then it plays it to see how it plays. And and if it’s good it moves on to the next so there’s already things that are out there
Of course, automation isn’t necessarily AI. I’ve seen that in terms of generative AI, people creating objects, whether that might be a weapon, that might be a power up or something like that, and that’s happening right now.
I’ve seen a CEO who’s spending $8,000 a month to give all of his or her, I don’t know which it was, employees, OpenAI, subscriptions to OpenAI, right, and said, hey, this is cheap, $20 a month for each employee, it’s a force multiplier. It makes them more efficient, more effective and gets more done. And so I think that we’re still early in those days, but I think that people are already getting some of those benefits.
How do you see AI being used in user acquisition?
Well, probably on the ads side, I mean, there’s already automation that is used for years with different tools out there. But for example, in Dive, we currently hook up data insight back and feed it back to Singular to optimize UA campaigns. You could possibly put an AI engine to do the whole thing, but to be honest, it’s really hard to trust a computer to do everything.
You have to like to supervise it somehow.
So I mean, it sounds great. It sounds really like a Skynet and stuff, but the bottom line is just you need someone to supervise it because if it goes crazy, it will just start, you know, creating campaigns all over the place and go wild.
Yep, yep, yep. And you talked actually as well about just plain old hardcore data analysis in concert with product to see what’s going on in the game.
Yeah, so I think there are several steps in data. First, getting it right, like getting the foundation right, because if the data is inaccurate, everything you build on top of it is just, we call it garbage in, garbage out. That’s like the insight you get.
But if the data foundation is clean and organized, then you can start getting the first insight with the data analyst working back to back with a good product manager and optimizing. Let’s check this feature, let’s check that feature.
Let’s see how the users do that, I don’t know … correlation, that’s something that I love doing a lot in games. Like the users who completed like this achievement on the second day. And what did they do in the days after that? Did it lead to more engagement, to more sessions, to more revenue? So that’s, there’s a lot of work that can be done with, with a couple of people that work a lot of hours in the same office or Zoom call or whatever.
And then the next step after that would probably be machine learning, mostly recommendation engines and regression analysis, where you try to predict models like churn, or predicting paying users, because even if you can predict something that eventually leads to a more engaged user, then that is enough to feed it back to the UA engine and optimize it based on that.
Instead of the boring, like when you do UA today, it’s super boring. It’s like either installs or revenue. And those who really like take it to the next level, big companies growth, that’s what you ask growth questions. They like to start feeding custom events, custom actions back to the UI, back to Google, back to Facebook, back to the … UA channel so that they can optimize according to that because they found the correlation of people that use this weapon end up playing more time a day.
Why? I don’t know. That was the data set. But there’s a ton of things you can do with machine learning and prediction and tying it back to the game or to UA or yeah, lots of things.
Do you often find some of those correlations of things that you wouldn’t have thought of? It’s the non-obvious sort of KPI. It’s the one that, you know, why does that work? It’s often those things that are most predictive.
To be honest, most of the time you do something and you have no idea how it’s going to affect and what it is going to affect exactly. Like you can, it can be a button, it can be a color. 20 years ago we used to A-B test colors, you know. So I know red would be, I don’t know, it’s an alert color. Blue is relaxing; green is like go for it. But then I don’t know red would be in one country would Like be one thing in another country. Maybe it would be another thing.
So the whole world of data insight and personalization is so huge that when you get to AI for me, it’s like so much down like the road … there’s so much work before that to be done that And most of the big companies, most of the data team is busy on that.
And then you maybe have like a small team that is doing the AI, the fun stuff that, you know, maybe don’t move the needle that much of the company. But, you know, it sounds good on podcasts and conferences.
Yeah, sounds good on the podcast. Looks good in the investor deck and all that stuff.
I suspect that will shift as we’re seeing massive innovation in AI, but of course there’s always a lag effect between shiny new tools and productive application of those in operational scenarios.
Exactly, exactly. It’s one thing to like, you know, it’s one thing is the buzz. And then another thing is how really useful it is, you know?
It takes time for it to sink in and it will probably sink in at some point because the early work is really amazing, but I don’t know. I mean, even if you’re a small to midsize game studio, I don’t know if that’s … again, it’s good for raising funds. but it’s a completely different story when you get to it. You can generate text and content, but maybe you’ll go over it and there are typo mistakes or the content is irrelevant.
Or, you know, small things that when someone is very engaged with the product he will find all details super fast, like if you watch your favorite series on Netflix and they would do something like an AI … I don’t know computer …
A computer making the script …
… a new computer or change something, you would say, hey, that doesn’t make sense. You will notice in a second.
Yes. I will say that ChatGPT by OpenAI is pretty amazing.
I’ve seen, I’m a subscriber, and whatever I’ve seen from it has been incredible and not obviously to the human eye machine generated.
Also, you can tell it to adopt a persona. So if you could at some point, and I don’t think you can do this with ChatGPT, but perhaps you will in the future with productized AIs around this to adopt a persona. To adopt, and that persona is, you know, kind of how you present your game or the flavor. It might work. I don’t know. I guess we’ll find out.
It might work, I don’t know.
A lot of this has been super interesting and super relevant. Maybe let’s … let’s end at the beginning. As we started and I started asking about scaling and growing, you said, you skipped the hard part. Talk briefly about the hard part.
You already built Westworld and you still …
Exactly. It’s perfect. Clash of clients. Here you go in your lap … market it.
Yeah, yeah, you already built it. Yeah, the initial work is insane.
And most of the game studios do it wrong. And in a way we work on data and live apps, but we constantly educate our clients to do things the right way. Go with this, go with Singular and not with the beep. There’s a lot of things like that. Sometimes they love like kids. They need to make mistakes by themselves to learn.
But some of them will listen and say, you know what you guys work with, you know, tens of games all over the industry or cross-platform. We do everything from mobile to PC web, Roblox, Minecraft games and web2. So we see everything all over. We have a really wide view of all the types of games and genres and stuff so some tools we know that fit better for different uses.
And if you use the right tools and build, like go the right path, then yes, then it’s easier at least than it used to be in the past because building companies and building products became easier, way easier and require less budget than it used to, if you look back 15, 20 years ago.
And yet, there’s so much out there, and it’s so noisy in the market. There’s so many games and apps being released every single day, even building something amazing. There’s no guarantee that it’ll make a dent in the universe, as Steve Jobs used to say. Anyways, this has been a lot of fun and super enjoyable. Thank you for joining us on Growth Masterminds.
John, that was great. Thank you, take care.