Podcast: The modern mobile gaming economy (with Phil Black)

On this week’s episode of the podcast, I am joined by Phil Black, one of the co-hosts of the Game Economist Cast, a podcast dedicated to game economy design, and a consultant with Game Economist Consulting. Previously, Phil held game economist and analytics roles at Amazon Games, DICE, and Scopely. Phil is also a panelist on the This Week in Games (TWIG) podcast.

Phil joins me on this episode to examine the shifting dynamics of the mobile gaming economy, from the consolidation of high-revenue genres to the strategic adoption of AI and direct-to-consumer models. Among other things, we discuss:

• How the concentration of mobile gaming revenue into puzzle and 4X genres reshapes the competitive landscape for independent developers
• Whether the integration of hyper-casual mechanics into core loops represents a fundamental evolution of the modern mobile gaming economy
• Why the transition from visual AI outputs to deep personalization remains the primary value proposition for future game monetization
• What the divergence between Android and iOS installation gates reveals about the health of the global mobile marketing ecosystem
• If the rise of third-party web shops and direct-to-consumer models can effectively counteract the platform fees of major stores
• How the emergence of warbonds and sampling-based monetization strategies signals a shift away from traditional battle pass reward structures
• When the mobile gaming industry will move beyond post-ATT recovery strategies

Thanks to the sponsors of this week’s episode of the Mobile Dev Memo podcast:

• ⁠⁠⁠⁠INCRMNTAL⁠⁠⁠. True attribution measures incrementality, always on.
• Xsolla⁠. With the Xsolla Web Shop, you can create a direct storefront, cut fees down to as low as 5%, and keep players engaged with bundles, rewards, and analytics.
• ⁠Branch⁠. Branch is an AI-powered MMP, connecting every paid, owned, and organic touchpoint so growth teams can see exactly where to put their dollars to bring users in the door and keep them coming back

Interested in sponsoring the Mobile Dev Memo podcast? Contact Mobile Dev Memo advertising.

The Mobile Dev Memo podcast is available on:

• YouTube
• Apple Podcasts
• Spotify

Transcript

Eric Seufert: Hello and welcome to the Mobile Dev Memo podcast. I am your host, Eric Seufert, and I am joined today by Phil Black. Phil, welcome to the podcast.

Phil Black: Happy to be here. Thank you for finally having me, Eric. We have traded podcasts now a couple of times.

ES: We have. I was on the Game Economistcast back in September or October. For the benefit of anyone who maybe does not know of you from the Game Economistcast or TWiG, please introduce yourself.

PB: I am Phil. I am a game economist. If you do not know what that means, neither do I. I am a consultant, or at least I have been for the last couple of years. I used to work at Scopely as an analytics games manager. I went over to DICE, which is how I moved to Europe and Sweden. I worked on Battlefront and Battlefield, as well as some other EA titles there. They had a lot of interest in game economies. Some of their games had very traditional game economies, like when you think about FIFA having a marketplace. There were a lot of questions they were interested in across a variety of products.

Then I worked at Amazon Games, spinning up their European operations, whatever that was. We could spend a lot of time talking about that. Now I do consulting, which is about one-third mobile, one-third PC, and then one-third other weird stuff, whether it be apps or theme parks. I think about strategy, economics, building models for a lot of these things, and a lot of experimentation.

ES: What are theme parks? Fill me in.

PB: If you were to go to Disney with your friends, one question you might have is why this is not a co-op adventure. It seems to have all the pieces you would need if someone could figure out how to connect it digitally or add that meta layer on top. How could we reimagine that as an economy?

ES: And that is a project you have undertaken?

PB: If you go to Dubai, there is one theme park where you will experience an Apex Legends banner system. There are stat trackers and different cosmetics you can get.

ES: How did you move into gaming and game economy design? I feel like a lot of people, when they want to work in gaming, say it is because they love to play video games. I usually tell them they are going to destroy the one joy they have in life if they work in gaming. Working in gaming takes a little bit of the veneer of mystery off. How did you make the jump?

PB: If you remember Yanis Varoufakis, the Greek finance minister who became the game economist for Valve. I say that loosely; he did two blog posts. He worked at Valve as a freelancer. Gabe Newell had read one of his books and asked him to come on and figure something out. He wrote two blog posts about Valve. One was whether Valve is a spontaneous order itself, and the other was whether you could solve for equilibrium with keys in the TF2 economy. That was pretty much it. He has been dining out on that contract for some time.

Respect as a consultant, by the way. Let him take the win. But that told me this is a thing you can do. Previously, I had just been in economics, a poly-sci and history nerd. I never really thought about doing games at all. I was a gamer, but I had not thought about combining these two things. Once I saw it was an opportunity, I started looking for anything written to validate that this is what I think it is. Your book was one of the things I read. I remember looking for someone speaking my language. When I started to get more encouragement that there are people doing this, I started spending more time playing these games. As an economist in this playground, you have this weird digital product which is a combination of art and science. You can run experiments on it. It feels like a fairy tale. You get addicted to it and enter through analytics, where there seems to be the most demand. Now there is more demand for the holistic understanding of what an economist does and what they can do for a company.

ES: You know what is weird is that my book is 12 years old, and there is not much else that has been written in that space. Russell Owens has an excellent book. I had him on the podcast, but I bought his book before I knew him. He wrote a fantastic book, but other than that, I cannot think of anything else that hits the analytical framing and scope from an economy design, analytics design, and instrumentation standpoint. I do not think anything else really exists. Mine might have been the first one covering freemium. It is strange because gaming is a massive industry. You would think there would be more interest in the underlying analytical infrastructure, but there just does not seem to be.

PB: One of the things I have ended up pivoting on is to ask not what games can do for economics, but what economics can do for games. We have it backwards. We have been trying the wrong quest. We share mutual friends in academia who study this field and are interested in what you can infer out of it. Everyone else who looks into it always ends up disappointed that it does not have all this external validity, which I am fine with because it is interesting in and of itself. It demands its own study and its own questions.

ES: If you went to find books on strategy for the film industry, there are probably a lot. It is very odd. It comes back to the question of why I wanted to get into games. Mobile gaming gives you access to a massive audience that emits an enormous volume of data. It gives you the opportunity to do scaled experimentation in an environment that is not possible in any other context. It brings together a bunch of disciplines in a market where you can make real money. In some ways, the actual job day-in and day-in is not that different, except that you are wearing shorts and there is a cardboard cutout of a character next to your desk.

PB: We say it combines art and science, and I think that is certainly true. When you look at the P&L, it is always somewhere in between. The fundamental economics sit between those different fields. But it has this commerce piece which I do not know if anyone has really respected. You can really express yourself in commerce through mobile games in a way that I think people underappreciate.

ES: I agree completely. I am in the middle of a series I am publishing called the Prosperous Society. That was the whole point of the second installment. Part of human expression is commerce. If you expect AI to fulfill that for you, you are taking away a big piece of human identity. I do not think that will materialize as a scaled application of AI. The process of buying things is a process of self-expression, and doing that in gaming gives you the infinite canvas. Funnily enough, I got the biggest ever royalty check for Freemium Economics last month.

PB: Wow.

ES: I think it is because a lot of what I was talking about there was speculative around the ability to do freemium with total personalization. At that point, it was a theoretical proposition because you could not build a content catalog that gets to that level of specificity. Now you can because you have generative AI. I think there is a new relevance for the freemium model. That is the promise of AI: personalization at the most fine-grained level. People get led astray with AI by focusing on the outputs and the visual representation, but the core value proposition is the ability to do user-level personalization at scale. That provides a much more fulfilling experience and allows you to monetize commensurate with that, which is the whole prospect of freemium. Freemium fails when people do not respect that and do not allow the monetization to actually match the consumer surplus. You give it all away and you do not actually match the economy to that.

I want to jump into the questions. You wrote a tweet on Friday that mobile gaming has evolved into essentially two genres: Puzzle and 4X. Walk me through how you got there.

PB: There are always three recurring actors in a mobile story. One is UA dynamics. The market has disciplined anyone who does not respect UA dynamics. You have East versus West, which is another paradigm in mobile that does not get respected enough. Then there is a process I call the Kuhnian genre evolution. There are these paradigm shifts inside of genres and crises these genres go through. They try to subsume one another. It is very close to biology models of game theory when different organisms interact in an ecosystem. This process has resulted in 43% of US iOS revenue, according to Sensor Tower, coming from 4X and Puzzle. That is remarkable. The other thing you want to pair it with is the recent revelation that Gossip Harbor, a game by a Chinese firm, has now outgrossed Candy Crush Saga. It has scaled to an enormous amount.

Thinking about where we are with UA dynamics, I would go back to one of the first pieces you published talking about where ATT actually reached majority iOS distribution. We look at what has happened since then to Puzzle and 4X versus other genres. These are some of the few places that are actually growing in mobile gaming. In 4X, they have figured out how to do a bunch of things. They have figured out how to absorb RPGs. They have figured out how to casualize the art style. They have also figured out how to build fake ad funnels. We called them fake, but they are integrated into the core loop now. It is part of an early onboarding experience. They have been able to scale to these massive levels built on the 4X game economy engine. It allows for almost unlimited LTVs. It is almost like an auction that happens inside of a 4X shard, and that has interesting dynamics when we think about bidding on UA. You have 4X driven by these Chinese firms doing all these things. They are killing us when it comes to live ops. They are making content at a rate that we cannot seem to keep up with. They are moving faster on AI and crushing it when it comes to ad creative volume.

Then you have what is happening with Merge. The expansion of Merge, this new game type, is actually combining a lot of old mechanics. This feels like an organic and natural Kuhnian genre evolution of things we have been playing with in other parts of puzzle subgenres. There has been a ton of experimentation in puzzle, and I would give credit to hyper-casual. They produce a lot of volume and are a great minor league farm for game mechanics. You might be able to trace some of the merge stuff back to that. That is a legitimate part of the market that is expanding. It is led by the Chinese and has been given up by the West, which feeds back into that East versus West story.

ES: Merge as a genre is really interesting. I was the interim CMO at Gram Games. I was living in London building my company, Agamemnon. I knew the founders and they asked if I could come in and be their CMO while they launched a game they had just bought. They had just bought Merge Dragons. Gram started as a hyper-casual studio. They were based in Turkey and their studio was like a lab. All Fridays were reserved for prototyping. They would make a prototype and then have a process for elevating the ones that got a certain number of votes into production. Merge Dragons was the first game that had a lot more of a core loop to it. The economy they were bringing to this game was unique. They did a lot of work professionalizing it from what they bought. That was 2016. What happened was the total collapse of hyper-casual, which then forced that UA discipline. UA discipline really is monetization discipline. It is building enough of a nexus of monetization to be able to deploy UA against it. That had enough scaffolding to support an economy. The floor came out on the casual side and both tails disintegrated. My point with the move to the middle was that both of those tails were going to disintegrate. A lot of hyper-casual game developers were very angry with me. People were writing articles saying I did not know what I was talking about. Well, look what happened.

Both tails essentially became unsustainable. The hyper-casual piece is entirely dependent on just buying new users every day. It reminds me of Lehman going into summer 2008; they were financing 30% of their expenses every day in the repo market. That is like hyper-casual. That just went away. The other tail, the very core piece like social casino, remains because they had players that were valuable and had retention. You can see it in Monopoly Go; you add or subtract that to casino and it changes the whole picture. You can look at Huuuge’s financials; they are printing more EBITDA than they ever have because they are not really investing in UA anymore. That is what had to happen. Do you think that is what supports Merge? Does it have that scaffolding to support the monetization while still providing this casual experience?

PB: It definitely has deeper monetization than Match 3. That is one thing it has to its benefit. One of the unique parts of a Match 3 economy is that engagement is always tied to monetization. The basic model of Match 3 is how many times you get to level ends, then how many times you convert on level ends, and what is the average price of that conversion. We have tried experimenting with the price, and it does not really do anything. Ultimately, we settle on how many times we can get you to level ends and what we can do with your conversion rate. Merge is a completely new economy paradigm that draws a lot from social casino because you have this energy multiplier and you can drain it down. They are able to compress time in a way that I do not think Match has been able to. That takes a lot of plays from social casino, maybe more from Coin Master, while combining it with a casual core. They got that casualization on the UA side, but I still think they were able to build a deeper economy. It is still early innings for the subgenre. It has not had its Royal Match moment. There is still going to be refinement to the core engine. It has not had its UA moment. You could have even more aggressive UA discoveries. The sky is the limit.

ES: Let’s hover there for a second. You talk about the energy mechanic being part of the economy. I think a lot of people outside of the space, but even actually within the space, think about a game economy as the price of gems or hard currency versus soft currency. Talk to me about why the energy mechanic actually is fundamentally the bedrock of the economy, or these mechanisms that moderate sessions and playtime and retention.

PB: One of the things you start to learn about in most, but not all, free-to-play economies is that ultimately you are selling time. There is always opportunity cost, which is your own labor that you could have used to get to whatever point in progression you want to get to. You can basically choose the allocation of labor and capital that you want to mix together to get to any point in a progression curve. Free-to-play games have always tried to figure out how we correlate price with whatever that point is. The thing you start learning about is that sometimes these monetization curves can be tied to engagement. The velocity of Match 3 limits how much spend you can really have per day based on how many levels you are playing. That is a fundamental fixed ceiling almost on how much money you can spend. If you go into social casino or even Monopoly Go, you can spend $300 in 30 minutes. They are able to compress time in a very different way that unlocks a lot on the monetization front. Power score economies like 4X, where combat is one of the core resolution mechanics and you just want the most attack, allow you to just sit there and basically speed through being able to get to the power score you want just by paying. That is a much more direct relationship than we have had in a lot of puzzle games beforehand.

ES: What is fundamentally different about Merge is that if you look at Candy Crush, the OG Match 3, when they went public I remember the stat because it was such a good rebuttal to the people that said freemium is corrosive and psychologically manipulative. It was something like 90% of the people who were on the highest level available in the game had never paid a cent. These are the people that are the most engaged and fanatical about the game. They are at the last level waiting for the update that week that gives them 15 more levels, and 90% of them have never paid a cent. That talks to you about that skewness of the monetization curve. The reality is you need to find a way to moderate the content consumption. If you just get on and eat all the levels, you can do it, and that is actually a very bad way to run a game. You need to actually control for that because ultimately the value that you need to trade for is retention.

PB: That is true, but here is what I point out. Match 3 has basically abandoned energy. This idea of the token arcade model where you insert 25 cents and go play a round has basically not become a binding constraint anymore. You are almost never coming up against it. It is very easy to get new lives because we figured out you can just make more money per level end. So let us just give you more level ends. What we will do is we will use difficulty as the rationing mechanic. That is one way we figure out how to ration people: if we just increase difficulty, it takes you more attempts per success. That is one way we can ration people, just through challenge.

ES: Another way is to give them access, but they are not rewarded for content consumption past some threshold. I give full credit to Supercell for that with Clash Royale and the chests. Your chests are full. You can keep playing, but you are not going to gain anything from playing.

PB: I am not the biggest fan of wage limits. It just sends a shiver down your spine. Maybe too much time in econ classes.

ES: Applied to a labor market, sure. But there is a secondary benefit which is just the satisfaction of playing. But that is the point because you could deplete that; that is a deteriorating effect.

PB: I look at progression as a wage.

ES: Then that is fair, and then it is a wage limit. But it is a time-based wage limit, so it is capped beyond any sort of realistic level. There are people that set a timer and they will come back to the game when the chests are openable. For most people, it is still capped beyond any sort of practical limit. In Match 3 now, you do level looping. You feed in early parts of the levels or you seed them in some way so you can basically recreate sequences. There is no such thing as end of content; you keep going. There have been clever solutions to this. We definitely do things like increase the marginal price per unit of power score, like in most of these games too. As you get to the end of the curve, it becomes increasingly prohibitively expensive to get to whatever the maxima point is in most games.

ES: I want to move on to downloads. Everyone saw a lot of articles going into 2024 about how mobile gaming is growing again, but it is a little deceptive because downloads have basically fallen off a cliff. Really, it is just these big players that have gotten better at monetization and they have these big franchises, but downloads are shrinking. Just because the market is growing does not mean there is a lot of opportunity because it is all being captured by the big companies. I think the issue is a little bit more nuanced than that.

I wrote about the mobile gaming have and have-nots like two years ago. If you looked at that point at the top-grossing games, there were new entrants like Scopely and Dream, with Royal Match and Monopoly Go. They were performing better than any other game. My point was that it is feast or famine, but a well-capitalized and experience-dense team can feast even if it is new. What I think people are maybe mistaking here in this analysis of the install number being down is that Android instituted a much stricter gate for new game publishing that is AI empowered. It is still a fully automated process, but it is much stricter. My sense is that you just cut a big chunk of the tail off for games that were not monetizing at all anyway. If you look at the differential between new app uploads on the App Store versus Android, I think that validates that. The App Store was up 84% year-on-year in Q1, and Android was not up at all. A lot of that AI stuff is not being caught in the filter on the App Store, but it is on Android. The App Store drives a lot fewer downloads on an absolute basis. My sense is that this new filter just excised a lot of installs that were essentially zero value. You are reducing the denominator, but you are not changing the numerator at all.

PB: That seems like a reasonable hypothesis. I can never get anything out of looking at downloads. I feel like there is always some sort of fraud or weird thing whenever I look at app store rankings. I always have to filter out hyper-casual. I have never found it to be trustworthy. How am I supposed to separate saturation from just a million different effects? And not only that, there are effects within each country too that I struggle with.

ES: That is my point. That is why just looking at these numbers broadly at a market level, people are drawing the wrong conclusion. They are drawing this conclusion that UA is impossible, it is all legacy games, and they are just getting better at live ops because they have reinvested money there. I do not think that is true. If you look at the top 10 games in the US App Store most downloaded, all but three were launched in 2024 or later. Tell me about a crisis in installs. Tell me about it being impossible to do UA. It seems like UA is maybe more accessible. That is actually a lot of turnover in the top downloaded chart. I wrote about this years and years ago; you looked at the top downloaded chart year-on-year and it was just the same seven studios. Now there is a lot of turnover.

It does not to me sound like it is impossible to launch a new game. It sounds like it is very difficult to launch a new game, and it takes a lot of density of expertise. You talk to anybody that works at Dream Games: what are they really good at? They are really good at monetization and they are really good at UA. Those two things work in concert. You can only be really good at UA if you are really good at monetization. I think what has happened is that a lot of the talent in the US market got absorbed into the bigger companies that have publishing regimes, but they also operate across platforms and mobile is not the primary focus. What happened in Europe was that the new studio formation just fell off a cliff because venture capital stopped investing in games. In Turkey, you have this density of talent, a lot of which came out of Peak Games. They went on to form numerous studios that are doing very well. Then China. In China, they have a lot of people that they can throw at this problem. To your point, they can produce thousands of creative variations every week. In a lot of cases, it is artists making the original variations and then they will obviously create variations from those with AI tools. That to me is what is happening. It is a totally new operating paradigm. It absolutely is possible to launch a new game into the top 10 downloaded this year if you wanted to, but you need this density of talent that just does not exist in the US. You need infrastructure that a lot of the big publishers in the US do not have, and you need the ability to adapt and to upgrade your infrastructure. In Europe, that was always driven by startups turning over, and by definition they had a lot of new and innovative approaches, but that is just not happening anymore. In Israel, it was mostly social casino. In Turkey and China, there is this explosive growth because there is more turnover and more willingness to update operating principles. They are adapting to the new operating norms. The idea that it is impossible to launch a new game because UA is dead is a misread of these numbers because it is a misread of what caused them.

PB: It is also interesting that every time a game scales in this category recently, it is always clouded by accusations of scaling unprofitably or negative ROAS. There is always some sort of air of suspicion around whether or not this is feasible. But don’t you think there is at least something to the idea of mobile gaming being closer to a perfectly competitive market? Things do not get bid down to the cost of production, but they move in that direction. There are very few moats, everyone copies from one another, and you cannot really hold on to any intellectual property. UA moves at the lightest speed. Do you think the United States has lost its comparative advantage in content?

ES: Yes, and I think it is worse than that. I think the United States has more or less lost its advantage in machine learning. That is a bigger problem. We do not produce nearly enough machine learning engineers, people that are fluent in that space. That is the modern industrial tech economy. That is it. It is actually very scary. But putting that aside, yes, I do think that these legacy publishers are operating in the Stone Age. I saw someone had a picture of a slide that they gave at some UA conference. I just sent it to a buddy of mine and said this could have been a photo from 2014. People are stuck in thinking from 12 or 13 years ago. It just has not evolved. You go and talk to people that work at Dream Games and they are operating on the frontier. You talk to these Chinese studios and they are operating on the frontier. Maybe they are not sharing. Anything that you push out in Facebook Ads Library is going to be copied mercilessly within a day. I had Matej Lancaric on a while back and he said you get a week with a winning creative, but that number is on the decline. In some cases, he was saying that it will not be a clone of your creative, it will be your creative. The creative will be the exact same thing; just the end card is swapped out. So it is not cloning, it is just stealing.

But that is not actually where the value is derived in UA. It is not at the output layer. It is in the underlying pLTV model, it is in the mechanism for doing signal engineering and passing back the most relevant and information-dense signal to the platform, and it is in cohort-level prediction. All this is ML now. That is where a lot of the studios in the US certainly have fallen behind. All this stuff gets copied and cloned, but that is not the actual power of the product. It is not the things that you can see; it is everything that empowers those being served to you in the game or at the first contact, which is the ad.

PB: Do you think the United States is behind in efficiency of buying UA traffic? If we made these optimizations that the Turks and the Chinese have, then we would be in a better place. If the Swedes had made this optimization for Candy Crush, they would have been able to act against Gossip Harbor. Is that what the West is missing?

ES: I do believe so. Western studios are not really doing scientific UA. A lot of my insight that is the input to that assessment comes from places like LinkedIn, which are biased toward slop. But my sense is yes, and this also comes from talking to a lot of people and doing selective consulting. I do think the approaches have not evolved since maybe that 2014 to 2017 era. That is a real operational and competitive disadvantage. I do think that explains a lot of the bifurcation that you are seeing in the market. But what you are not seeing is an inability to launch a new game to be a hit. You are just not seeing that. Look at the dates of launch for the top 10 most downloaded games in the US App Store.

ES: We are seeing mobile gaming’s M&A pick up again. Does this result in further genre concentration or could we see diversification as a result of this?

PB: Look, is it really picking up? It is always hard to tell; these are rare events. Outliers can really drive a lot of this analysis. I guess it is picking up, but it does not feel like valuations are up. It feels like it is more like it is residing. It is shrugging toward being back again.

ES: Well, we have had a number of big blockbuster deals. Moonton being the biggest, but that is like a Chinese sell-off. That is captured in the headline numbers. We have seen a lot more studio acquisitions. I just know this from my friends at Aream; they are posting a tombstone once a month. These are not billion-plus stuff that makes the Wall Street Journal, but still it seems like there is more of a cadence of M&A now.

PB: Pixel Federation, I think, was really exciting, but that game has underperformed since it was purchased. Moonton was selling from a position of extreme weakness, I would argue, rather than strength. Action RPG is way down; their portfolio was way down. I guess Dream is kind of interesting, but what does that acquisition do? I guess it returns money to people. That is interesting, but does it enable them to do anything that they could not have otherwise done? It just does not seem exciting. Scopely and Savvy Games Group seemed the most exciting. They have Monopoly Go; I thought that was really fun. But a lot of their other acquisitions have also gone by the wayside. Remember Tag Games? What happened to them? Digits has kind of fallen by the wayside. Pixel Federation does not look like it is going anywhere. There have been some huge misses. Miniclip, Easybrain. It just feels like the exits should be bigger here and that would also help valuations.

ES: Let’s say that it does pick up. Do you think that just results in more concentration or is that going to lead to category expansion and adjacency exploration? That was the outcome of the 2017-ish boom of M&A.

PB: I go back to the perfect competition model. If there are economic profits to be had, new firms enter the market and compete for those profits. This market responds very well to incentives and it responds pretty quickly. If companies get sold on one end, then there are going to be new companies coming in on the other end.

ES: What about AI-generated games? Are we seeing any traction there or any games being launched with AI that are worth talking about?

PB: Fuck no. Complete waste of time. No one is doing this. It is really disappointing. You wrote a piece a while ago about the stages of UA creative. I definitely see it there. I see it in my Instagram feed. You can see the AI ads. You can see the move toward this Pixar-style art, which had already been happening beforehand. Chinese and Turks are mastering this. But AI games have been disappointing. I think we have really abandoned the engineer founder. I think they are completely by the wayside. You look at the founders in games; they are not the original engineer. They are not the Tim Sweeneys who had some technical background. It is not the Carmacks of the world. It is not Will Wrights. I think we are losing our connection to the actual technology, which is the driving force in video games. We talked about art and commerce, but technology is the most important; it is the thing that makes it a video game. I think we have lost sight of that. We are not seeing it through the funnel. We are not seeing the ideas. Mobile studios do not take risks like this. This will happen in indie Steam. That is where this will start. We have seen so many trends in mobile gaming start in indie Steam. Hyper-casual, I would go even one step earlier on the farm system; it would be indie Steam. This is where so many hobbies games have come from. That is where the UA for Kingsland came from. It is an incredible farm league in and of itself, but it is not going to happen in mobile.

ES: What did you make of the Crimson Desert furor with the AI-generated placeholder art and people are up in arms about it because they did not disclose it?

PB: The Reddit agenda represents a certain voice on the internet and particularly in HD gaming. Does this voice matter? For mobile gaming, it never does. There was a subreddit I remember dedicated to The Walking Dead; it was just like “Hate Scopely Tuesdays” or something along those lines. This has always been the case; there is a loud furor against free-to-play games by some consumers. On Reddit, it is worse for PC console games. It is just not a large share of DAU. So they have aligned on an agenda, which is anti-AI assets. It won’t matter for mobile gaming and no one will talk about it. They do not want to make the issue more salient in PC console, which I think is a smart strategy. The same thing we played for monetization when there was loot box controversy: reduce the salience.

ES: I agree with you. Mobile gamers are not fanatical in that way. I think the impact of AI is invisible. There has been a tremendous impact on studio operations in terms of, certainly, the UA space. I actually think art creation is the lowest value part of that, but I have seen a lot of really interesting things being done that bridge monetization and player lifecycle management with AI.

PB: When does that hit a GDP number? I feel like I have seen magic, but when does that hit a top-line number that we can observe if it is so magical? I feel like this is where it should land, in some top-line number that AI has lifted. Why aren’t downloads back up if we got better at the discovery process through all this UA creative that AI can generate?

ES: But that is my point. First of all, I think it has; it has at the UA layer certainly. But again, keep in mind who is investing in this stuff. It is the big companies. It is not small studios that can do this right now. I do think there is a deck going around in some big company that shows we instituted this system and it drove 4% or 5% incremental conversions. That is important. You look at Meta’s earnings; they drove a 10% improvement in Instagram Reels engagement from an optimization to their recommendation system. Those are important boosts. That trickles down to every other aspect of monetization. You are asking when it is apparent in revenue numbers broadly? My point is I think it is. You won’t be able to parse apart the impacts at the individual company layer because no one is going to explain that to you. No one is going to tell you how they are using it or what the outcome is because right now, this is secret sauce stuff. This is not something that you explain in a GDC presentation or a blog post. This is very sensitive competitive intelligence or advantage. You wouldn’t just give it away. I do think studios are applying this to great effect now. When does it become so impactful that you see it in the top-line numbers? I think probably when we look back at 2026 revenue, we will see effects of that. But again, it is the biggest companies that can implement it. When does it become available to the small startups? It becomes available when the biggest companies make it available. That will be like, ask yourself, has AppLovin increased revenue broadly for mobile gaming? Yes, it has. So that is a GDP number, and that is AI. That is an impact of AI. I do think studios are applying this to great effect now. It won’t be SaaS tools. It will be the biggest companies making these available mostly at the UA layer but also maybe with production. Then it will be smaller studios that build this stuff as part of the game development process and they get gobbled up; they get acquired.

PB: If this is the case, and this is the same question we talk about with D2C, do all these gains go to the ad network or do some of them also go to the game provider? It is the same thing with D2C. Are all the D2C just getting bit away? Because once LTV is higher, then you just bid higher in the ad auction. So you have to think about what is the marginal propensity for the ad network when you get gains like this. Do you think it is 100%?

ES: Like 95%. But where does the value come from in the first place? It comes from UA. It comes from these platforms. If they did not exist, you would have zero. So thank them.

PB: I completely agree with you. I think what you want to do is you want to form a cartel to restrict supply and raise price. I think that is the only way out of this, but this is the mobile gaming nature. It is always going to be brutal.

ES: Sure, but you drive your revenue up, you drive the value of your studio up. The cash just gets recycled back into UA, but my studio is worth more because the top line is worth more. That has always been the case. That has always been the operating model. These margins are thin; it is just the volume is high when you account for UA spend. But I think that goes back to M&A and why private equity is interested in it. It goes back into what you see as mobile gaming as a piece of software or as a piece of entertainment. It has a very unique profile in economics.

ES: Talk to me about D2C and how that has impacted the ability to do the kind of analysis that we are talking about, looking at revenue and revenue per download. How is that changing things?

PB: It is always similar to what we talked about with ad networks. You never had ad revenue in Sensor Tower, so it always felt like a hole and you were undercounting hyper-casual. You were not really getting the full picture. Again, you still aren’t in many games. We could be off by 30% to 40%, which is why sometimes you have to look at downloads too because usually downloads are the result of that process you just talked about, that cycle. So if those are scaling or decreasing, usually it gives you a hint. So you have the same problem where you are not observing the data anywhere, but I can tell you it is very much real. Every almost single one of my clients has it and it is driven by games in which there is a socio-competitive benefit to optimizing on spend. This is why almost every 4X game has it: if I can get a cheaper price per unit of hard currency, that means I can actually progress faster than my clanmates. That is why you see this huge adoption of D2C in a lot of mid-core games while casual sometimes sits around 30%. You are seeing this war between some of these payment processors, different platforms, everyone competing for this space. I think margins are basically getting bid down to zero on D2C transactions, or they are really getting squeezed on how much a third-party provider can take versus you just doing it in-house. But that is certainly something real. Again, if it ends up increasing LTVs because you are taking more home, then if you are a studio, you are going to bid more on UA traffic and what does that mean at equilibrium? CPI goes up and the ad network gets to reap the gains.

ES: Across the companies you have worked with that are doing this, what is the largest percentage of revenue you have seen being generated from D2C?

PB: 50%. I would say the other thing you just have to keep in mind is that there are time series effects. The share of your DAU who are customers or people who have made a payment at any one point will increase just by survival bias because people who tend to make payments have longer retentive life cycles. They stack over time. Some games can have 80% of their DAU be payers. D2C revenue can also increase over time. I have personally observed as much as 60% in a 4X title.

ES: And how are they implementing that? Was that just moving the existing catalog into a web shop or were they applying some science there?

PB: Always incentives, always paying people. Science was not done in an effective way in that there was a controlled AB test where we know the causal uplift of all these effects versus people just naturally discovering it. But usually there is always some sort of VIP program straight out of the Eastern playbook from 10 years ago they were running. There is usually easy Apple Pay checkout. That is extremely important. Why do you keep hitting yourself? When you get to the checkout for many of these D2C platforms, you can use Apple Pay. You do not have to pay them. It is Apple Pay on the web, so you can basically guarantee that if you are on iOS, you are going to be able to solve this. The linking out is becoming far more aggressive. Clash Royale has a link out where you claim a free daily reward by going to the web store, which loads very quickly. Scopely has something similar with the Tycoon Club. You can get far more aggressive in getting people to the web stores than you ever could. Usually loyalty programs, discounts, sometimes a special SKU just to get you into the habit of getting into the web store. But they are definitely subsidizing it.

ES: Talk to me about those incentives because you still have to. This whole fake fear over Apple kicking out Cal AI, the calorie tracker app that got bumped by MyFitnessPal for 50 million and they got kicked out like two weeks in. Apple is overstepping and it is violating the Yvonne Rogers injunction. No, it wasn’t. They were doing all of this unscrupulous stuff at the checkout. Second, there are still rules. We haven’t been thrust into anarchy here as a result of the Epic trial. There are still rules. The reality is what actually, and David Barnard made this point on Twitter, the fact that Apple could not implement the pricing on the link out stuff was actually a benefit. Now that they can, they are going to start taking a cut of the linked out transactions. That whole idea that Apple is going to jail because they violated… I saw people on Twitter driven to hysterics over this. People at Apple are going to go to jail because they violated… No, they are not. This app was blatantly violating the rules. They were doing it in an in-app browser, which you cannot do if you link out. You have to offer the IAP. You cannot just have exclusively these alternative payments processes. How, if you are talking about they are getting more aggressive with the link out, where are people getting right up to the line there?

PB: I mean you can just straight up press one button and then boom, I am in the web store. My information is usually already saved or linked out beforehand and then I am in checkout in a couple of taps. It never used to be this aggressive beforehand. I was never made aware of the web store. It is not the original Project Liberty screenshot. If you remember Tim Sweeney and the original screenshot, it was the crossed-out price. That to me was always the vision he was fighting for, and I have not seen that. But what I have seen is just straight-up links to the web store inside of their traditional hard currency store in many games or where the hard currency SKUs sit.

ES: And have you seen people do any sort of content optimization there?

PB: CRMs still exist. You’ll still get emails that might push you to the web store again so they can get you outside of you just clicking the app icon or being in the app experience to get you to the web store. But that has pretty much been the extent of it. If you remember, you can use WhatsApp numbers too. AFK Journey used to do that for me to push me toward the web store. It sent me a WhatsApp text message and then it linked me into the Lilith store.

ES: The emails were the original tactic because that just totally sidestepped any sort of limitations in the app. You just send emails. That was actually what got allowed in the first place. Remember there was that class action lawsuit from small developers that had sued Apple in a class, but Apple had to allow apps to collect the email address early on in the onboarding and then they had to allow people to be able to communicate with users that way. That was how a lot of people managed that; they just did the web store via email discovery.

I have two more questions. Talk to me about GDC. I have not been to GDC; this is the second year I did not go. I do not know how well the LinkedIn chatter has represented the actual boots-on-the-ground experience. Tell me about GDC and what you perceive as the prospects of GDC going forward.

PB: Do you use it as a bellwether to evaluate the whole industry or just the mood?

ES: Mood, and I think people are more dour than they need to be. I think if you had a better representation of studios there, you would probably be more optimistic.

PB: I did not find that to be the case. Maybe it is survival bias and all the marginal people got removed so now it is only the excited people who are happy to be there. I found the mood to be high. I thought it was a great experience. It certainly shrunk. Sensor Tower reported it was down like 10,000. Don’t quote me on that, though. The hotel lobbies were not as packed as they normally are. I had a great time. I think everyone else who went usually has a great time. I didn’t think there were any dour moods or sour moods. I think there is always news about unemployment becoming a vicious cycle. The thing I always like to point out is that we do not know what hiring is like. I don’t think hiring is outpacing the closure of studios. I am certain there has been a net loss. We overestimated demand, and labor markets are going to correct. They take longer, they are sticky. This has been a slow process, but we don’t know what hiring is. It is not zero. Until that ends, that is going to be like water torture for developers. That was always in the mood, but besides that, AI was more affirmatively on the menu than I expected. More people are curious; they are AI curious.

ES: Talk to me about what you saw that you thought was inspiring or insightful related to AI.

PB: They had an AI session. They have these open sessions where everyone just kind of basically gathers in a room and then they throw a cube with a microphone around. It ends up being a pretty well-moderated discussion. I sat in on two days’ worth of that. Of course, you have people who self-selected into a conversation about AI. Almost all of it was super productive conversations. It was trying to think through whether there was any legal exposure, which remains on a lot of people’s minds. Do I have copyright over the assets that AI produces? Trying to figure out where you hand things off between a model and actually hiring someone to do the work. Integration with the 15 million different tools. But it was mostly about trying to unlock this as a productivity tool than anything else.

ES: No one is thinking about what the AI-first game designs are. How do you use AI to create new experiences that you could not otherwise have? This is the only way I see us growing our TAM at the moment. I do not think it is going to come through making a faster horse. Does TAM need to grow? Isn’t it already kind of saturated?

PB: Look, I think Roblox hit a fundamental limit when you look at how many kids they have. The first decrease in DAU in forever. They were sitting at one in four kids in the United States. I forget the age range, but it was half depending on the cut of kids. I always looked at that as a positive demographic signal for us because how many of those people can we retain going forward? Think about it like a lifecycle. This to me is us basically stacking highly interested gamers. If you just multiply this out, if we stack this wave of Roblox players, even if it decreases, that means total TAM will increase over time. I think, does it need to expand? No, but I’ll take anything that can help us win. I think this is one thing that is positive.

ES: Here is what I think we will get from AI. I do not think it is TAM expansion; I think it is engagement expansion. A lot of the analysis that points to sports betting and more participation in stock markets as taking away engagement from mobile gaming is a misread. The reality is those things are not mutually exclusive. These are bite-sized sessions. My sense is, like we talked about for 20 minutes, look at the top 10 downloaded chart. Seven out of 10 launched since 2024. But it is a lot of that, it is not that it is impossible to launch a hit new game, it is impossible for the people who do not know how to do it. We had this very disruptive exogenous shock with ATT. We are five years past that, and you are still hearing people talk about how to market in a post-ATT world. No, the companies that are successful learned that four years ago. If you are still asking that question or going to talks with that title, you lost the plot. You are so far behind it is going to be really difficult to catch up. I think it would just make a lot of sense to look at the companies that are doing that well and try to reverse engineer what they are doing. Roblox is down 20% after their earnings. They had a weak 2026 guide too. There is a lot riding on them. I think that also felt like it wasn’t the right bellwether for gaming. A lot of people repackaged their Metaverse enthusiasm into Roblox as a story. If that crumbles too, then maybe that whole project never had any merit, which I don’t think it did.

What is the most interesting thing you are seeing in economy design?

PB: I am going to be personally biased here. Warbonds, I think, are really interesting. If you have ever played Helldivers 2, it is their version of the battle pass and it solves a lot of fundamental problems with a battle pass, which I would argue was probably the biggest innovation in monetization design in what, a decade? Battle pass was basically popularized in Dota 2, then Fortnite takes it. Your traditional battle pass, every three months, you charge $10, usually there was a hundred tiers worth of content that you could upgrade as you go along the pass. You could choose at any time to pay $10 and get the premium lane and you would earn all the rewards that you had gotten up until the level that you had progressed to. You only had a certain amount of time to earn the rewards, usually three months. New battle pass would come. That ended up forming the delusion of many mid-price premium studios as what you could use as a revenue stream. People figured out it was more of an engagement driver than a monetization driver. How do you solve for this? Build a model. What is $10 divided by 90 days on a three-month cycle? That is your average daily monetization cap per day, and that is going to be far less than a dollar. So that is very low for a free-to-play game. You cannot support the LTVs you need.

The warbond system is an iteration, an evolution of trying to solve some of those problems, which is that we will stack battle passes. Battle pass was only available for players who would play in those three-month periods. If the Disney Vault is the revenue-maximizing strategy, would HBO make more money if they only had Game of Thrones available for 30 days in a month? Of course they wouldn’t. We think that is an absolute farce. Yet we believe this is the right strategy for content in games. We have started to move away from this. We start to have more permanent stores. Warbonds falls in this same mechanic where you can buy as many of you as you want of them; they stack over time. You could go into the game day one, maybe it has been a year since development, and you can buy four warbonds at once. So that is one piece, just stacking the warbond so you do not have to be pressured into them at any one time. Then you also have the ability to choose which items you want in a warbond. So rather than this lane we were talking about, you choose from a page, and then when you have reached a certain amount within a page, you can flip to the next page. So you have some level of autonomy. This to me has been a huge innovation in monetization design for a lot of HD games. It solves a lot of problems. You can stack content, you can stack monetization. I think it is a big winner and I think you are seeing more games do it.

ES: Anything else that is more mobile-specific?

PB: Absolutely we have seen more event box design. This really comes out of the East, where you basically have a fixed pool of items and then what is happening is you are sampling without replacement. You are going in, you are removing an item from the pool, you are getting rewarded, and then the drop weights or the drop rates are getting recalculated now that it has been removed from the pool. Sometimes what the Eastern games will do is they will increase price proportionally with that, so the game also becomes more expensive to play. You are seeing a lot more Eastern RNG or gambling design come into live ops and those games be popular in the West.

ES: Phil, this was great. Tell people where they can interact with you online, where they can consume your insights online.

PB: I am on LinkedIn, unfortunately or fortunately. I have my own blog, gameeconomistconsulting.com/blog, where I subscribe. I’m also on the Game Economistcast, which is my own podcast, and then there is also This Week in Games on Deconstructor of Fun.

ES: Thank you so much.

PB: Thank you for having me.