If you build 80% of the right game, you’re going to get like $10,000 in sales. If you build 100% of the right thing, you might get $100 million in sales.

That is Ran Mo — CEO of Proxima, creator of Suck Up!, and arguably the person with the most credibility to be an AI maximalist in gaming right now. Suck Up! was the first commercially successful game to use AI at runtime. It generated over 100 million YouTube views with zero marketing spend. Before founding Proxima in 2021, Ran led product teams at EA working on The Sims franchise and spent time at YouTube and BCG.

He has every reason to be all-in on the AI hype. He is not. And his restraint is the most instructive signal in the current landscape.

I should be transparent: I am an AI maximalist. I spent weeks sleeping two to four hours a night learning vibe coding. I believe development costs will drop 10x by the end of this year and that 2026 may be the most consequential year for AI's impact on our industry. Ran and I disagree on pace and magnitude — but the conversation we had sharpened my thinking on AI's impact on product velocity: the difference between speed and direction. AI gives you speed. It does not give you direction. And in a power-law market, direction is everything.

Here is the thesis of this piece: AI tools deliver real, measurable productivity gains on specific coding tasks. Those gains will not, in isolation, translate into more successful games. They will translate into more games — most of which will fail — while the winners pull even further ahead. The bottleneck in games was never production. It was always taste, direction, and distribution.

The Trunk and the Leaves

The most useful mental model for AI deployment in game development is actually quite simple: a tree.

The trunk is your core architecture — the game loop, the netcode, the core systems that everything else depends on. The leaves are discrete, isolated features — a grab animation with quaternion-based rotation, a UI element, a shader variant, a localization pipeline.

AI is exceptional at leaves. Ran demonstrated this during our conversation by sharing his screen — a quaternion-based object rotation system with sine-curve acceleration and slerp interpolation, built live in Unity with Claude plugged into a Rider IDE terminal. A senior engineer builds it in a day. A junior engineer needs two weeks just to learn what a quaternion is. Claude built it in under 30 minutes. That is a genuine 30x speed gain on a specific, well-scoped task.

But the trunk? Ran tested this himself — full vibe-coded Unity projects produce functional prototypes that can be architecturally catastrophic. His analogy: it's like using MS Paint to draw a square by painting every individual pixel instead of using the square tool. The result looks identical. The underlying structure is unusable the moment you need to scale. The code compiles. It runs. And then it becomes unmaintainable spaghetti when you need to iterate on systems that touch everything. "You're going to burn a million tokens and get nowhere," as Ran put it.

This reflects a current-day structural reality: AI tools are built for local optimization, and trunk-level architecture is a systemic design problem. Every tech leader I talk to is fighting this battle internally. The pressure to show AI-driven productivity gains is immense. But if you let AI-generated code infect your architectural core, you are building on sand.

At least today, building production-quality code for a complex game remains an extremely high-skill endeavor that few people in our industry can handle.

So what does Ran's actual productivity gain look like? His answer surprised me, given my maximalist leanings. On leaf-level tasks — the grab animation, the UI polish, the boilerplate — he sees 5-10x or even 30x gains. But on trunk work — re-architecting a codebase, refactoring systems — using AI can actually slow you down, because the time spent injecting context, supervising output, and auditing for silent breakage can exceed the time it would take a senior engineer to just do it themselves. The overall gain, across all work, is real but more muted than the X headlines suggest.

The strategic move is ruthless segmentation. AI on the periphery. Senior engineers on the spine.

The $100 Million Gap Between 80% and 100%

Here is where the conversation gets uncomfortable for the "AI will democratize game development" crowd.

Games are not houses. If you build 80% of a house, you can sell it for 80% of the price. Games operate on an exponential attention curve where "almost right" is commercially indistinguishable from wrong. The last 20% — the part that is pure creative judgment, directional conviction, finding the right mechanic, the right feel — is where all the value concentrates.

AI makes the first 80% cheaper and faster. It does nothing for the last 20%. Which means AI tools are making it cheaper to produce games that fail.

The data confirms the power law. The global games market generates approximately $184 billion annually. Fewer than 5% of Steam titles produce meaningful commercial returns. AI tools will dramatically increase the number of games uploaded to Steam and mobile app stores. The denominator explodes. The numerator — total player attention — stays fixed.

Ran drew the parallel to YouTube, where he worked before entering games. The barriers to entry for video content collapsed years ago. Anyone can record themselves eating breakfast and upload it tomorrow. What happened? Extreme power law distribution. The top creators captured the vast majority of attention. The long tail got nothing.

Games are heading to the same place. You must be premier television or Mr. Beast. There may not be a viable middle.

The Technically Capable Designer: Who Actually Wins

This brings us to a question most AI-and-games commentary skips entirely: if AI compresses the coding bottleneck, who is the ideal game maker in this new environment?

Ran's answer is specific: the technically capable designer. Not a 10x engineer. Not a business founder who hires coders. A designer who understands code architecture well enough to direct AI tools, read what they produce, and catch when they're building the wrong thing — but whose primary skill is creative judgment and taste.

"The limiting factor for games is not code, it's the creative output of the design," Ran said. "You can have a designer who understands code, but not very good — he would need two engineers to realize his vision two years ago. Today he might not need anyone. But he has to be somewhat technically capable."

This is the new archetype. Not the vibe coder who prompts their way to a prototype and calls it a product. Not the senior engineer who uses AI to ship 3x faster but has no design instinct. The person who sits at the intersection — who can evaluate a quaternion implementation and knows whether the mechanic it enables is worth building in the first place.

Design as Differentiator, Not Speed

Ran Mo does not use AI for art. He does not use AI for design. He does not use AI for concept work. Not even for brainstorming.

This is not Luddism from someone who doesn't understand the tools. This is a strategic decision from the founder of an AI-native game. His reasoning: in a world of AI-generated abundance, the human element of your creative output is the only differentiator. AI creative work is average by definition — it is a statistical composite of its training data. When everyone has access to the same generative tools, average becomes invisible.

"You can have the most ugly chess set and it's still chess," Ran told me. "But if you have this beautiful tile thing that isn't chess, it doesn't matter."

It’s a sharp rebuttal to the idea that AI-generated environments and world models will replace game design. The mechanics are load-bearing. A stunning procedurally generated world with no game loop is a tech demo, not a product.

And here is where Ran offered what I think is the most contrarian and underappreciated insight of our conversation — one that carries particular weight because he built the first successful AI-runtime game.

The Probabilistic Gameplay Problem

When Google announced world model capabilities, Unity stock dropped approximately 10%. The market panicked. Several VCs Ran spoke with were convinced this would change everything.

Ran's response: they're solving the wrong problem.

"People, especially people outside of games, misunderstand what games are," he said. "Games are mechanics wrapped in logic. What's really important is the mechanic. And when you build AI systems — whether it's world models or AI NPCs — they're probabilistic. If people can't predict how the probability functions, it's very difficult for them to decide what to do next."

He used a chess analogy that crystallized it for me: imagine playing chess, but the knight might do something different on the next move — you just don't know what. Would you play that game? Probably not. The reason games work is because players can form mental models of the rule system and make strategic decisions within it. Probabilistic AI breaks that contract.

This matters because Ran has more standing than almost anyone to say otherwise. Suck Up works because the comedy context tolerates unpredictability — the AI NPCs are a comedy stage, not a strategic system. But he's clear-eyed that this is the exception, not the template. "We haven't seen too many games out of Suck Up who have really gotten the stuff beyond," he acknowledged. "There's probably less there than people think there is."

The implication for studios chasing AI-in-gameplay as a differentiator: make sure you're building for a context that benefits from probabilistic behavior, not one that requires deterministic rules. Most game genres require the latter.

The Velocity Trap

This is where Ran's philosophy diverges most sharply from the prevailing Silicon Valley consensus — and from my own instincts, if I'm being honest.

Ran is deliberately lengthening his iteration cycles. Not because he is lazy. Because he learned the hard way — nearly burning out himself and his team after Suck Up — that micro-level velocity often masks macro-level misdirection.

"We would get a playtest and figure out the changes and send feedback by 8 PM, and by 8 AM the next morning we'd have the new build," he told me. "The iteration cycle was crazy fast. But then you step back after three months and you're like — we went down entirely the wrong direction."

Now he forces himself to sleep on a build for an extra day before committing to a direction. He described it through the lens of a Taoist philosophy called wu wei — the art of deliberate non-action. It sounds like mysticism until you map it to product development: in an environment where everyone is sprinting, the competitive advantage belongs to the team that pauses long enough to confirm they're sprinting toward the right destination.

"A lot of people are just kind of pedaling to go nowhere," Ran said. "And it's better just to not pedal at all in some cases."

I pushed back on this in our conversation — I think speed matters enormously right now, and the teams that figure out AI-augmented development first will have a structural advantage. But Ran’s experience reinforced something I’ve already written about: product velocity is a vector — magnitude and direction. A team moving slowly in the right direction will outperform a team moving at blazing speed in the wrong direction. AI gives you magnitude. Taste gives you direction.

Distribution Is the New Production

If creative taste is the moat, distribution is the drawbridge. And most studios still treat it like an afterthought.

Suck Up's distribution story is now well-known, but the mechanics of it deserve closer examination. "We had $0 in marketing because we were poor," Ran said, "but we had one guy who had 100,000 subscribers on YouTube play it and he got a million views in a day."

Within a week, 30 creators were playing it. Within a month, top creators with tens of millions of followers had picked it up. The game generated over 100 million YouTube views organically.

This was not luck. It was architecture. Suck Up was designed to produce entertaining content when played. The AI NPCs created a comedy stage for content creators — a canvas for drama and improvisation. The game's core loop was inherently watchable and shareable. Distribution was a design constraint, not a marketing function.

And then Proxima made a hire that signaled exactly where the industry is going: a head of marketing with approximately one million TikTok followers. Not someone from a brand marketing background at a major publisher. Someone who had built an audience themselves.

"The biggest indicator if they understand marketing today is can they build an audience themselves," Ran said. "Because until you build an audience yourself, you don't really know how to build an audience."

This is the new competency test for marketing hires in the games industry. Can you build an audience from zero? Not manage a UA budget. Not optimize CPIs. Actually create content that earns organic attention. If the answer is no, you are hiring for the old model.

The Proxima playbook — embedding distribution-native thinkers into the development cycle from day one, not bolting on marketing six months before launch — might just be the new meta. When production costs deflate, the capital you save must be redeployed toward creative risk and distribution capability. Studios that pocket the savings and ship the same game faster are simply accelerating their path to irrelevance.

The Contrarian Case — and Where It Falls Short

I want to address the strongest counterargument directly, because it is not wrong. It is just incomplete.

The argument goes like this: incumbents also get the cheap tools. A 40-person studio using AI to operate like an 80-person studio does not get squeezed — it gets leverage. It keeps its institutional advantages: project management maturity, existing player communities, platform relationships, live-ops experience. The desktop publishing revolution of the late 1980s was supposed to kill mid-tier design agencies. Instead, the agencies adapted, shed staff, and became more profitable.

This is fair. And it will be true for some mid-tier studios — the ones that restructure ruthlessly, redeploy savings into distribution and creative risk, and have the discipline to kill projects that are not working. Supercell has operated this way for years: small cells with institutional infrastructure, shared technology, and a marketing machine behind them. The model works.

But here is where the analogy breaks down. Desktop publishing produced bad newsletters. AI-enabled game development will produce playable games that compete for the same finite shelf space on Steam and the App Store. The flood is not just noise — it is functional noise. When Steam's discovery algorithms are overwhelmed by AI-generated content, they will favor quality signals: review scores, playtime depth, refund rates, developer track record. Those signals favor established studios. They also favor small teams that have found genuine product-market fit and earned organic attention. The studios they crush are the ones in the middle — too large to be nimble, too small to compete on spectacle, shipping derivative products into an ocean of content.

The Bottom Line

Every era of cost deflation in creative industries produces the same pattern. Production becomes cheap. The market floods. Attention concentrates. The winners are not the fastest producers. They are the ones with the clearest creative vision and the strongest distribution instinct.

AI does not change this pattern. It accelerates it.

The studios that survive the near future will be built around a specific kind of person and a specific kind of discipline: technically capable designers with strong creative taste, distribution thinking embedded from day one, and the wisdom to slow down when everyone else is sprinting toward the wrong destination. They will use AI ruthlessly for the leaves and protect the trunk with senior engineering judgment. They will hire marketers who have built audiences, not managed budgets. They will resist the pressure to ship fast and instead ship right.

Ran Mo nearly ran out of cash three months before Suck Up launched. One of his VCs told him they backed Proxima because the team "clung on and didn't die." That is not a glamorous endorsement. But it captures something essential about what wins in a power-law market: the conviction to stay on the right path even when the pressure to sprint in any direction is overwhelming.

Ran's parting advice was drawn from the Taoist concept of wu wei: "Sometimes it's important to slow down to speed up. Not everyone needs to be pedaling all the time."

I'm still an AI maximalist. But I left this conversation thinking less about how fast I can build and more about whether I'm building the right thing. That might be the most valuable shift the AI x gaming discourse could use right now.

The gaming industry spent 2024 and 2025 obsessing over which jobs AI would eliminate. But the more interesting question — the one with actual strategic implications — is what AI can’t replace. Solomon Ruiz-Lichter figured out the answer three years ago, before the current AI frenzy even started: curated human relationships in a high-trust community.

Solomon is the founder of the Product LiveOps and Growth Symposium, now entering its third year during GDC. What began as a vendor-side experiment has become one of the most sought-after events during GDC week — and its growth trajectory holds lessons that go far beyond event planning.

The core insight: In an AI supercycle where content costs approach zero and noise becomes infinite, the ability to build trusted, curated communities around specific professional audiences is becoming the single most durable competitive advantage in the games business — for operators, service providers, and studios alike.

🎧 Listen on Spotify, Apple Podcasts, or Anchor

Speakers:

• Joseph Kim, CEO at Lila Games

• Solomon Ruiz-Lichter, Founder at Dual Pulse Advisory

The 10% Gap That Created an Entire Community

The Symposium didn’t start with a budget or a venue. It started with a hypothesis.

Solomon and his co-founder Tim Hong (then at AWS) analyzed the GDC content lineup and found that product management, LiveOps, and growth content accounted for roughly 10% of sessions. This for an operational discipline that directly drives the revenue of every live-service game on the market.

That gap between audience need and content supply created the opening. “The hypothesis was that product leadership, product management, and particularly LiveOps was an underserved audience at not just GDC, but events in general,” Solomon explained. “There just wasn’t a lot of content around it.”

The validation process was straightforward: map the existing content landscape, identify the underserved audience, and build a programming concept around the gap. The first year’s concept — how AI would change games — proved prescient. At the inaugural 2024 Symposium, Travis Oatman from Carbonated was already discussing AI-automated localization, a technique that major publishers only rolled out across their enterprises in 2026.

The lesson for anyone considering building a community event: start with a falsifiable hypothesis about an underserved audience, not with a sponsorship budget.

Year 3: What Breaks and What Holds

This year tested whether the Symposium had built something real or something fragile.

Three things went wrong simultaneously. Tim Hong left AWS mid-planning cycle, creating uncertainty about the anchor sponsor relationship. GDC attendance dropped — particularly from European attendees. And new competitors emerged, running events clearly inspired by the Symposium’s format.

Here’s what happened: the event came together in roughly 21 days of serious execution. Registrations came in strong. Speakers committed quickly. The brand held.

“People were hitting me up about a month out, asking if we were doing the Symposium this year,” Solomon said. “That’s the brand we’ve built.”

The 2026 event expanded to two days for the first time, anchored by prime sponsors Xsolla, AWS, and AppsFlyer, with Travis Oatman delivering the keynote and a full programming slate covering both strategic vision (Day 1) and tactical, actionable content for studio operators (Day 2).

What made this resilience possible wasn’t budget. It was the community’s expectation that the event would exist and deliver value. That expectation — earned over three years of consistently prioritized content over sales pitches — is the actual asset.

The “Don’t Sell” Discipline That Builds Trust

Solomon’s first piece of advice for anyone building a community event is counterintuitive for anyone on the service provider side: don’t sell.

“Everyone says, ‘oh yeah, I know, don’t sell.’ But no — you really have to be aware of it. You have to be very purposeful about it,” he emphasized. The Symposium team enforces this rigorously. When an abstract feels even slightly promotional, someone flags it. The content curation process filters for genuine operator value, not vendor messaging.

This isn’t naive altruism. It’s strategic patience. The buying consideration cycle in enterprise game tech is long. By giving genuine value — connecting operators with actionable insights they can take back to their studios — the Symposium builds the kind of trust that eventually converts to business relationships. But the conversion happens on the community’s timeline, not the vendor’s.

The second discipline: take the time to learn the space deeply before you try to build in it. “If you’re a game tech provider and you think, ‘how hard could it be?’ — hard,” Solomon said. “The gaming industry really resonates with people who took the time to learn the space. You need to earn the right to be out doing things in the market.”

This maps directly to a pattern I’ve observed repeatedly in mobile gaming: the service providers who build lasting businesses are the ones who can speak credibly about product metrics, LiveOps calendars, and monetization design — not just their own tool’s feature set.

Why Community Value Is About to Spike

Here’s where this conversation shifts from event-planning advice to strategic insight.

Investor Chris Camillo — who built a $300 million portfolio from a $20,000 start — recently described the current AI moment as a “supercycle.” His framework for what survives disruption centers on moats. And his top moat? Relationships and trust.

This aligns with a thesis I’ve been developing: as AI drives the cost of content creation toward zero, the signal-to-noise ratio in every professional community degrades. Generic content becomes infinite. Targeted, curated, high-trust content delivered through human relationships becomes scarce.

The math is simple. When anyone can generate a 2,000-word industry analysis in 30 seconds, the analysis itself loses value. What gains value is the curation layer — knowing which analysis matters, who is credible, and which insights are actionable for your specific context. That’s what a well-built community provides.

The Symposium illustrates this at a practical level. Travis Oatman’s keynote isn’t valuable because it’s content about AI in gaming — there are thousands of those pieces. It’s valuable because Travis has been building production AI tooling at Carbonated for three years, he’s far enough ahead that you could spin out three tech companies from what he’s built, and the Symposium audience knows this context because they’ve been tracking his work since the first event.

That depth of context — speaker credibility validated over time within a trusted community — is something no AI-generated content feed can replicate.

The Window Is Open: Why You Should Consider Starting Your Own Event

If the value of curated communities is inflecting upward, the logical next question is: should you build one?

I’ve thrown GDC dinner events for five years running. I paid for a GDC pass for nearly a decade and averaged 1.5 sessions per year before finally stopping. The formal conference content wasn’t where the value was — the value was in the rooms I curated myself. And I’m not alone. The entire GDC week has become a constellation of external events: content symposiums, private dinners, happy hours, and yes, those nightclub events where the music is deafening and nobody can actually talk (skip those).

Here’s my argument for why right now is the moment to start: the attention economy is about to get dramatically worse. AI-generated slop is flooding every digital channel. The cost of producing generic content is approaching zero, which means your audience’s inbox, feed, and search results are about to become significantly noisier. Against that backdrop, the ability to curate a room of 50-200 professionals around a specific, underserved topic isn’t just valuable — it’s becoming rare.

The Symposium’s playbook offers a practical template. Find a connector — someone with a platform and genuine relationships in the space. Solomon initially leveraged my GameMakers audience to accelerate the Symposium’s early traction; today, LinkedIn is full of people in mobile gaming with substantial followings who could serve the same role. Validate a content gap with actual data, the way Solomon and Tim mapped GDC’s session breakdown and found the 10% LiveOps gap. Program ruthlessly around operator value, not sponsor messaging. And start small — Solomon’s first proof of concept was a dinner in Seattle with a featured speaker and a handful of the right people, not a 400-person two-day conference.

One thing Solomon mentioned that I want to underscore: don’t gatekeep. The Symposium deliberately reserves spots for people between gigs, students, and aspiring operators — not just senior executives at potential client studios. That inclusivity isn’t charity. It’s how you build a community with real depth and long-term loyalty, rather than a transactional networking event that dies the moment the sponsor budget disappears.

The window for building these community assets is open right now, while the value of in-person curation is rising and before the landscape gets crowded. Solomon kicked open the doors, and now others are following. That’s the signal, not the noise.

The Renaissance Model for the AI Age

The parallel that keeps coming back to me is the Renaissance workshop model. The artists who produced the greatest work weren’t isolated geniuses. They clustered in specific cities, shared techniques within tight communities, and collectively raised the bar for everyone in their orbit.

In the early days of mobile gaming, the same dynamic played out. The studios that understood incentivized installs on Tapjoy before their competitors had a massive edge. Game of War beat Kingdoms of Camelot not through superior creative vision, but through an information advantage — they studied Chinese monetization techniques (VIP systems, gold banks) and implemented them first, pushing LTV from $8-$10 to $12-$15 and sweeping the market.

That information advantage came from being in the right rooms with the right people. The same dynamic is playing out now with AI. The studios and operators attending events like the Symposium aren’t just networking — they’re building the information networks that will determine who adapts fastest to the AI supercycle and who gets left behind.

Game development itself may be AI-resistant at the creative level. But the competitive landscape around it — who ships faster, who monetizes smarter, who adopts the right tools first — will be won by the people with the best information networks and the highest-trust professional relationships.

Takeaways

• Community is a moat, not a marketing channel. The Symposium survived its co-founder leaving the anchor sponsor company, a GDC attendance drop, and new competitors — because the community itself expected and demanded it continue. Build community assets that transcend any single company or sponsor.

• Start with an underserved audience, not a budget. The 10% content gap for product and LiveOps at GDC created an opening that no amount of marketing spend could have manufactured. Map the content landscape for your target audience and build where supply is genuinely thin.

• The “don’t sell” discipline is a strategic choice, not a moral one. In long buying-cycle industries like game tech, trust built through genuine value delivery compounds over years. Every promotional session you cut from your event is an investment in that compound return.

• Information advantage matters more during transitions. Just as early mobile operators won by being first to adopt incentivized installs or Chinese monetization mechanics, the operators who are first to adopt practical AI workflows — learned through trusted community channels — will have disproportionate advantages in the current supercycle.

• The window to build community events is open now. The value of in-person curation is rising while the cost of digital content drops to zero. You don’t need a massive budget — you need an underserved audience, a connector with reach, and the discipline to program for operator value over sponsor messaging. Start with a dinner. See what happens.

• Curated human networks become more valuable as AI content becomes cheaper. When the cost of generating generic analysis drops to zero, the premium shifts to knowing which insights are credible and actionable. Invest in the communities that provide that curation layer.

I told you last week that the gaming industry is losing its moat. Now here’s the data.

The GDC 2026 State of the Industry report surveyed over 2,300 game professionals. It’s the most comprehensive snapshot we get of what’s actually happening inside this industry. The report presents data without a thesis — that’s its job.

My job is different. I’ve been developing an argument about where this industry is headed, and the GDC data confirms it almost point by point.

Here’s the argument: the gaming industry’s historical moats — capital, talent, technology — are all collapsing at the same time. Not slowly. Not theoretically. Right now.

Three moats. Let’s go.

1. The Money Moat

For decades, you needed money to make games. Real money. Publisher advances. Venture rounds. Platform deals. The development cycle was long, expensive, and required capital just to get in the room.

That’s ending. Not slowly. Not theoretically. But it’s not over yet — and the studios that recognize they’re in the middle of this shift, not past it, are the ones with the advantage.

35% of developers say self-funding is their primary source of funding. For solo devs, it’s 86%. Publishing deals? 20%. VC and private investment? 5% each. Crowdfunding, seed rounds, bank loans, accelerators — all at 1% or below. Platform-based funding sits at 2%.

Read those numbers again. The entire infrastructure of game financing — publishers, platform deals, VC — accounts for less than half of how games get funded. Most developers are funding themselves.

And it’s not by choice. The money isn’t just expensive — it’s gone.

“We attempted to secure venture capital... and got 0 bites and a lot of: ‘Wow, if you had pitched this 2 years ago, you’d already be funded.’”

— Director (Game Design), California

“There’s basically no funding at our scale: we’re too expensive for indie, too small for AA/AAA. Award-winning studio, ‘Triple I,’ prestige and history, but there’s a gap.”

— Senior Employee (Game Programming), Virginia

“Platform deals have dried up (Game Pass, PS+ etc.), which used to be a great way to finance indie games.”

— C-Level (Independent Studio 5+ Years), California

The “Triple-I” squeeze is real. Studios too big to bootstrap but too small for AAA publishing deals are getting crushed. The funding infrastructure wasn’t designed for a world where the cost of building dropped faster than everything else.

But here’s the thing. If building games is cheap, why does anyone need money at all?

» Distribution. «

When the GDC report asked what matters most when choosing a platform, audience reach was #1 at 78%. Discoverability: 43%. Top marketing methods? Social media (65%), streamers (39%), paid advertising (31%).

The money you used to spend building the game? You’re about to spend it making anyone notice it exists. I wrote about this last week — when anyone can build an app, and nobody has patience, you get a billion apps and seconds to prove yours matters.

The moat didn’t vanish. It moved.

2. The People Moat

You used to need a big team. Dozens of artists, engineers, designers, QA testers, producers, and PMs to coordinate them all. Scale was the price of entry.

Not anymore.

77% of newer indie studios have fewer than 20 employees. 11% of respondents at newer studios work entirely alone. Across the full survey, 8% identify as solo developers — and they’re shipping real products, not hobby projects.

I wrote about this in September — The Great Western Game Development Divergence.

The breakout hits on Steam and PC aren’t coming from 200-person studios. They’re coming from teams of 1 to 20. Lethal Company. Schedule 1. Project Zomboid. The GDC data confirms that pattern at scale.

Meanwhile, the big teams are bleeding out.

28% of respondents have been laid off in the past two years. In the US, it’s 33% — one in three. And it’s getting worse: 17% laid off in the past 12 months, up from 11% the year before. For AAA specifically, 19% personally laid off in the past year — more than double the prior period.

Of those laid off, 48% still haven’t found another job. For people laid off 1-2 years ago, 36% are still looking for work. The “course correction” everyone hoped for didn’t arrive. The jobs aren’t coming back.

The companies say it’s “restructuring” (43%), “budget cuts” (38%), and “market conditions” (38%). The people who got fired have a different take:

“Executives who have never actually worked as a dev are pulling up the boards on their ships, throwing people overboard, and expecting these scuttled ghost ships to keep making them infinite money.”

— Former Senior Employee (Game Design), New York

And the pipeline is breaking too. 74% of game dev students are worried about their future job prospects. 87% of educators expect a negative impact on placement.

“Most of my students will not have a career in game development.”*

— Educator (Computer Science), Michigan

“There aren’t any jobs. Everyone’s getting fired. It’s fucked.”

— Student (Game Design), California

The people moat used to mean you needed a hundred skilled professionals to ship a competitive game. Now you need five. Maybe one. The human cost of that transition is enormous — but the structural reality is clear.

3. The Technology Moat

This is the one nobody wants to talk about.

For decades, proprietary technology was one of the hardest moats to cross. Custom engines, institutional knowledge, and toolchains built over multiple product cycles. id Tech, Frostbite, Rockstar’s RAGE — these weren’t just tools, they were competitive weapons.

That era is ending.

Unreal Engine now sits at 42%, surpassing Unity (30%) for the first time. Godot is gaining traction with newer indie studios at 11%. Proprietary engines have dropped to 19%, with almost all of that in AAA. The tools to build a competitive game are available to anyone with a laptop.

But the real disruption isn’t which engine you use. It’s what’s happening to the people using them.

52% of game workers think AI is having a negative impact on the industry — up from 30% last year and 18% the year before. Visual artists: 64% negative. Designers: 63%. Programmers: 59%. And executives? 19% say it’s positive.

The open-ended responses capture what the percentages can’t:

“AI is theft. I have to use it, otherwise I’m gonna get fired.”

— Senior Employee (Visual and Technical Arts), Ukraine

Read that again. A person who believes the technology is morally wrong uses it anyway because the alternative is unemployment. That’s not adoption. That’s capitulation.

And it’s everywhere:

“I guess I’m hypocritical about it. I’m outspokenly against AI, but then do use ChatGPT as a sounding board for some of my marketing ideas.”*

— Manager (Marketing/PR), Canada

“I don’t tell anyone how I use AI, for fear of backlash.”

— Contributor (Game Design), California

Meanwhile, on the other end:

“We are intentionally working on a platform that will put all game devs out of work and allow kids to prompt and direct their own content.”

— Senior Employee (Machine Learning Ops), New Mexico

The technology moat isn’t collapsing from outside. People in the industry know it’s collapsing. Some are accelerating it. Most feel powerless to stop it.

That’s what makes this one different from the money moat or the people moat. Those are structural shifts you can adapt to. The technology moat’s collapse comes with an identity crisis. When your tools can be replicated by anyone, what makes you you?

The Moats Didn’t Vanish. They Moved.

The money moat shifted from development to distribution. You don’t need $50M to build a game anymore. But you might need it to make anyone notice.

The people moat shifted from team size to taste and speed. Small teams win because they move fast and build with their audience, not for them.

The technology moat didn’t shift — it’s shattering. Everyone has the same tools. Everyone has access to AI. The industry hates it and uses it anyway.

Studios still defending the old moats — raising money to build big teams using proprietary tech — are optimizing for a world that no longer exists. The new advantages are different: taste, speed, community, and distribution. Most studios aren’t built for them.

The industry isn’t in a downturn. A downturn implies things bounce back. This is a restructuring of who makes games, how they’re funded, what tools they use, and who survives.

The moats aren’t vanishing — they’re moving. And most studios are still standing where the old ones used to be.

Right now, 81% of game professionals use AI for research and brainstorming — but only 5% use it for player-facing features. The gap between AI as a thinking tool and AI as a shipping tool is enormous. The studios that understand that gap, and keep human judgment on everything past the line where AI’s usefulness drops off, are building a compounding advantage. But the window is closing. When adoption hits critical mass and the tools mature, understanding AI’s edges won’t be a moat — it’ll be table stakes.

That gap is your window. It won’t stay this wide.

This is the first in a series going deeper into each of these shifts. Next up: what the data says about who actually survives a moat collapse — and it’s not who you’d expect.

*📩 Subscribe to Gamemakers

One critical takeaway we all agreed on: change isn’t just coming, it’s here right now.

Here are the things I wanted to highlight — and the one nobody wanted to say out loud:

1. The Attention Economy: The noise problem is about to get much worse

2. Going Small: Small teams aren’t a trend. They’re the new default.

3. Progression Compression: Players are more impatient than ever — and that breaks your progression.

4. AI or Bust: If you’re not using AI, you’re ngmi.

👉 Watch the full panel on YouTube below!

🎧 Listen on Spotify, Apple Podcasts, or Anchor

Speakers:

• Oscar Clark, CEO at Fundamentally Games

• Cody Hard, Design Director at EA

• Joseph Kim, CEO at Lila Games

• Brett Nowak, CEO at Liquid & Grit

• Ariel Navon, Head of Competitive Intelligence at Playtika

Below, I’ll break down what stuck with me — and what I was thinking but didn’t say on the panel.

The Noise Problem Is About to Go Parabolic

Bruh, check out this chart of App Store submissions from a16z below.

Post-AI, the curve is bending massively upward! After relatively flat growth over three years, new iOS app releases surged 60% year-over-year in December. They attribute it to “agentic coding” — what people are calling “vibe coding.”

By the end of this year, how much higher do you think this goes?

Here in Silicon Valley, nobody codes anymore. That’s not hyperbole — it’s just how people build software now. Web apps, SaaS products, internal tools — all AI-assisted, built in days instead of months. Games are harder, and game dev will likely be the last software category to go fully AI-native. But we’re already seeing it accelerate.

If you’re interested in why, I can post on this in a future episode.

Think I’m being hyperbolic? Here are just 2 projects ( I actually have others I am not publicly sharing), I personally coded from zero as an old guy who hasn’t coded in over 20 years, and over just a few days:

• ClawTalk: A remote TUI into OpenClaw that enables context-specific conversations, model usage data, multiple agents in context-specific discussions, live voice chat, and more. It’s open source; check it out, especially if you use OpenClaw.

• Synthetical Research: A service we use at Lila to synthetically test art concepts. I created an easy way for anyone to run this research paper: https://arxiv.org/abs/2510.08338

The cost of software and content is trending to zero. Not reaching zero — there are still operating costs, infrastructure, human judgment — but the marginal cost of producing *something* is collapsing by orders of magnitude.

And it’s not just a supply-side problem. On the demand side, players are more impatient than ever. We saw this clearly with ARC Raiders — the progression loop has been compressed dramatically. You don’t grind 20 games to unlock the next map anymore. You play 5, and you’re in. You craft something after your first battle. The “do something, get something” loop has been squeezed to near-zero delay. That’s the attention economy at work — if your game doesn’t deliver value immediately, players bounce to one of the billion other options that will.

So studios are caught in a double squeeze: more competition flooding in from the supply side, and pickier, less patient players on the demand side.

What happens when anyone can build an app, and nobody has patience? You get a billion apps and seconds to prove yours matters.

Oscar Clark put it well on the panel. He compared the moment to the rise of folk bands — people seeking authenticity and trust in response to the synthetic flood. When everything looks the same, the thing that’s clearly made by humans who give a damn stands out.

But I’d push that further. Authenticity alone isn’t enough. You need to signal three things almost immediately: 1. quality, 2. a strange attractor (that’s a Hollywood concept — the thing that makes your project undeniably unique), or 3. mystery (think Simon Stalenhag). If you don’t hit at least one of those in the first 15 seconds (for mobile), you’ve already lost the player’s attention. And that was true before the AI flood. Now it’s existential.

—————————————————————————

The New Winners: Small Teams That Build Communities

I published a post last September (see Alpha Insight #2: The Great Western Game Development Divergence) that looked at what the most successful Western games had in common. Not the $200 million AAA titles — the breakouts. And a clear pattern emerged, especially on Steam and PC.

The winners were small teams. Teams of 1 to 20 people. Lethal Company — one developer, $70M+ net revenue, zero marketing spend. Schedule 1 — one developer, 8M+ units sold. Project Zomboid — about 10 people, hundreds of millions in lifetime revenue over 13 years of early access.

Two things they all had in common:

1. They engaged their community obsessively. Not as a marketing tactic — as a development methodology. Building in public, iterating based on player feedback, making players feel like co-creators. The community wasn’t a distribution channel. It was the product.

2. They moved fast. No 18-month development cycles. No waiting for approval from seven layers of management. Ship, learn, iterate, ship again.

And here’s the thing — it’s not just small teams winning. Delta Force generated 60-70% more revenue than PUBG with a fraction of the DAU (over 22X less) — through layered LiveOps that stack events, progression, and monetization in ways PUBG never attempted. The winners aren’t small or big. They’re fast and smart.

Here’s what this means for 2026: if you’re a new studio, your path to breaking through the noise isn’t a bigger marketing budget. It’s a deeper community and faster iteration cycles. Build something people care about, with people who care about it, and let the community do what no amount of paid UA can.

—————————————————————————

The Thing Nobody Wanted to Hear

I said on the panel that I’m an AI maximalist. I compared the current stigma around AI in games to the early days of online dating — everyone thought it was gross, nobody wanted to admit they did it, and now it’s how most couples meet.

AI in gaming is on the same trajectory. Players care about it today. They care less about this issue for mobile. And in one to five years, it won’t matter at all. You can’t fight the wave. It’s like trying to fight the Internet.

There’s going to be a split: studios that lean into AI to create compelling experiences — faster, cheaper, better, and some things that aren’t even possible without AI — and studios that don’t.

I believe the second group simply doesn’t make it.

—————————————————————————

Next week, I’m going deep on the GDC State of the Industry report — because the data in it confirms a thesis I’ve been developing about what the gaming industry looks like when its moat disappears. The industry isn’t in a downturn. It’s losing its moat. Stay tuned.

The disconnect between fundamentals and valuation tells you everything about what’s really broken at Ubisoft. Public markets aren’t pricing in a business downturn. They’re pricing in a complete loss of confidence in the company's leadership.

🎧 Listen on Spotify, Apple Podcasts, or Anchor

Speakers:

• Matthew Kanterman, Blue River Financial Group

• Joseph Kim, Lila Games

• Eric Kress, Deconstructor of Fun

• Brian Peganoff, Timber Advisors

The Efficiency Gap That Explains Everything

Matthew dropped a stat in our discussion that crystallizes Ubisoft’s problem better than any other: revenue per employee (we warned you about this 2.5 years ago).

Ubisoft generates approximately $123,000 in bookings per employee. EA generates over $500,000. Take-Two exceeds $400,000 (and in a GTA year, that number climbs even higher).

That’s not a gap. That’s a chasm. And it explains why a company generating $1.85 billion in annual bookings can trade at a market cap barely over $1 billion while EA is selling to Saudi investors for $55 billion.

The story gets worse when you dig into Ubisoft’s recent cost-cutting. From March 2023 to March 2025, they eliminated 3,000 employees and cut fixed expenses by €205 million. They’re planning another €100 million in cuts (roughly 1,500 more people) by March 2027. Despite all this surgery, they’re still projecting break-even operating profit.

The business isn’t generating returns. It’s just surviving.

The Tencent Deal: Lifeline or Exit Strategy?

In November, Tencent invested €1.16 billion (~$1.25 billion) to acquire a 26.32% stake in a new subsidiary, Vantage Studios. This carve-out contains Ubisoft’s crown jewels: Assassin’s Creed, Far Cry, and the Tom Clancy brands (including Rainbow Six).

Brian’s framing of this deal was blunt: “Your company is worth a billion dollars. You’re looking at this rich Chinese investor and saying, ‘You can have my whole company for a billion dollars.’ Instead, they say, ‘How about I just give you a billion dollars for 25% of your three biggest franchises?’ And they shake hands. This only ends one way.”

The deal terms tell the story:

• Ubisoft must retain a majority stake in Vantage for at least two years

• Tencent has a five-year lockup, while Ubisoft maintains the majority

• Roughly 25% of the cash went to paying down debt

This isn’t a turnaround investment. It’s a call option. Tencent gets exposure to franchises that have generated $9.5+ billion over the last decade, with minimal downside risk. If the Guillemot family continues to mismanage these IPs, Tencent can wait and potentially acquire the rest at distressed prices.

Matthew suggested an alternative endgame: “What’s more likely to me than Tencent buying it out is in two years they flip Vantage to private equity. They sell Vantage to private equity and leave Ubisoft shareholders holding the bag on everything else.”

How Did Europe’s Gaming Champion Implode?

Three structural failures destroyed Ubisoft’s position.

First: The Copy-Paste Design Philosophy

Starting in the 2010s, Ubisoft perfected open-world game design. The problem? They applied the same template to Assassin’s Creed, Far Cry, Watch Dogs, and Ghost Recon without meaningful iteration. What worked well for years eventually became a liability as players grew fatigued with the formula.

Assassin’s Creed Valhalla (2020) was actually the franchise’s best-selling entry. The RPG-style pivot in Odyssey and Valhalla worked commercially, even if longtime fans complained the games didn’t feel like “real” Assassin’s Creed anymore. But Ubisoft couldn’t sustain the momentum. And critically, they never built a third Assassin’s Creed studio to maintain an annual release cadence—something Matthew flagged as an inexplicable strategic failure.

Second: Death by a Thousand Pet Projects

Eric's list of misallocated capital is damning: The Crew, Starlink (Toys to Life, launched years after the category died), Rocksmith (which Eric acknowledged was "a great product" but "the juice wasn't worth the squeeze" for profitability), Haze, Avatar, the Quartz NFT platform, XDefiant, and Skull & Bones.

Skull & Bones deserves special mention. An 11-year development cycle. Estimates between $500 million and $1 billion in total investment. The project grew from a minigame in Assassin’s Creed Black Flag—boats were fun, so Ubisoft decided to spin off an entire game. Meanwhile, Xbox launched Sea of Thieves and captured the exact market Skull & Bones was chasing.

Star Wars Outlaws adds to the destruction. The game’s mechanics were solid, but design decisions alienated the target audience. It sold roughly 1 million units—disastrous for a AAA Star Wars title, given whatever minimum guarantee Ubisoft paid Disney.

Eric’s indictment: “EA, all the other publishers were focusing on three or four big franchises and doing a few risky bets. These guys were taking risky bets all over the place.”

Third: Management and Culture

This is where the conversation got uncomfortable.

Serge Hascoët, Ubisoft’s Chief Creative Officer and the creative force behind Assassin’s Creed, Ghost Recon, and Far Cry, resigned in 2020 amid sexual harassment scandals. Three executives ultimately received suspended sentences—Hascoët (18 months), former VP Editorial Thomas François (3 years), and former game director Guillaume Patrux (12 months). The culture Eric described was “endemic.”

But Hascoët’s departure created a different problem. Yves Guillemot, who Eric describes as “not a design person,” took over creative direction for the company’s biggest franchises. Family members—including Yves’s son Charlie, now co-CEO of Vantage Studios overseeing Assassin’s Creed, Far Cry, and Rainbow Six—have proliferated through leadership positions.

The departure of Yannis Mallat (Managing Director of Ubisoft’s Canadian studios) and Cécile Cornet (Global HR) in 2020 marked the end of objective management outside the family. Since then, every major decision has been made by the Guillemot family members and their loyalists.

Eric’s assessment: “This whole thing was set up so they could maintain the family. Yves has always wanted his family to inherit this company. And this is what this Tencent deal gave him—that security. But it’s definitely not good for the franchises, certainly not good for the shareholders.”

The France Problem

One structural challenge Ubisoft cannot solve: 25% of its workforce (roughly 4,000 people) is based in France, where labor laws make headcount reductions extraordinarily difficult.

Eric’s provocation: “If you actually shut down Paris, if you shut down all operations in France, you would lose almost no value to this company. They literally have no development and no management required to manage their biggest franchises in France.”

All of Ubisoft’s valuable IP development happens in Canada. The Tencent deal essentially creates a path where “Canada will be Ubisoft and France will just basically get shut down.”

The Only Bull Case

Is there any path to recovery?

There is likely great talent within Ubisoft studios. Eric emphasized that the development teams—particularly in Montreal—are capable of making great games. The Division team is excellent. The Assassin’s Creed team has delivered when given proper direction.

According to Eric, the problem isn’t the people making games. It’s the decisions being made above them.

Clair Obscur: Expedition 33, the critically acclaimed RPG from Sandfall Interactive, was led by a former Ubisoft producer who left the company. The talent likely exists. The question is whether it can express itself under the current leadership.

Rainbow Six Siege proved that Ubisoft could build and operate a genuine live service success. That template exists—they just never replicated it.

But our collective view was pessimistic. The company remains under the same management that destroyed over $15 billion in shareholder value. The Tencent deal provides runway, but it also reveals how limited outside confidence in a recovery truly is.

Brian summarized what public markets have already concluded: “This is not a mystery to anybody anymore. This company is using this as a slush fund for the family to operate and do a terrible job, and they don’t give a shit anymore.”

Key Takeaways

• Ubisoft’s revenue per employee ($123K) is 4x lower than EA’s ($500K+)—this efficiency gap explains the valuation collapse better than any single failure

• 70% of Ubisoft’s bookings come from back catalog titles—they’re living off past hits, not building new franchises

• The Tencent deal is a call option, not a rescue—€1.16B (~$1.25B) for 26% of the company’s best IPs is positioned as a put, not a bet on recovery

• All valuable IP development happens in Canada—the French operations contribute minimal value while employing 25% of the headcount

• Serge Hascoët's 2020 departure removed the creative brain—Yves and family members have made design decisions since, with catastrophic results

• The talent is still there—studios like Montreal can make great games under the right leadership; the problem sits at the executive level

What’s your take on Ubisoft’s future? Can the company recover under current management, or is the Tencent deal just the beginning of the end? Leave a comment below.

AppLovin is the apex predator in mobile right now. My business is proof that they’re not being prohibitively high in their pricing—because if they were, I wouldn’t be succeeding. But we do lack independence, and it does come across as though we are operating inside somebody else’s system.

That’s Josh Chandley, CEO of WildCard Games and one of the sharpest operators in mobile monetization, describing the uncomfortable reality of building games in AppLovin’s ecosystem. His portfolio generates millions in revenue through AppLovin’s platform—and he’s still nervous about what comes next.

The gaming industry’s most controversial company just crossed $250 billion in market cap. The stock is up 127% year-to-date. EBITDA margins hit 82% last quarter. And yet many executives and investors in the mobile and gaming industry ask the same question: Is this the beginning, or the end?

I assembled what may be the most qualified panel possible to answer that question: Chandley with his hands-on operational data, Matthew Kanterman with a financial analyst lens, and Brian Peganoff with an institutional investor perspective. What emerged wasn’t the typical bull-bear debate—it was a blueprint for understanding how dominance actually works in ad tech.

🎧 Listen on Spotify, Apple Podcasts, or Anchor

Speakers:

• Josh Chandley, President & CEO at WildCard Games

• Matthew Kanterman, Director of Research at Blue River Financial Group

• Brian Peganoff, Former TMT Investor, Founder Timber Advisors

• Joseph Kim, CEO at Lila Games

The Numbers Tell One Story. The Operators Tell Another.

AppLovin’s financials look like a typo. Revenue and EBITDA are up 72% through the first three quarters of 2025. The stock trades at 32x forward revenue and 39x forward EBITDA. At a PEG ratio of approximately 1, the valuation is technically justified by projected 2026 revenue growth of 35-40% and EBITDA growth of 40%.

“It’s been a loser’s bet basically for the last three years to bet against Adam,” Kanterman notes. “The valuation has re-rated. The question is what’s next.”

But here’s what the financials don’t capture: the structural lock-in that makes AppLovin’s margins possible in the first place. Chandley laid out the mechanics with uncomfortable specificity.

AppLovin owns MAX, the dominant ad mediation platform. They charge competing networks a 5% fee to bid on inventory—while exempting themselves. “They don’t have to be as good as Meta or Unity,” Chandley explains. “They just have to be 95% as good.” In machine learning-driven auctions where advantages compound at the margin, that’s an insurmountable head start.

The lock-in deepens through Axon, their demand-side platform. For ad-monetized games, it’s completely normal to have over half of your users coming from Axon. And here’s the key: you can only use Axon if you use MAX. Publishers who leave MAX lose access to Axon—and with it, roughly half their user acquisition pipeline.

“That’s why over the last couple of years we’ve seen them able to double their revenue publicly without actually really increasing their share of voice very much,” Chandley says. “Their take rate is just doing incredibly well.”

Translation: AppLovin isn’t growing by winning more auctions. They’re growing by extracting more value from the auctions they already win.

The E-Commerce Expansion: Where the Real Action Is

The bull case for AppLovin increasingly rests on its expansion into e-commerce advertising. Chandley’s data from his own portfolio is striking: at one point on iOS, all 10 of the top creatives in his games were AppLovin e-commerce ads. Around Black Friday, AppLovin passed Google Ads on Android for e-commerce—on Google’s home turf.

“Most importantly for their stock, that dominance—the share of voice, the number of impressions that AppLovin is winning as a percentage—is barely changing despite all of this new demand flowing in,” Chandley explains.

Think about what that means. E-commerce advertisers are flooding in, willing to pay premium prices. But AppLovin isn’t passing that value to publishers. The gap between what advertisers pay and what publishers receive is widening—and that gap is pure margin.

The structural advantages that made AppLovin dominant in gaming translate directly to e-commerce. Their rewarded video inventory is fundamentally different from Meta’s feed ads. “An ad on AppLovin’s rewarded video is a 30-second infomercial,” Chandley notes. “The attention is guaranteed.” That matters enormously for e-commerce conversion.

Meta and Google face three barriers to competing effectively in this space. First, iOS: AppLovin’s share of voice is 30% on Android but 55-60% on iOS. ATT gifted them that position, and five years later, Meta and Google still haven’t recovered. Second, creative relevance: 15-second thumb-stopping content optimized for social feeds doesn’t perform as well in 30-second captive-audience slots. Third, organizational DNA: Meta and Google built trillion-dollar businesses on first-party data within walled gardens. Their entire infrastructure—ML models, sales organizations, account management—struggles outside their owned-and-operated properties.

“Google and Meta are fighting distracted giants in a venue where those giants can’t really use their greatest strengths,” Chandley summarizes.

The Bear Case: Expectations Already Assume Victory

Brian Peganoff frames the bear case precisely: “It’s not like this is a broken company with some hidden structural flaw. The bear case comes down to the valuation already reflecting success.”

At a $250 billion market cap, the stock isn’t expensive given the stretched multiple. It’s expensive because earnings already assume a lot goes right. The risk is disappointment, not multiple compression.

Peganoff identifies five specific concerns. Growth durability—upside depends on self-serve adoption, Axon performance gains, and advertiser wallet share continuing to compound. If that flywheel slows even modestly, operating leverage flips from tailwind to limiter. Expansion optionality—e-commerce and connected TV are attractive but not yet scaled or transparent enough to underwrite as core drivers. Scale effects—the last two years benefited from simplification and low-hanging improvements that become harder to find. Competitive convergence—Google, Meta, Amazon, and Trade Desk don’t need to beat AppLovin; they just need to narrow the gap.

The SEC investigation adds a layer of uncertainty. The agency’s Cyber and Emerging Technologies unit is examining AppLovin’s data collection practices, which were triggered by a whistleblower complaint and amplified by short sellers (Muddy Waters, Fuzzy Panda, Culpert Research). The allegations center on fingerprinting and potentially fraudulent attribution practices.

The panel’s consensus: the risk of deplatforming is real but low. “The likelihood is probably 2-3%,” estimates one panelist. I think it’s higher—maybe 20-30%—but it's still not the most likely outcome. The more likely result is platform restrictions or fines that affect the business model without destroying it.

But there’s a nuance worth considering. “Anytime an ad tech platform shows this kind of performance, people are going to ask where that data access is coming from and could it get shut off,” Peganoff notes. “It’s the whole Cheetah Mobile conundrum.”

The Infrastructure Risk Nobody’s Discussing

Chandley raises what he considers the only significant negative risk: losing control of supply.

Right now, publishers stay loyal to MAX because they want access to Axon’s ad ROAS campaigns. But that loyalty has a price. If another platform could increase LTV enough to overcome the loss of Axon, publishers would leave. Or if Axon’s advantage shrinks enough that existing platforms could jump the hurdle, there would be a flight from MAX.

Both scenarios are happening simultaneously, Chandley argues. AppLovin’s take rates are extremely high—if they decreased from 50% to 30%, publishers would see a 40% increase in LTV, enabling much more aggressive scaling on Meta and Google. Meanwhile, MAX itself appears to be aging. “It has been largely underdeveloped. Axon 2 and the e-commerce opportunity have been front and center for them.”

The specifics get technical but matter enormously. Some networks are exploiting open auctions to manipulate bid prices. Unity opened up its LevelPlay ad quality tool for MAX, which seems helpful until you realize it gave Unity direct access to data flowing through AppLovin’s platform. Other parties are nibbling away at AppLovin’s data advantage through product offerings that fill gaps MAX has neglected.

“This quarter, WildCard has spent less on AppLovin than we expected,” Chandley reveals. “Some networks are starting to cross over. Competitors are improving—Unity, Moloco, and to a lesser extent Mintegral improved their performance dramatically in 2025.”

The structural question: Can AppLovin scale e-commerce without threatening their ad ROAS moat in gaming? Can a relatively small company by headcount (1,500 employees versus Unity’s 5,000) execute on Axon, e-commerce expansion, and MAX improvement simultaneously?

The Competitive Landscape: Three Arenas, Different Dynamics

Mobile Gaming & Mediation: Unity theoretically has advantages AppLovin lacks—70% engine market share, direct IAP data visibility, and the IronSource acquisition. In practice, execution has lagged. “I’ve noticed over the years that there has been an execution gap at Unity,” Chandley acknowledges. But he’s careful to credit the Unity Grow team specifically: “Vector has shown it’s really good. It’s been picture perfect. We’ve seen it more than double share this year.”

The structural opportunity for Unity remains tantalizing: if they can execute on passing back engine data through Unity 6.2, they’d have the same kind of first-party data advantage in IAP games that AppLovin has in ad-monetized games. But timeline risk is real—these data pipelines take longer to build and debug than anyone expects.

E-Commerce: The owned-and-operated networks (Meta, Google) have significant demand but face iOS-related challenges and creative relevance challenges. Moloco is expanding but lacks the mediation infrastructure that gives AppLovin its structural advantage. For now, AppLovin is running largely unopposed.

Connected TV: This is genuinely a wild card. Trade Desk dominates programmatic CTV but had execution issues last year as Amazon took share. AppLovin’s six times Trade Desk’s EBITDA on a little more than double the revenue—they have the resources to compete. But CTV presents different challenges: zero percent of commerce occurs on the TV screen, attribution is less clear, and the targeting advantage that works in mobile gaming may not translate.

“Connected TV is kind of a dumb surface,” Peganoff notes bluntly. “No transactions occur there. The personalization problem is fundamentally different.”

What Each Panelist Actually Thinks

Brian Peganoff (Institutional Investor Lens): “Directionally positive. Scale has reduced risk, not increased it. The company is simpler than it was two years ago, margins are real, execution has been consistent. You’re underwriting continuation, not a new strategy. But expectations are already high—upside has to come from earnings continuing to surprise, not from the market paying more for the same dollar of profit.”

Matthew Kanterman (Financial Analyst Lens): Largely agrees with Peganoff, but emphasizes macro risk. “AppLovin has a beta of three. A lot of that is negative downside skew. If something bad happens or there’s a massive liquidity drain from the market, AppLovin drawdowns have been violent and aggressive. We’re one tweet away from the president sending the market down 20%.”

Josh Chandley (Operator Lens): “AppLovin is an incredible business. Every person I know at that business is exceptional. They’ve taken huge margins and returned huge value for their customers at the same time. 2026 will start with a victory lap, and the e-commerce opportunity is going to look really good. But 2026 is also when the next chapter starts to emerge—when it becomes clear whether there’s a real competitor. What are Google and Meta going to do on iOS? Can Unity execute on engine data? The fog will start to lift.”

Joseph Kim (Mr. Handsome): I’m probably the most bearish person here, though even I think they roll in the short term—maybe to $500 billion market cap. My concerns center on growth deceleration in e-commerce as Meta and Google eventually respond, and I don’t think the deplatforming risk is quite as negligible as others do. But the structural moat is real, and in the short term, momentum wins.

The Strategic Implications

For Game Studios: You’re operating inside someone else’s system. That’s not necessarily bad—WildCard is proof you can succeed within AppLovin’s ecosystem—but diversification matters. Test Unity’s Vector for IAP campaigns. Maintain relationships with multiple ad networks. Watch Meta and Google’s iOS strategies closely; if they crack probabilistic attribution at scale, the dynamics shift.

For Investors: This is a quality compounder, not a momentum trade. The multiple expansion story has played out. From here, upside requires earnings surprises, which in turn require e-commerce expansion and self-serve adoption to exceed already-high expectations. If the market de-rates broadly, AppLovin will come with it—but likely from a position of strength.

For Competitors: The window is narrowing but not closed. Unity has the data assets to compete if execution improves. Meta and Google have the demand if they can solve the iOS problem and the creative relevancy challenges. But every quarter of inaction is another cohort of games locked into MAX, another round of advertisers trained on AppLovin’s infrastructure.

The Quarter-Trillion Dollar Question

Here’s what makes AppLovin genuinely fascinating: the business is simultaneously the most dominant ad tech platform in gaming, the most controversial stock in the sector, and the most misunderstood by traditional investors who don’t grasp the technical mechanics of how their moat actually works.

The bull case is that the flywheel keeps compounding—more publishers on MAX means better data for Axon, which means better performance for advertisers, which means more demand, which means higher take rates, which means more publisher lock-in. E-commerce expansion is gravy on top.

The bear case is that expectations already assume the flywheel works perfectly. Any slowdown in growth, any competitive response from Meta or Google, any regulatory action that affects their data practices, and the multiple compresses faster than earnings can grow.

Both cases can be true. AppLovin can be a great business and a risky stock at current valuations. The dominance can be real, and the runway can be shorter than investors expect.

Chandley puts it most precisely: “I find myself stating the bear case a lot more than I probably should, just because it’s so under-talked about. But like every person I know at that business is exceptional. They’ve executed to perfection so far—at least in the short to intermediate term window.”

For the industry, the implications extend beyond one company’s stock price. AppLovin’s success demonstrates that owning the infrastructure matters more than owning the inventory. That vertical integration from mediation to demand-side to attribution creates compounding advantages that traditional ad networks can’t match. The mobile attention economy is consolidating into fewer, more powerful hands.

Whether that’s good or bad depends on where you sit. But it’s definitely happening.

Key Takeaways

• AppLovin’s moat isn’t performance—it’s infrastructure. The MAX/Axon lock-in means they don’t have to be better than competitors; they need to be 95% as good.

• E-commerce expansion is working, but not being passed to publishers. Share of voice is flat while revenue doubles—the gap is pure margin extraction.

• The bear case is about expectations, not fundamentals. At $250B market cap, the stock already prices in continued dominance.

• Deplatforming risk is real but probably low. The SEC investigation matters, but platform restrictions seem more likely than complete removal.

• 2026 is when competitive clarity emerges. Unity’s engine data play, Meta and Google’s iOS response, and new entrants will either materialize as threats or fail to launch.

• For operators: diversification matters. Success within AppLovin’s ecosystem is possible, but dependence on any single platform creates risk.

What’s your read on AppLovin’s next chapter? Reply with your thoughts—I read every response.

Royal Match has generated over $4 billion in lifetime revenue (according to AppMagic). When Dream Games CEO Soner Aydemir explained their success in a TechCrunch interview, he said something that sounds obvious but has radical implications:

“If you don’t have good enough metrics, even with all the money in the world, it’s impossible to scale.”

This isn’t false modesty. It’s a fundamental truth about how F2P mobile actually works—and few teams act like it’s true.

Today, I’m breaking down why “Product Is Marketing” and the three specific implications this has for your studio’s mindset, velocity, and organizational structure.

Why F2P Is Different

In retail, e-commerce, or SaaS, marketing and product can operate relatively independently. When Nike sells you a shoe, the transaction happens before you experience the product. Marketing creates desire, you pay, then you wear the shoe. If marketing is great, you might overpay for a mediocre product.

Premium games work the same way. I have over 300 games in my Steam library. I haven’t played the vast majority for more than 30 minutes. Publishers marketed to me successfully with products I didn’t actually engage with.

F2P mobile is fundamentally different.

The transaction happens during the product experience—often days or weeks after initial engagement. If your product doesn’t retain users and convert them to payers, no amount of marketing genius can save your unit economics. You’re just paying to acquire users who churn before they monetize.

Your retention curve, conversion rate, and ARPU aren’t just quality metrics. They determine your marketing budget. They are your marketing capability.

The Math That Dictates Your Ceiling

The relationship between CPI and ad spend volume is non-linear. As you scale spend from $100/day to $100,000/day, your CPI rises dramatically. At some point, it taps out. You get priced out of the auction.

If your LTV is mediocre, your marketing ceiling is low. If your LTV is superior, you can weaponize it.

The classic case study: Machine Zone (Game of War) vs. Kabam (Kingdoms of Camelot) in the early 2010s.

Kingdoms of Camelot was a successful 4X game with an LTV around $8-10 (according to word on the street at that time)—incredible for that era. Machine Zone copied the formula but studied Chinese MMOs, layering in aggressive monetization features like VIP tiers and gold banks. Game of War hit $12-15 in LTV (again, based on word on the street).

MZ used that delta to sweep ad inventory. They bid aggressively, monopolized the available supply, and effectively nuked Kingdoms of Camelot by starving it of traffic. Game of War’s LTV advantage became a weapon.

Product metrics don’t just measure health. They dictate your ability to compete.

Implication #1: From Fatalism to Agency

Too many teams treat product metrics like the weather—something that just “happens” to them.

The mindset is: “We built a great game. Now let’s soft launch and measure retention to see if it works. I hope the numbers are good.”

This is an attempt to divert accountability. It’s an intentional disassociation of the team’s effort from the outcome.

Here’s what’s odd: everyone agrees that soft launch is used to optimize product metrics. Teams work to drive these metrics. That’s not controversial. But then those same teams treat their ability to drive metrics as somehow separate from the product itself—as if the team is on one side and the results are on the other.

They’re not separate. They’re the same thing.

The winning mindset is agency and self-determinism. Dream Games didn’t stumble into Match-3 dominance. They run approximately 60 A/B tests annually. Aydemir said they spent “so much time on tiny details, so many tests over several years to create the dynamics of the game.”

Implication #2: Product Velocity Is the Killer Variable

If the product sets your marketing ceiling, how quickly you improve it determines how quickly you can scale marketing.

Product velocity is the killer variable. If it takes you months to implement a feature that should take weeks, you are dead in the water. If you can iterate faster than your runway burns, you win. If you can’t, you lose.

I struggle with this too. At my own studio, we recently spent 42 days getting product analytics (PostHog) into production. That should have taken a few days. Let me be clear: that lack of velocity puts us on a path to failure unless we fix it.

I’m not going to pretend it’s out of my control. It’s not fate. It’s me. It’s our team. It’s our ability to generate creative solutions and execute them with speed.

Velocity requires looking at your team with a critical eye. It means conducting a root cause analysis of why things are slow. Often, this means having uncomfortable conversations about people and processes—something that rarely happens because criticizing people at most organizations is playing with fire.

But if you don’t fix velocity, you cannot fix the metrics that unlock marketing scale.

Implication #3: The End of Silos

At most gaming companies, Marketing and Product are separate kingdoms. Different leaders, different meetings, different dashboards, different incentives.

This leads to the CYA game:

Marketing says: “CPI is low. We’re doing our job. The product just doesn’t retain.”

Product says: “Retention is fine. Marketing is bringing us garbage users.”

If Product is Marketing, this structure is obsolete. You’re creating artificial friction between two functions that need to operate as one system.

You need shared reporting—one leader owning the full loop from UA efficiency to monetization outcomes. You need shared dashboards—product must see UA data, UA must see product KPIs. And you need shared accountability—when the game fails to scale, it’s the team’s fault, not a specific department’s fault.

I know this is hard to implement, especially at larger companies with entrenched structures. But at minimum, get these teams in the same room, looking at the same data, with the same goals.

The Hopeful Flip Side

When Aydemir said it’s impossible to scale without good metrics, that’s the brutal truth of F2P mobile. But here’s the flip side:

If your product metrics are good enough, you can scale. Small teams can beat entrenched competitors with massive war chests. Royal Match competed in Match-3—one of the most competitive genres in mobile—and won.

Marketing is just execution. The product is the unlock.

And improving those metrics is within your team’s power. Not fate. Not luck. Ownership and relentless experimentation.

Because in mobile F2P, the product isn’t just what you sell—it’s how you sell it.

Takeaways

• Your product metrics ARE your marketing budget. LTV determines how much you can spend to acquire users and how far you can scale before getting priced out. Better metrics = higher ceiling.

• Stop treating metrics as lottery results. Dream Games runs 60+ A/B tests annually. The mindset of “let’s see if the product is good enough” is fatalism. Iterate until it’s good enough.

• Product velocity is survival. If you can improve faster than your runway burns, you win. Fixing velocity means confronting uncomfortable truths about people and processes.

• Kill the silos. Shared reporting, shared dashboards, shared accountability. When product and marketing blame each other, the real problem is your org chart.

• Competitive advantage compounds. As Game of War demonstrated, even small LTV advantages let you outspend competitors and starve them of inventory. Product metrics aren’t just about profitability—they’re weapons.

Over the years, I’ve watched the same validation mistakes repeatedly kill games or waste millions in development costs. Different studios, different genres, different budgets—same failure patterns.

Studios conducted extensive player tests, including focus groups, surveys, and playtests with hundreds or thousands of players. The data came back positive, with high satisfaction scores, statistically significant results, and greenlights from leadership. And yet, the games launched and failed.

Meanwhile, some of the biggest hits—games that became genre-defining—barely tested at all before committing to their vision. PUBG didn’t survey players about whether battle royale would work. Brendan Greene just built it.

The studios that succeeded often understood something the failures didn’t: validation isn’t science, it’s art. They know when to trust one designer’s taste over a thousand players’ opinions—and when to do the opposite.

Here’s the framework I’ve developed for knowing when to trust your taste and when to trust your players.

Executive Summary

• Validation is art, not science. You must trust tastemaker judgment for radical innovation, but use player data for proven models.

• Most studios don’t fail by picking the wrong approach. They fail from “invisible” problems in their process.

• These invisible failures are:

  • Broken Methodology: Running tests that look scientific but are flawed, yielding worthless data.

  • Wrong Tastemakers: Committing to a “top of the pyramid” vision when the person at the top has bad taste.

  • Validation Theater: Using over-intellectualized frameworks that look rigorous but give your team nothing they can actually build or execute.

The Validation Pyramid: When to Trust Taste vs. Data

I’ve spent years working with designers across game studios, trying to understand why some validation approaches work while others fail spectacularly. This framework emerged from those conversations and observed patterns. It’s not backed by comprehensive academic research—it’s a mental model, a thinking tool for navigating the art/science tension in product validation.

Think of validation using the “Pyramid Design Model”. At the top sits one person—a designer, creative director, tastemaker—making design decisions based on judgment. This is Steve Jobs territory: “People don’t know what they want until you show it to them”. At the bottom sit thousands of players generating data through feedback and behavior—the world of surveys, A/B tests, and statistical significance.

The critical question isn’t which approach is right. It’s which approach matches your level of innovation.

Brendan Greene created battle royale. Before his work in ARMA and H1Z1 King of the Kill, there was no “battle royale hill” to climb. Players couldn’t give feedback on whether the concept would work because the reference point didn’t exist. Brendan Greene had to trust his vision that dropping 100 players onto a shrinking island would be compelling.

Riot turned the Warcraft III DotA mod into League of Legends using extensive player feedback to refine champions, items, and meta. Players provided meaningful signals because they understood MOBAs. Marc Merrill describes their philosophy: “When we’re faced with difficult decisions, we always ask: What’s best for the player?”

The difference? Player feedback optimizes you toward the top of your current hill. Players evaluate based on what they know. Perfect for optimization. But if you’re trying to find a new peak—a new genre, an untested mechanic—asking players for direction is asking people at the bottom of one hill to tell you if a different hill they can’t see is taller.

I’ve seen this pattern destroy studios: they use optimization techniques (player feedback, A/B testing, incremental iteration) when they need discovery (vision, taste, willingness to jump hills). They optimize their way to the top of the wrong hill.

When Your Tastemaker Is Wrong

Here’s the uncomfortable reality that makes validation even harder: operating at the top of the pyramid only works if the tastemaker has good taste. What if your creative director is wrong?

I’ve seen projects start with a tastemaker who seemed right—strong portfolio, confident vision, team buy-in—only to discover months or years later that their taste didn’t match what would actually work. The team committed to the top-of-the-pyramid approach, trusted the vision, and climbed the wrong hill for months or years.

This is one of the most challenging problems in validation because it is difficult to test taste in advance. But there are warning signs:

Fall back on first principles. Why should this concept work? What player needs does it address? If the tastemaker can’t articulate the logic beyond “trust me,” that’s a red flag. Vision needs rationale, not just confidence.

Seek external tastemakers with relevant expertise. Show the concept to 3-5 people whose taste you respect in the relevant domain. Do they see the vision? Do they get excited? If they’re lukewarm or confused, that’s a signal. You’re not looking for approval—you’re stress-testing whether the vision makes sense to other people with good taste. Note, however, that the biggest opportunities will be the most controversial: a “secret.”

Validate the validator. Past success in a different domain doesn’t guarantee taste in your current one. A designer who shipped a successful match-3 game might not have the right instincts for a hardcore strategy game. Domain expertise matters.

Watch for vision drift. If the creative vision keeps changing fundamentally every few weeks, that’s not iteration—that’s a tastemaker who doesn’t actually know what they want. Vision should evolve and sharpen, not pivot constantly.

The hardest decision: if your tastemaker is wrong and you’re committed to the top of the pyramid, you might need to change who’s at the top. Better to face this early—painful as it is—than waste months or longer climbing the wrong hill. I’ve seen studios delay this decision because it feels brutal, but keeping the wrong tastemaker is more brutal. It kills the project.

The 50-50 Gender Split That Revealed Worthless Data

Here’s why broken validation is more dangerous than no validation.

A mobile studio I know ran a CSAT survey to validate a new feature before greenlighting production. They distributed it through an offerwall—players took the survey in exchange for hard currency in another game. Thousands of responses. Statistical significance achieved. High satisfaction scores. Green light given.

One problem: the players didn’t read the questions. They tapped randomly to get their currency reward. The entire dataset was garbage.

The team caught it because someone noticed the male-to-female ratio was 50-50. For that game, it should have been 80-90% male. That anomaly triggered an investigation. They discovered players were gaming the survey.

Here’s what makes broken tests so dangerous: they’re invisible. They look identical to good tests in your spreadsheet. Both give you numbers. Both achieve significance. Both generate the confidence that executives desperately want. The only difference is methodology—and most people never audit that.

I’ve seen this pattern repeat in different forms: Wrong incentives where players game surveys. Wrong context where low-fidelity prototypes can’t be meaningfully evaluated. Wrong wording that biases specific responses. Wrong sample where the wrong personas judge your game. Wrong interpretation where correlation becomes causation.

Any of these invisibly contaminates your data. In science, peer review catches flawed methodology. In game development, launch day is your peer review. By then, you’ve already bet on flawed data.

The Danger of Over-Intellectualization

There’s another way validation fails that’s particularly insidious: it looks sophisticated, scientifically rigorous, and impressively complex—but yields absolutely nothing operational or gives a wrong result.

The Head of R&D for a major game studio recently told me, “The vast majority of these studies are useless. These complicated motivational frameworks are generally not useful. You can do a lot of surveys, but they tend to be fluffy and lead to things your team will never understand how to execute.”

I’ve seen this pattern repeatedly. A studio hires a research firm or builds an internal research team. They develop elaborate player personas based on psychographic segmentation. They create motivational frameworks with names like “The Achievement-Oriented Competitive Socializer” and “The Mastery-Driven Aesthetic Explorer.” They conduct studies that produce detailed reports with charts, statistical analyses, and sophisticated academic methodologies.

Then they try to use it. And nothing happens.

The design team examines the personas and asks, “Okay, so what do we build for the Achievement-Oriented Competitive Socializer?” The answer is unclear. The framework is too abstract. The motivations are too general. The insights don’t translate into concrete design decisions.

The problem isn’t that the research is wrong—it’s that it’s not useful. You can have perfectly valid psychological frameworks that tell you nothing about whether your game loop is fun, whether your monetization will work, or whether players will understand your core mechanic.

This type of validation failure is particularly dangerous because it creates the illusion of rigor. Leadership sees impressive decks. Money gets spent. Time passes. Teams feel validated in their approach because they have “data-driven insights.” But when you ask what actually changed in the game because of the research, the answer is often nothing—or worse, the wrong result based on academic fluff.

The solution isn’t to avoid research—it’s to demand operational clarity. Before you invest in any validation methodology, ask: “What decision will this help us make? How will the answer change what we build? Do we actually understand and believe these fancy academic frameworks?” If you can’t answer clearly, the validation is probably intellectual theater rather than practical and correct insight.

Simple, concrete validation—“Players churned at this step in the tutorial. Let’s test three different approaches and measure funnel drop-off”—almost always beats elaborate theoretical frameworks. Focus on answering specific questions that directly change what you build.

When Asking Players About Innovation Kills It

I’ve seen two different studios make the same validation mistake. Both were designing new games and wanted to validate concepts with players.

They surveyed hundreds: “Which genres do you enjoy? MOBA? City-building? RPG?” Players selected multiple options. The data came back clear—players loved all three genres. They even believed that all three could be achieved in a single game.

Both studios created games combining three genres. After all, the data showed players wanted this. Both games launched. Both failed.

The problem wasn’t the data—it was interpretation. Players who like three different genres don’t necessarily want those genres mashed together. And it’s difficult for players to grasp what that hybrid would actually mean. Players want great MOBAs. Great city-builders. Great RPGs. Not a confusing hybrid that tries to be everything and becomes nothing.

Players will tell you what they like. But “what do you like?” isn’t the same as “what should we build?” The first question measures taste. The second requires vision. You can’t crowdsource vision from survey data.

Three Ways Player Feedback Generates Noise

Even when your methodology is sound, you face signal vs. noise problems. Through working with studios on validation, I’ve identified three ways player feedback generates noise instead of signal:

The conceptual gap. Players can’t evaluate what doesn’t exist yet. You’re not testing your game—you’re testing their ability to imagine what your game could be. Novel concepts need top-of-pyramid validation with small groups who can envision what doesn’t exist.

The persona mismatch. Asking the wrong players about your game is like asking vegans which hamburger they prefer. Test a hardcore strategy game with casual match-3 players and you’ll get clear, statistically significant, ultimately useless feedback. The data is accurate. The sample is wrong.

The articulation gap. Even expert players may struggle to explain what’s wrong. They feel something is off. They can tell you they’re not having fun. But they can’t necessarily articulate why or what would fix it. Their feedback is real—their diagnosis might be completely wrong. This is why behavioral data matters more than stated preferences.

Quality beats quantity in early validation. Five players who match your target persona and can envision what doesn’t exist generate more signal than a thousand random players. As your concept becomes more concrete, these noise problems decrease and large-scale player feedback becomes more valuable.

Why Perfect Validation Can’t Predict Execution

Here’s what no validation methodology will tell you: whether you’ll execute well enough for any of it to matter.

BioWare’s Anthem playtested well. EA CEO Andrew Wilson reported players were satisfied through 30 hours of content. The validation wasn’t fundamentally broken—players did enjoy what they played. The problem was execution.

The game spent seven years in development but didn’t truly enter production until the final 18 months. Leadership couldn’t provide consistent direction. The Frostbite engine created technical challenges. Content systems weren’t ready. The live service loop felt like a grind.

Validation couldn’t predict any of this. It couldn’t tell you whether the endgame would have enough content, that technical issues would plague launch, or that thousands of small decisions in UI, progression, and social features would fail to come together.

Even perfect validation doesn’t guarantee success. Between concept and launch sit thousands of execution decisions that validation can’t predict. This is why treating validation as pure science is dangerous. It creates false certainty. You gather data, reach significance, and believe you’ve eliminated risk. But you’ve only validated part of the equation.

The best studios I’ve worked with combine smart validation with a ruthless focus on execution. They never confuse validation with execution. Validation tells you what to build. Execution determines whether you build it well.

Matching Validation to Development Stage

Here’s what I’ve learned about validation across development stages:

Pre-Production: Trust Tastemaker Vision (But Validate the Tastemaker)

You’re validating whether the core idea resonates. Operate at the top of the pyramid. A lead designer or small team decides the vision based on taste and judgment. But make sure your tastemaker has good taste—test with 3-5 external tastemakers whose judgment you respect. Use fake ad CTR tests to signal interest, but the tastemaker still decides what to build.

The signal you’re looking for: Does this feel genuinely new and compelling? Does the logic for why it should work hold up under scrutiny?

Production: Mix Designer Judgment with Curated Feedback

You’re validating whether mechanics are fun and the game is learnable. Internal playtesting with target-persona staff (watch behavior, not words). Small-group playtests with 20+ players—qualitative feedback and surveys for directional feedback.

Avoid over-intellectualized frameworks. Focus on concrete, operational questions: “Do players understand this mechanic? Where do they churn? What changes retention?”

Soft Launch: Let Data Drive Economics

You’re validating whether players will engage and if unit economics work. Test with real money in limited markets. Hundreds to thousands of players generating behavioral signals. Early retention via funnel analysis, A/B test features, pricing, and monetization. Analyze LTV, ARPU, retention cohorts, and spender behavior.

Signal to trust: Actual spending behavior, not stated willingness to pay.

Hard Launch: Pure Market Validation

You’re validating what creative drives cost-effective installs, ROAS, and the ability to drive full project economics. Test marketing across channels to identify where your audience resides and how far you can scale before encountering diminishing returns.

Who decides: Market response determines what works.

The progression: You start at the top of the pyramid (tastemaker vision) in pre-production, then gradually move down through production and soft launch, reaching the bottom (pure data) by the time of the hard launch. Match your methodology to your stage, validate your tastemaker early, and avoid intellectual theater that yields no operational results.

Takeaways

Validation is art, not science—and it requires the right artist. Novel concepts require tastemaker judgment (PUBG), but only if your tastemaker has good taste. Proven models require player data (League). Match your approach to your innovation level, then validate that your validators are actually good at validation.

Broken validation tests are invisible. Statistical significance doesn’t mean your methodology was sound. Question your incentives, sample, context, and interpretation. The most dangerous validation isn’t no validation—it’s broken validation that looks rigorous.

Avoid validation theater. Over-intellectualized frameworks and complicated personas often yield little in terms of operational value. Before investing in any validation methodology, ask: “What decision will this help us make? How will the answer change what we build?” Simple, concrete validation beats elaborate theoretical frameworks.

Quality beats quantity in early validation. Player feedback generates noise through conceptual gaps, persona mismatches, and articulation gaps. Five right players generate more signal than a thousand random ones.

Validation predicts interest, not execution. Even perfect validation can’t tell you if you’ll execute well. The gap between validated concept and delivered product determines outcomes. Combine smart validation with ruthless execution focus.

This week, I’m in Chennai, India, attending India’s largest game development conference, IGDC.

Five years ago, I cofounded a mobile gaming studio in India. People told me I was crazy then. And that the vision for what I wanted to try to do would be impossible.

Today, I’m asked how best to start a game studio in India. Furthermore, over the past five years, I have witnessed the growth of the local gaming player market, unlike anything I have ever experienced before.

Check out the data for yourself below!

In light of IGDC coming up, we talk about the two opportunity spheres associated with India in a two-part series:

1. The Indian Gaming Market (Part 1): The MASSIVE growth and opportunity presented by the booming local gaming market in India.

2. The Indian Game Studio Opportunity (Part 2): The opportunity for India to become the next China for game development, initially serving the burgeoning AA and eventually the AAA market for games.

The Massive Indian Gaming Market Opportunity - Part 1

Everyone saw India ban Real Money Gaming in August 2025. Almost nobody is tracking where $2 billion in annual gaming spend is about to flow.

While headlines focused on the massive impact on Dream11 and MPL (a collective loss of approximately $10 billion in market cap), a structural shift was happening underneath: 140 million paying gamers in India—20% of the total gaming population—suddenly lost their primary entertainment outlet. These weren’t casual players. They were users comfortable with UPI payments, habituated to daily gaming sessions, and proven to convert at rates Western publishers dream about.

The capital, the users, and the infrastructure don’t disappear when you ban a category. They redirect. And right now, they’re redirecting into free-to-play gaming at a scale that makes India the most underpriced gaming opportunity in global markets.

Where does the margin flow?

Executive Summary

India’s RMG ban eliminated billions in revenue but created a larger opportunity for F2P gaming. With 520 million gamers, 835 million internet users, a median age of 29, and ARPPU growing 9x since 2020, India’s digital gaming market is projected to more than double from $1.9 billion in FY25 to $4.3 billion by FY30. The combination of macro tailwinds—rising discretionary income, mobile-first infrastructure, and regulatory clarity—positions both global and local developers to capture users and spend fleeing the banned RMG category.

Do the current even further macro tailwinds drive the future potential of India’s gaming market even more?

India Didn’t Ban Gaming. It Banned One Business Model.

The August 2025 Promotion and Regulation of Online Gaming Act banned online money gaming nationwide, eliminating categories such as fantasy sports and skill-based card games, which accounted for 98% of India’s gaming revenue in FY23. The headlines treated this as a death blow to Indian gaming.

The data tells a different story.

Digital gaming (excluding RMG) grew at 31% CAGR from FY23 to FY25 even before the ban. Post-ban, growth is expected to accelerate, with the digital gaming market projected to reach $4.3 billion by FY30—an 18% CAGR from FY25. Meanwhile, the total Gaming and Interactive Media market is expected to grow even faster at 26% CAGR, reaching $7.8 billion by FY30.

This isn’t speculative. The structural reset is already visible in user behavior. According to BITKRAFT and Redseer’s October 2025 report, the share of paying gamers in India increased from 80 million in FY21 to 140 million in FY25, and is projected to hit 230 million by FY30. These users aren’t leaving gaming—they’re switching categories.

Where RMG required users to deposit money upfront and accept financial risk, F2P games entertain first and provide wider monetization flexibility. For 140 million users accustomed to daily gaming habits and mobile payment rails, F2P represents a lower-friction, higher-entertainment alternative. The infrastructure they built—app installs, payment methods, gaming routines—transfers directly.

The Macro Tailwinds Are Unprecedented

India’s gaming boom isn’t happening in isolation. It’s supercharged by macroeconomic shifts that dwarf the RMG ban in significance.

Consumption is surging. According to the report, Private Final Consumption Expenditure (PFCE)—the share of GDP driven by household spending—now accounts for 61% of India’s GDP. This matches the United Kingdom (61%) and approaches levels seen in consumption-driven economies like the United States (68%) and Brazil (64%), while far exceeding other major emerging markets like China (40%), Indonesia (55%), Saudi Arabia (45%), and UAE (46%), based on IMF and World Bank data cited in the report.

Translation: India’s economy runs on consumer spending, not just exports or manufacturing. This creates a foundation for discretionary purchases, such as gaming.

And that discretionary spending is accelerating. The report shows that discretionary spending as a percentage of monthly expenditure increased 5 percentage points over the past decade: from 24% to 29% in urban areas and from 22% to 28% in rural areas, based on India’s National Sample Survey Office (NSSO) data for 2011-12 and 2023-24.

Demographics favor digital entertainment. India’s median age is 29, with 51% of the population under the age of 30. Of 835 million internet users in FY25 (projected to reach 1.12 billion by FY30), 520 million play at least one game. That’s a 66% internet-to-gamer conversion rate—and it’s accelerating as Tier 2 and Tier 3 cities come online.

Infrastructure costs collapsed. India has the lowest mobile data costs globally at approximately $0.15 per GB—6.7x cheaper than the US and 12.7x cheaper than the UK. Smartphones in India cost roughly half the price of equivalent devices in the US and UK (which are both 1.9x more expensive). Meanwhile, Indians spend an average of 5 hours per day on smartphones—the third-highest globally behind Indonesia (6.3 hours) and Brazil (5.3 hours).

This combination—rising incomes, young demographics, cheap data, and high engagement—creates what BITKRAFT calls a “mobile-first ecosystem” where gaming becomes the default leisure activity for a generation coming of age entirely on smartphones.

Battle Royale Dominance Signals, Not Limits, the Opportunity

The Indian gaming market today looks highly concentrated. Battle Royale games—primarily BGMI (Battlegrounds Mobile India) and Free Fire—account for 47% of total gaming revenue in FY25. Simulation games dominate downloads at 22% of the top charts. Casual and hyper-casual games make up 54% of market share by segment.

This concentration isn’t a weakness. It’s a leading indicator of massive untapped diversification potential.

When one genre captures nearly half of all revenue in a market projected to double in five years, it signals that user demand is ahead of supply-side innovation. Battle Royale succeeded because it solved for India’s two dominant gaming motivations: socializing (cited by 40-50% of mobile gamers) and familiarity (50% of gamers in FY24 preferred local games with regional content).

As the market matures, developers who address other motivations—such as competition, progression, status, and creativity—will unlock new segments. The early signs are visible: hybrid casual games, which blend casual mechanics with mid-core progression systems, are the fastest-growing segment. Mid-core and core games are expected to grow from 41% of the market in FY24 to 54% by FY30.

Genre diversification historically follows monetization maturity. In India, ARPPU increased 9x from $3 in FY20 to $27 in FY25. In-app purchases (IAPs), which accounted for 38% of revenue in FY23, are projected to reach 66% by FY30, overtaking advertising as the primary monetization lever.

Translation: Indian gamers are increasingly willing to pay for progression, customization, and competitive advantage—the monetization mechanics that drive mid-core and core games globally. As developers layer these systems onto locally relevant IP and themes, the revenue per user will expand, and the genre mix will diversify.

Local vs. Global: Two Different Opportunities, Same Tailwind

The India opportunity bifurcates into two distinct plays:

For global developers, India represents a massive, undermonetized audience with proven payment behavior. The 520 million gamers in FY25 already exceed the entire US population. The 140 million paying gamers match the paying user base of many mature Western markets, but at 1/10th the ARPPU. As purchasing power increases and payment friction decreases (UPI adoption is near-universal), the ARPU gap will compress. Developers who localize for India now—local language interfaces, regionally relevant content, social features optimized for voice chat—will capture a disproportionate market share as the category scales.

For local Indian developers, the opportunity is even larger. According to BITKRAFT/Redseer analysis of Play Store download data, five of the top 10 games in India by downloads in FY24-25 were from Indian publishers: Ludo King (1 billion+ downloads), Zupee Ludo, Indian Bikes Driving 3D, WinZO Ludo, and Cricket League. These games succeeded by solving for cultural familiarity. Ludo King’s premise—a digitized version of a board game that every Indian household owns—required zero explanation. The social features (voice chat, multiplayer lobbies) mapped directly to how Indian families and friends already interacted.

[INSERT GRAPH: Top Games by Downloads and Revenue, Page 25 of report]

But here’s the underappreciated insight: over 80% of Indian game developers’ revenue comes from overseas markets. Studios like Gameberry Labs (Ludo Star), Moonfrog Labs (Ludo Club), and Nazara (CATS: Crash Arena Turbo Stars) generate the majority of their revenue outside India. They’ve proven they can compete globally on game design, live ops, and monetization.

The RMG ban removes the distraction. For five years, Indian gaming talent was pulled into building skill-based gambling mechanics optimized for user acquisition and retention on a 30-day timescale. Now that capital and talent can redirect toward building IP-driven, progression-based games with multi-year retention curves. The 500+ gaming studios founded in India over the past five years—many staffed by developers trained at global publishers—are now competing for the same user base that RMG vacated.

The result: a Cambrian explosion in supply-side experimentation, with both global and local developers racing to serve the same 520 million users.

Takeaways

• The RMG ban didn’t destroy demand—it redirected it. Approximately 140 million paying gamers, accustomed to daily gaming sessions and mobile payments, are transitioning from skill-based gambling to free-to-play (F2P) entertainment. Digital gaming is projected to grow from $1.9 billion (FY25) to $4.3 billion (FY30) at an 18% CAGR.

• Macro tailwinds are unprecedented. Discretionary spending has increased by 5 percentage points over the past decade. PFCE accounts for 61% of GDP, and India has 835 million internet users with a median age of 29. Data costs are the lowest globally, and smartphone usage averages 5 hours per day.

• Battle Royale’s 47% revenue share signals opportunity, not saturation. When one genre dominates revenue in a market projected to double, it reveals that user demand is ahead of supply-side innovation. As ARPPU grows 9x (from $3 in 2020 to $27 in 2025) and IAPs overtake ads by FY30, genre diversification into mid-core and core will accelerate.

• Local developers have a home-field advantage but global ambitions. Five of India’s top 10 games by downloads are from local publishers, but 80% of Indian developers’ revenue comes from overseas. Post-RMG, that talent is redirecting toward IP-driven, progression-based games that can compete both locally and globally.

• Localization is non-negotiable for global developers. 50% of Indian gamers prefer local games with regional content. Voice chat usage in BGMI is higher in India than in any other country. Developers who solve for cultural relevance, localized UI, and social features optimized for Indian behavior will capture outsized market share.

Next in Part 2: Why India could become a world-class game development hub similar to China—the rise of the White Tiger.

Sources: BITKRAFT Ventures & Redseer Strategy Consultants, “The Gaming and Interactive Media Opportunity in India,” October 2025.

Key Takeaways

The Problem: Game studios face an increasingly complex landscape when choosing analytics infrastructure. Traditional SDK analytics vendors promise ease of use but deliver vendor lock-in, unpredictable contract negotiations, and escalating costs. Building custom solutions offers control but requires dedicated engineering resources that most studios don’t have.

The Stakes: Your analytics stack determines how fast you can iterate, how well you understand your players, and ultimately how much you’ll pay to run your game’s data infrastructure at scale.

The Solution: The analytics landscape has fundamentally shifted. Modern warehouse-native platforms like PostHog now offer the best of both worlds—ease of implementation with data freedom—plus transparent, usage-based pricing that eliminates contract negotiations entirely.

Bottom Line: For new and scaling mobile game studios, PostHog represents the new meta: free to start and $100-700/month vs. $1,000-3,000+/month for SDK analytics vendors; real-time insights; built-in experimentation and feature flags; warehouse-native architecture; and true pay-as-you-go pricing without contracts.

The Analytics Infrastructure Dilemma

The game analytics landscape has historically been stratified by studio size and resources:

Small/Indie Studios defaulted to free tiers:

• Firebase Analytics or Unity Analytics (both free)

• Basic dashboards, limited customization

• “Good enough” until you outgrow them

Medium Studios graduated to SDK analytics vendors:

• Amplitude, Braze, CleverTap, or MixPanel ($1,000-5,000+/month)

• Better analytics, but vendor lock-in

• Expensive as you scale

Large Studios built custom infrastructure:

• In-house data engineering teams (5+ engineers)

• BigQuery or Snowflake warehouses

• Complete control, massive investment

But this rigid progression is now obsolete. New warehouse-native technology has collapsed these tiers, allowing studios to start with enterprise-grade infrastructure without the enterprise costs, complexity, or contract negotiations.

The old problem: You had to pick your tier upfront, and moving between tiers was painful. Outgrow Firebase? Migrate everything to Amplitude. Outgrow Amplitude? Rebuild everything on Snowflake.

Each transition meant:

• Months of engineering work

• Lost historical data or complex migrations

• Retraining teams on new tools

• Rewriting queries and dashboards

The new reality: Modern platforms like PostHog combine the ease of free-tier tools with the power of data warehouses—letting you start small and scale indefinitely without forced migrations, architectural rewrites, or contract negotiations.

Critical Issues Every Studio Must Consider

1. Reliability and Transparency

SDK analytics platforms like Amplitude, Braze, CleverTap, and Mixpanel have a dirty secret: they fail silently.

Marketing-first analytics platforms abstract away data processing, making debugging nearly impossible. When something breaks, you won’t know for weeks or months. As one engineer put it: “They are built to fail silently, and it takes weeks/months to notice these issues.”

What to look for:

• Open-source or transparent data processing

• Visible error logs and payload inspection

• Ability to see exactly how events are ingested and processed

2. Contract Negotiations vs. Pay-As-You-Go

Here’s a pain point many studios don’t realize until they’re already committed: most SDK analytics vendors require contract negotiations based on traffic estimates.

The traditional SDK analytics vendor model:

• Sales calls and demos before you can even start

• Estimate your MAU, events, or data volume for the year

• Lock into an annual contract at that rate

• Guess wrong? Pay for capacity you don’t use, or get hit with overage charges

• Want to scale up? Renegotiate mid-contract

• Want to scale down? Too bad, you’re locked in

This creates impossible decisions for game studios:

• Overestimate traffic: Waste budget on unused capacity

• Underestimate traffic: Face surprise overage bills when your game succeeds

• New game launch: No idea what traffic will be, forced to guess anyway

The new model (PostHog, Firebase):

• Start using immediately, no sales calls required

• Pay only for what you actually use

• Scale up or down automatically

• No annual commitments

• Transparent, public pricing

• Billing caps to prevent surprise costs

This isn’t just about flexibility—it’s about risk. When you’re launching a new game, you literally cannot predict traffic. Pay-as-you-go means you’re not penalized for success or punished for slower growth.

What to look for:

• Public, transparent pricing

• Usage-based billing (pay for actual events/users)

• No annual contracts or commitments

• Ability to set billing caps

• Self-service signup and start

3. Development Velocity and QA

Analytics is often integrated late in the development cycle and gets only a “basic QA pass.” This creates a dangerous pattern:

• Mistakes surface over longer time periods

• Issues remain undetected unless actively monitored

• The ability to flush data during development becomes critical

Without immediate data visibility, you’re flying blind during QA. Traditional SDKs batch events for battery optimization, creating 15-60+ minute delays before you can verify your implementation works.

What to look for:

• Instant event visibility in dev/test environments

• Manual flush capabilities

• Local event inspection without waiting for cloud sync

4. Dashboard Friction and Iteration Speed

Many SaaS analytics tools create artificial bottlenecks. For example, requiring account manager involvement for dashboard configuration slows down product exploration.

The ideal solution:

• Self-service dashboard creation

• Iterative experimentation without support tickets

• No-code insights with optional SQL for power users

5. The Real-Time Fallacy

Here’s a counterintuitive truth: real-time analytics is rarely helpful early on. What actually matters is:

• Data visibility - Can you see and export your data?

• Clean pipelines - Is your data schema well-organized?

• Easy analysis - Can non-engineers explore data without SQL?

Real-time dashboards are impressive demos but poor substitutes for robust data infrastructure.

6. Cost Predictability at Scale

SDK analytics vendors use pricing models designed to extract maximum value as you grow:

• Per-MTU (Monthly Tracked Users) fees that require annual estimates

• Per-event charges that compound

• Tiered feature access requiring enterprise plans

• Contract renegotiations as you scale

A studio with 100k MAU and 50M events/month might pay:

• SDK analytics vendors: $1,000-3,000/month (via annual contract, must estimate in advance)

• Firebase + BigQuery + Mode: $2,000-4,000/month (pay-as-you-go)

• Unity Analytics: $1,000-4,000/month (MAU-based tiers)

• PostHog: $500-2,000/month (pay-as-you-go, usage-based)

• Snowflake: $12,000-19,000/month (committed spend + overages)

The Five Architecture Options

Option 1: SDK Analytics Vendors (Amplitude, Braze, CleverTap, Mixpanel)

The Pitch: Plug-and-play behavioral analytics with rich dashboards, segmentation, and messaging.

Major Players:

• Amplitude - Market leader in product analytics (NASDAQ: AMPL), very popular with game studios

• Braze - Mobile engagement leader with analytics (NASDAQ: BRZE)

• CleverTap - Strong in mobile CRM + analytics, especially gaming and Asia markets

• Mixpanel - Major independent player, strong in tech and gaming

Strengths:

• Cross-platform SDKs for Unity, React Native, mobile, and PC

• Advanced funnels, retention analysis, and churn prediction

• Built-in push notifications and in-app messaging (CleverTap, Braze)

• Sophisticated segmentation and targeting

• Minimal data engineering required

• Integrated CRM + analytics workflows (CleverTap, Braze)

Critical Weaknesses:

• Contract negotiations required: Must estimate MAU/events and commit annually

• Unpredictable costs: Overage charges if you exceed estimates, wasted spend if you don’t

• Cost scales aggressively with users/events

• Data locked in vendor’s cloud (export restricted or paid)

• Limited transparency into data processing

• Silent failures and debugging challenges

• Vendor lock-in makes switching extremely difficult

Pricing Reality:

• Requires sales calls and contract negotiation

• Annual commitments based on estimated traffic

• Amplitude: Typically $1,000-3,000+/month for growth stage studios

• Mixpanel: Similar range, $1,000-2,500+/month

• CleverTap: $1,000-2,500+/month, includes CRM features

• Braze: $1,500-3,000+/month, premium pricing

• Overages can add 20-50% to costs

• Feature access is often gated by pricing tier

Best for: Teams prioritizing speed over long-term flexibility, with budget for premium SaaS solutions and confidence in traffic forecasts. Marketing-first teams who value integrated CRM + analytics.

Cost Reality: $1,000-3,000/month at 100k MAU (annual contract), potentially much higher at scale or with overages.

Option 2: Firebase Analytics + BigQuery + Visualization Tool

The Pitch: Free event tracking, warehouse power for custom analysis, Google ecosystem integration.

Stack Components:

• Firebase Analytics (free telemetry SDK)

• Google BigQuery (cloud data warehouse)

• Mode/Tableau/Looker (visualization layer)

• Firebase Remote Config for A/B testing

Strengths:

• Firebase Analytics is completely free for unlimited events

• BigQuery enables custom SQL, ML models, and complex analysis

• Native integration with Google Ads, AdMob

• Full data ownership and export capabilities

• Mature platform with extensive documentation

• Pay-as-you-go pricing: No contracts, pay only for queries and storage used

Critical Weaknesses:

• High complexity - requires 3 separate platform setups

• BigQuery’s nested JSON structures notoriously difficult

• 1-2 hour delay for event data (batched exports)

• Mode/Looker visualization licenses add $1,500-5,000/month

• Manual work required to activate segments/cohorts

• No built-in feature flags or advanced A/B testing

Pricing Reality:

• Firebase: Free

• BigQuery storage: $40/TB (~$50-200/month)

• BigQuery queries: $6.25/TB scanned (1TB free/month), typically $200-800/month

• Visualization: Mode Pro $1,500-2,500/month or Looker $3,000+/month

• No contracts required - usage-based billing

• Can scale costs down if traffic decreases

Best for: Studios with data engineering resources, already invested in Google Cloud, needing ML/advanced analytics.

Cost Reality: $2,000-4,000/month (pay-as-you-go), plus engineering overhead.

Option 3: Unity Analytics

The Pitch: One-click Unity-native analytics, remote config, A/B testing, and LiveOps.

Stack Components:

• Unity Analytics (telescoped, Unity-native)

• Unity Remote Config

• Unity Cloud Messaging

• Unity Gaming Services ecosystem

Strengths:

• Seamless Unity integration (literally 1-click setup)

• Non-developers can run A/B tests and remote config

• Built-in push messaging and LiveOps tools

• Free tier up to 50k MAU

• Role-based UI designed for PMs and designers

• Clear tier-based pricing: No contract negotiations for standard tiers

Critical Weaknesses:

• Unity-only - doesn’t support React Native, web, bare metal games

• Steep pricing tiers: Free to 50k MAU, then jumps significantly

• No SQL access or custom BI

• Limited data export (data “trapped” in Unity Cloud)

• No advanced journey analysis or session replay

• Limited CRM integration

Pricing Reality:

• Free: Up to 50k MAU

• Standard tier: Reported $1,000-4,000/month for 100k+ MAU

• MAU-based pricing can scale steeply

• Self-service for standard tiers, enterprise requires negotiation

Best for: Exclusively Unity-focused studios under 50k MAU where non-engineers need direct dashboard access.

Cost Reality: Free to 50k MAU, then $1,000-4,000/month with steep MAU-based scaling.

Option 4: PostHog - The New Meta

The Pitch: Open-source, all-in-one product analytics, experimentation, feature flags, and session replay built on a streaming warehouse (ClickHouse).

Stack Components:

• PostHog SDK or HTTP API (universal ingestion)

• ClickHouse (built-in warehouse)

• PostHog dashboards (no-code + SQL)

• Feature flags and experimentation engine

• Optional integrations (Braze, CleverTap, etc.)

Strengths:

• All-in-one platform: Analytics + A/B testing + feature flags + session replay + CDP

• True pay-as-you-go: No contracts, no sales calls, no traffic estimates required

• Transparent pricing: Public pricing calculator, pay only for events used

• Billing caps: Set maximum spend to prevent surprise costs

• Real-time data: Events available instantly (no batch delays)

• Warehouse-native: Direct ClickHouse storage with SQL access (HogQL)

• No-code + SQL: Power users get SQL, everyone else gets intuitive dashboards

• Transparent: Open-source with code-level visibility

• Developer-friendly: Immediate data flushing in dev/test environments

• Best cost at scale: $0.00005-0.00002/event (1M events free/month)

• Data freedom: Easy export to any destination, no lock-in

• Advanced experimentation: Multivariate testing, sequential testing, automatic significance

LiveOps Features:

• Behavioral segmentation and cohorts

• Feature flags for targeted offers and personalization

• A/B/n testing for game mechanics, pricing, UI

• Session replay for debugging player issues

Weaknesses:

• No built-in CRM or push notifications (requires Braze/CleverTap integration)

• Unity SDK requires custom HTTP implementation (1-2 days work)

• No native mobile attribution (use Adjust/AppsFlyer separately)

• Some operational setup required

Pricing Reality:

• No contracts or negotiations: Start using immediately

• 1M events free/month: Enough for 10-30k MAU to start

• Usage-based beyond free tier:

  • 1-2M events: $0.00005/event

  • 2-15M events: $0.0000343/event

  • Scales down further at volume

• Transparent calculator: See exact costs before starting

• Billing caps available: Set maximum monthly spend

• Scale up or down automatically: Pay only for what you use

Example costs:

• 10M events/month: $50-400

• 50M events/month: $1,000-1,500

• 100M events/month: $1,750-2,000

Best for: Cross-platform studios (Unity + React Native + PC), teams wanting long-term flexibility, anyone prioritizing data freedom and predictable costs.

Cost Reality: $100-700/month starting out, $500-2,000/month at 100k MAU/50M events (pay-as-you-go, no contracts).

Option 5: Snowflake Data Warehouse

The Pitch: Enterprise-grade data warehouse with unlimited scalability and multi-cloud support for studios that need complete customization.

Stack Components:

• Custom telemetry collectors → Snowflake

• Snowflake data warehouse (compute + storage)

• Mode/Looker/Tableau for visualization ($1,500-5,000/month)

• RudderStack/Hightouch for CDP/activation ($500-2,000/month)

• Custom-built analytics, segmentation, and activation

Strengths:

• Ultimate scalability: Handles billions of events without breaking a sweat

• Multi-cloud support: AWS, Azure, GCP - run anywhere

• Best-in-class performance: Fastest queries at massive scale

• Unlimited customization: Build exactly what you need with SQL

• Enterprise compliance: SOC 2, HIPAA, GDPR certified

• Data governance: Fine-grained access controls and audit logs

• Future-proof: Integrate any data source, run any model

Critical Weaknesses:

• Extreme cost: $12,000-19,000/month minimum realistic spend

• Requires dedicated data team: Need 5+ full-time data engineers

• Annual commits typically required: Snowflake credits purchased annually

• Complex pricing: Credits, storage, compute all separate

• 3-6 month implementation: Significant time investment before value

• Everything is custom: Analytics, segmentation, activation - all built from scratch

• Massive overkill: 99% of game studios don’t need this

• No built-in LiveOps: Must build experimentation, feature flags, messaging separately

• Operational complexity: Cost optimization requires constant monitoring

Pricing Reality:

• Snowflake typically requires annual credit purchases

• Credits cost $2-4 each (on-demand rates)

• Medium warehouse = 4 credits/hour = $8-16/hour

• Running 12 hours/day = $2,880-5,760/month just for compute

• Storage: $40/TB (~$50-200/month)

• Plus visualization tools, CDP tools, engineering staff

• Total minimum: $12,550-19,200/month

• Realistic costs: $15,000-25,000/month

Cost Breakdown (100k MAU example):

• Snowflake compute: $500-2,000/month (committed credits)

• Snowflake storage: $50-200/month

• Visualization tool (Mode Pro or Looker): $1,500-5,000/month

• CDP/activation tool: $500-2,000/month

• Data engineering staff costs: $10,000+/month

• Total: $12,550-19,200/month minimum

Reality Check: This is EA/Activision/Riot-scale infrastructure.

Best for: Large studios (500k+ MAU, $50M+ revenue) with 5+ data engineers who need multi-source data integration beyond game events, or studios with specific enterprise compliance requirements.

Cost Reality: $12,550-19,200/month minimum (annual commits), realistically $15,000-25,000/month at scale.

The Comparison: Side by Side

The Contract Negotiation Tax: A Hidden Cost

Let’s talk about something SDK analytics vendors don’t advertise: the time and opportunity cost of contract negotiations.

The Traditional SDK Analytics Vendor Sales Cycle:

Week 1-2: Initial Contact

• Inbound inquiry or outbound sales contact

• Schedule discovery call

• Demo presentation

• “What’s your expected MAU/traffic?”

Week 3-4: Scoping and Pricing

• Estimate your traffic for the next 12 months (impossible for new games)

• Get custom quote based on estimates

• Negotiate features, support tier, overages

• Legal review of contract terms

Week 5-6: Approval and Procurement

• Internal budget approval

• Finance/legal review

• Contract signing

• Payment processing

Week 7: Finally Start Using

• Technical integration begins

• 6+ weeks after you first wanted analytics

The Real Cost:

• 6+ weeks delayed insights while you negotiate

• Opportunity cost: Decisions made without data

• Risk: Locked into estimates that might be completely wrong

• Pressure: Sales tactics to commit to higher tiers “just in case”

The PostHog/Firebase Pay-As-You-Go Cycle:

Day 1:

• Sign up (5 minutes)

• Start integrating

• Data flowing same day

That’s it. No calls, no negotiations, no estimates, no contracts.

The Business Impact

For a new game launch, this difference is massive:

Scenario: You’re 6 weeks from launch

Traditional SDK Analytics Vendor Path:

• Week 1-6: Sales cycle and negotiations

• Launch day: Just now starting integration

• Post-launch: Scrambling to get analytics working while managing live issues

• Result: Flying blind during your most critical window

Pay-As-You-Go Path:

• Day 1: Start integration

• Week 2: Analytics fully integrated and tested

• Week 3-6: Iterating on events, testing dashboards, QA

• Launch day: Full analytics already running smoothly

• Result: Data-driven decisions from day one

The contract negotiation tax is real: 6 weeks of delayed insights, opportunity cost, and risk. For fast-moving game studios, this alone can justify choosing pay-as-you-go solutions.

The Integrated Platform Debate: All-in-One vs. Best-of-Breed

A common counterargument: “Platforms like CleverTap and Braze offer analytics + CRM + messaging in one integrated system. Doesn’t that eliminate the data sync problems you’d have with PostHog + separate CRM?”

The short answer: It depends on your priorities.

The Case for Integrated Platforms (CleverTap, Braze)

Valid strengths:

• No data syncing - Analytics and messaging share the same data store

• Real-time reactivity - Trigger campaigns instantly on user behavior

• Operational simplicity - One vendor, one contract, one platform to learn

• Built-in workflows - Drop-off detection → segment creation → campaign trigger happens seamlessly

• Eliminates mismatch issues - No discrepancies between analytics and CRM data

• Faster time-to-campaign - From insight to action in one platform

When this makes sense:

• Your primary need is CRM and lifecycle marketing

• Analytics is secondary to messaging

• You want plug-and-play simplicity over customization

• Your team is marketing-focused, not data-focused

• You value operational convenience over data ownership

There’s real value in an integrated system where analytics informs messaging without data syncing. The “sum is greater than the parts” argument holds if your primary job-to-be-done is lifecycle marketing.

The Case for Best-of-Breed (PostHog + CRM)

Valid strengths:

• Superior analytics - Dedicated analytics platforms are better at analytics than CRM platforms

• Data ownership - Your data lives in a warehouse you control

• Flexibility - Swap CRM providers without losing analytics history

• Advanced capabilities - Better experimentation, session replay, SQL access, feature flags

• Cost efficiency - Often cheaper to combine best tools than buy all-in-one

• Avoid double lock-in - Not locked into both analytics AND CRM with one vendor

Modern CDC minimizes sync issues:

• PostHog’s CDP can push events to CleverTap/Braze in near real-time

• Event streaming (not batch ETL) means minimal lag (seconds, not hours)

• Data mismatches are rare with proper implementation

• Real-time reactivity is still possible with modern integrations

When this makes sense:

• Your primary need is deep analytics and experimentation

• You want data ownership and warehouse-native architecture

• You’re okay with 2 platforms instead of 1

• You value flexibility to switch CRM providers

• You have engineering resources to manage integrations

The Real Question: What’s Your Primary Job-to-Be-Done?

If your primary goal is lifecycle marketing: → CleverTap/Braze integrated platforms make sense → You’re buying CRM that happens to include analytics → Accept the trade-offs on analytics depth and data ownership

If your primary goal is understanding player behavior and experimentation: → PostHog + CRM makes sense → You’re buying analytics that happens to integrate with CRM → Accept the trade-off of managing two platforms

The Lock-In Factor

Here’s where the debate gets interesting: vendor lock-in is worse with integrated platforms.

With CleverTap/Braze (integrated):

• Your analytics, segmentation, AND messaging are locked in

• Switching means rebuilding all three

• Data export is limited

• You’re locked in at multiple levels

• Historical analytics data may not be exportable

With PostHog + CleverTap (best-of-breed):

• Your analytics lives in PostHog (warehouse-native, exportable)

• Your CRM is CleverTap (can swap to Braze later)

• If you switch CRM, your analytics history is preserved

• You’re locked in at only one level

• Future flexibility is preserved

The Contract/Pricing Critique Still Applies

Even accepting the “integrated is better” argument for CRM-focused teams, the contract negotiation and pricing model issues remain:

• Still requires sales calls and annual contracts

• Still requires traffic forecasting

• Still has overage risks if you guess wrong

• Still lacks transparent, public pricing

PostHog’s pay-as-you-go model applies to the analytics layer whether you integrate with CleverTap or not.

Our Recommendation

For marketing-first teams:

• Consider CleverTap or Braze for integrated CRM + analytics

• Accept the trade-offs on analytics depth and data ownership

• Still evaluate contract terms carefully

• Good fit if lifecycle campaigns are your primary KPI

For data-first teams:

• Start with PostHog for analytics + experimentation

• Add CRM (CleverTap/Braze/others) when needed

• Accept the trade-off of managing two platforms

• Good fit if player behavior insights are your primary KPI

For most game studios:

• PostHog + CRM is the better long-term architecture

• The data ownership and flexibility benefits outweigh integration convenience

• Modern CDC tools make the “sync problem” largely solved

• You’re optimizing for 2-3 year horizon, not immediate convenience

Choose based on whether you’re analytics-first or marketing-first.

The Recommendation: Think in Phases

For New Studios (0-50k MAU)

Winner: PostHog

Start with PostHog’s free tier (1M events/month) to validate product-market fit. You get:

• Complete analytics infrastructure

• Experimentation and feature flags

• Session replay for debugging

• Real-time insights

• No upfront costs

• No contracts or negotiations

• Pay only if you exceed free tier

Implementation: 1-2 days to build Unity HTTP integration, immediate value.

Month 1-6 Roadmap:

1. Week 1-2: Core analytics (DAU/MAU, retention, revenue)

2. Week 3-4: A/B testing setup (pricing, gameplay, UI)

3. Month 2: Segmentation and personalization

4. Month 3: CRM integration (add CleverTap or Braze)

Why not the others?

• SDK analytics vendors: Unnecessary cost when bootstrapping, plus sales cycles delay launch

• Firebase + BigQuery: Too complex without data engineers

• Unity Analytics: Viable alternative if Unity-only, but less flexible long-term

• Snowflake: Absurdly expensive overkill for early stage

For Growing Studios (50k-500k MAU)

Winner: Still PostHog

At this stage, PostHog scales efficiently while competitors get expensive:

• PostHog: $1,000-1,500/month for 100M events (pay-as-you-go)

• SDK analytics vendors: Often $3,000-5,000+/month (annual contracts)

• Firebase + BigQuery: $3,000-5,000/month (pay-as-you-go)

• Unity Analytics: $2,000-5,000/month (steep scaling)

• Snowflake: $15,000-25,000/month (still overkill)

Key additions:

• Add CleverTap or Braze for CRM ($500-1,500/month)

• Add AppsFlyer or Adjust for attribution ($500-1,000/month)

• Total stack: $2,000-4,000/month (still cheaper than SDK analytics vendors alone)

• All pay-as-you-go: No contracts, scale with actual usage

Why PostHog wins here:

• Real-time LiveOps becomes critical at scale

• Segmentation and experimentation drive revenue

• Data freedom lets you build custom models

• Warehouse-native design prevents future migration pain

• No contract renegotiations as you scale

When to consider Firebase + BigQuery:

• You’ve hired data engineers

• You’re heavily invested in Google Cloud ecosystem

• You need custom ML models beyond PostHog’s capabilities

For Large Studios (500k+ MAU)

Winner: PostHog or Snowflake (depending on complexity)

At enterprise scale, you have three viable paths:

Path A: PostHog + Data Lake (Most Studios)

• PostHog for operational analytics and experimentation

• Export to S3/BigQuery for ML and long-term analysis

• Cost: $2,000-5,000/month (pay-as-you-go)

• Best for: Studios focused on games, not data infrastructure

Path B: Firebase + BigQuery + Custom Tools

• Full data engineering team (2-3 people)

• Custom ML pipelines and predictive models

• Cost: $5,000-10,000/month (pay-as-you-go)

• Best for: Google Cloud-native studios with existing data teams

Path C: Snowflake + Full Custom Stack

• Dedicated data engineering team (5+ people)

• Multi-source data integration (not just game events)

• Cost: $15,000-25,000/month (annual commits required)

• Best for: Multi-game portfolios, complex enterprise requirements

When to choose Snowflake:

• You have 500k+ MAU across multiple games

• You need to integrate data beyond game events (finance, customer support, web, etc.)

• You have 5+ dedicated data engineers

• You require enterprise compliance (HIPAA, SOC 2)

• Your analytics budget is $150k+/year

• You’re a public company or planning to IPO

• You’re comfortable with annual commit negotiations

Reality check: Snowflake is for EA/Activision/Riot scale. If you’re asking “should we use Snowflake?”, the answer is probably no. Studios that need Snowflake already know they need it.

Why Traditional SDK Analytics Vendors Are Losing

The analytics market has fundamentally shifted. PostHog represents what we might call “analytics 2.0”—combining the ease of SDK analytics vendors with the power and flexibility of warehouse-native architectures, plus transparent, contract-free pricing.

The Old Model (SDK Analytics Vendors):

1. Easy integration

2. Black box processing

3. Vendor lock-in

4. Escalating costs

5. Data silos

6. Contract negotiations

7. Annual commitments

8. Traffic forecasting required

9. Limited to what vendor offers

The New Model (PostHog):

1. Easy integration

2. Transparent processing

3. Data freedom

4. Predictable costs

5. Warehouse-native

6. No contracts

7. Pay-as-you-go

8. No forecasting needed

9. Extensible and customizable

SDK analytics vendors made sense when warehouses were expensive and complex. In 2025, with ClickHouse and modern data infrastructure, that trade-off no longer makes sense. The contract negotiation overhead is the final nail in the coffin.

Special Case: When Each Option Makes Sense

Choose SDK Analytics Vendors (Amplitude, Braze, CleverTap, Mixpanel) if:

• ✅ You’re primarily CRM/marketing-focused and want integrated analytics + messaging (CleverTap, Braze)

• ✅ You’re already locked in and migration costs are prohibitive

• ✅ You specifically need their proprietary ML models

• ✅ Budget isn’t a primary concern

• ✅ You can accurately forecast traffic for annual contracts

• ✅ You value operational simplicity over data ownership

Reality: These still have a place for marketing-first teams, but very few new data-focused studios should choose this path. The contract negotiation overhead alone makes it unattractive for analytics-first use cases.

Choose Firebase + BigQuery + Mode if:

• ✅ You have dedicated data engineers (2+ people)

• ✅ You’re already deep in Google Cloud ecosystem

• ✅ You need custom ML beyond what PostHog provides

• ✅ You want Google Ads/AdMob attribution integration

• ✅ You prefer pay-as-you-go pricing without contracts

Reality: Good choice for Google-native studios with engineering resources.

Choose Unity Analytics if:

• ✅ You’re building exclusively Unity games (no cross-platform)

• ✅ You’re under 50k MAU (free tier is excellent value)

• ✅ Non-engineers need direct dashboard and LiveOps access

• ✅ You want minimal setup friction

• ✅ You’re okay with tiered pricing as you scale

Reality: Solid option for Unity-only indies and small studios.

Choose PostHog if:

• ✅ You’re building cross-platform games (Unity + React Native + PC)

• ✅ You want long-term data ownership and flexibility

• ✅ You need real-time analytics and experimentation

• ✅ You value cost efficiency at scale

• ✅ You want warehouse-native architecture

• ✅ You want to avoid contract negotiations and sales cycles

• ✅ You prefer transparent, usage-based pricing

• ✅ You don’t want to forecast traffic for annual commits

• ✅ Analytics and experimentation are your primary focus

Reality: Best choice for 80% of mobile game studios, especially data-first teams.

Choose Snowflake if:

• ✅ You have 500k+ MAU across multiple games

• ✅ You have 5+ dedicated data engineers

• ✅ You need multi-source data integration (beyond game events)

• ✅ You require enterprise compliance (HIPAA, SOC 2)

• ✅ Your analytics budget is $150k+/year

• ✅ You’re optimizing for ultimate customization over cost

• ✅ You’re comfortable negotiating annual credit purchases

Reality: Enterprise-only option. If you need to ask, you don’t need it yet.

Real-World Cost Scenarios

Let’s look at what each option actually costs as your studio grows, including the impact of contracts:

Scenario 1: Indie Studio (10k MAU, 5M events/month)

Winner: PostHog or Unity Analytics (if Unity-only)

Contract Impact: SDK analytics vendors require annual commit even at this tiny scale, creating unnecessary risk.

Scenario 2: Growing Studio (100k MAU, 50M events/month)

Winner: PostHog (best cost-to-value plus pricing flexibility)

Contract Impact:

• SDK analytics vendors lock you into $24k-48k/year

• If your game doesn’t hit projections, you’re still paying

• If your game exceeds projections, you’re renegotiating mid-contract

• PostHog scales naturally with actual usage

Scenario 3: Successful Studio (500k MAU, 200M events/month)

Winner: PostHog for cost-efficiency, Snowflake if you need ultimate customization and have the team

Contract Impact:

• SDK analytics vendors: $60k-120k/year commitment

• Snowflake: $180k-300k/year in credits

• PostHog: Pay actual usage monthly, no commitments

Scenario 4: Enterprise Studio (2M+ MAU, 1B+ events/month)

Winner: PostHog for most use cases, Snowflake for multi-game portfolios with complex data integration needs

Contract Impact:

• Even at this scale, PostHog’s pay-as-you-go is competitive

• Contract negotiations become board-level decisions

• Annual commits of $200k-600k+ for SDK analytics vendors or Snowflake

The Migration Question

“We’re already on [SDK analytics vendor]. Should we migrate?”

Migration is expensive, but vendor lock-in and contract commitments are expensive too. Here’s when to consider it:

Migrate to PostHog if:

• Your current bill is over $3,000/month

• You’re frustrated by contract renewal negotiations

• You’re blocked by data export limitations

• You need real-time capabilities

• You want to add experimentation/feature flags

• Your team is frustrated by black-box debugging

• Your annual contract is coming up for renewal

Migration timeline: 4-8 weeks for parallel tracking and validation.

Best time to migrate: 3-6 months before your annual contract renewal, giving you time for parallel testing and negotiation leverage.

Stay put if:

• Your current solution works and is under $2,000/month

• Migration would distract from core product work

• You’re deeply integrated with vendor-specific features

• Your contract has less than 6 months remaining (wait it out, then decide)

Implementation Roadmap: Your First 90 Days with PostHog

Phase 1: Core Analytics (Week 1-2)

Goal: Ship basic telemetry and dashboards

Tasks:

• Build custom Unity HTTP client → PostHog API

• Implement core events: app_opened, level_completed, purchase, session_start

• Add user properties: subscription_tier, level, total_spend

• Create dashboards: DAU/MAU, retention, revenue

Deliverable: Real-time visibility into core metrics

Cost: Still within free tier for most new studios

Phase 2: Experimentation (Week 3-4)

Goal: Enable A/B testing for key decisions

Tasks:

• Integrate PostHog feature flags SDK

• Test pricing experiments (different offer amounts)

• Test gameplay (difficulty curves, tutorial variations)

• Test UI (button colors, layouts)

Deliverable: Data-driven product decisions

Cost: Feature flags included, may push beyond free tier ($50-200/month)

Phase 3: Segmentation & Personalization (Month 2)

Goal: Build targeted player experiences

Tasks:

• Create cohorts: whales, dolphins, minnows, churning players

• Build targeted offers using feature flags

• Implement dynamic difficulty based on segments

• Personalized content recommendations

Deliverable: Increased monetization through personalization

Cost: $200-500/month as event volume grows

Phase 4: CRM Integration (Month 3)

Goal: Complete marketing automation

Tasks:

• Add CleverTap or Braze for push notifications (~$500-1,500/month)

• Set up PostHog → CleverTap event sync

• Build lifecycle campaigns: onboarding, retention, re-engagement

• Implement win-back campaigns for churned players

Deliverable: Full LiveOps capability

Total stack cost: $700-2,000/month (still no contracts)

The Bottom Line

The game analytics landscape has matured beyond the old SDK analytics vendor vs. custom build dichotomy. PostHog represents a third way: warehouse-native analytics that’s as easy to implement as traditional SDKs but offers the flexibility and cost efficiency of owning your data, plus transparent pay-as-you-go pricing without contracts.

For most mobile game studios:

• Start with PostHog’s free tier

• Scale up as you grow ($500-2,000/month at 100k MAU)

• Add CRM (CleverTap/Braze) when needed

• Export to data lakes if you reach enterprise scale

• Never negotiate a contract or forecast traffic

Avoid traditional SDK analytics vendors unless:

• You’re primarily marketing-first and want integrated CRM + analytics (CleverTap, Braze)

• You’re already locked in and migration costs are prohibitive

• You specifically need their proprietary ML models

• You’re comfortable with annual contract negotiations and traffic forecasting

Avoid Firebase + BigQuery unless:

• You have dedicated data engineers

• You’re already deep in Google Cloud

• You need custom ML beyond what PostHog provides

Avoid Unity Analytics unless:

• You’re Unity-only for the next 2-3 years

• You’re under 50k MAU (free tier is excellent)

• Non-engineers need direct LiveOps control

Avoid Snowflake unless:

• You’re at 500k+ MAU with multiple games

• You have 5+ data engineers

• You need multi-source data integration beyond game events

• Your analytics budget is $150k+/year

• You’re prepared for annual credit commitments and enterprise procurement cycles

The new meta is clear: warehouse-native, open-source, real-time analytics with built-in experimentation and transparent, contract-free pricing. PostHog hits this mark. Traditional SDK analytics vendors serve a role for marketing-first teams, but are increasingly obsolete for analytics-first use cases. Snowflake is powerful but overkill for 99% of studios.

Your data infrastructure should accelerate your studio, not constrain it with contracts and negotiations. Choose accordingly.

What analytics infrastructure is your studio using? Disagree? Let me know! Share your experiences in the comments.

Your design team churns out disconnected paper designs. Engineering over-engineers solutions that miss the actual need. Art generates assets without validation, leading to endless iteration and performance issues. QA receives broken builds with unclear requirements, producing bug reports that may or may not matter.

Here’s the uncomfortable truth: It’s probably not a talent problem. It’s a structural problem.

Chris Casanova has seen this pattern repeat across Microsoft Xbox, Mojang, Relic Entertainment, Offworld Industries, and, most recently, Hypixel Studios. And after watching studios shut down and games fail—not from lack of skilled developers, but from organizational friction—he wrote a manifesto that’s challenging how the entire industry thinks about team structure.

The core insight? We’re optimizing for craft silos when we should be optimizing for outcomes.

The Hidden Cost of “Craft-Based” Teams

Traditional game studios are organized by discipline: a centralized design department, an engineering department, or an art department. On paper, it makes sense. In practice, it’s a friction factory.

Here’s what actually happens:

• Design creates specs without engineering context → over-engineered solutions that miss the mark

• Art produces assets without real-time validation → wasted work and performance nightmares

• QA gets tossed half-baked builds → mountains of irrelevant bug reports

• Every handoff loses fidelity → the original intent dies in translation

“So many times I’ve seen different teams where there’s so much friction getting in the way of great people. And so much of the root cause is a bad structure. As leaders, we focus on symptoms and not the root cause.” — Chris Casanova

The symptoms look like:

• Endless iteration cycles

• Communication breakdowns

• Misaligned priorities

• Teams working at cross-purposes

• Features that don’t meet the actual need

But here’s what most studios miss: This isn’t a people problem. It’s a design problem. And just like we iterate on game design, we need to iterate on org design.

The Military Already Solved This Problem

There’s a compelling analogy from Jocko Willink’s Extreme Ownership: Team Boat 6 in Navy SEAL training. They swapped the team leaders of the last-place and first-place boat crews. The last-place team became first place.

Not because the rowers changed. Because the structure and leadership changed.

Chris referenced Team of Teams by General Stanley McChrystal:

“It was so interesting to read that book and hear the similarities between how a squad or fire team in the military is cross-functional, cross-discipline, craft across craft.”

Both books arrive at the same insight: Small, cross-functional units with clear missions and authority to execute vastly outperform larger, specialized departments coordinating through hierarchy.

The military figured this out in Iraq and Afghanistan. Navy SEALs prove it in training. Game studios? We’re still fighting with the old model.

The “Outcome Team” Framework: What Actually Works

Chris advocates for what he calls “outcome-based teams” (others call them cross-functional, cross-disciplinary, or squad-based). Here’s the core structure:

Team Composition: 5-9 People

This isn’t arbitrary. Research on psychological safety shows that 5-9 people is the sweet spot where:

• Everyone can contribute meaningfully

• Communication overhead stays low

• Team cohesion remains strong

• Decision-making is fast

“When you get to 10, 12, 15 people on a team, that’s when you start to see cliques form. People start to fade into the background. That’s when you need to split.”

Cross-Disciplinary by Design

Each team should have everything they need to achieve their outcome:

• Designer(s)

• Engineer(s)

• Artist(s)

• Producer

• QA embedded (not a separate function)

The key principle: Minimize handoffs. Maximize collaboration.

Clear Mission and Charter

Every team needs to know:

• What is their outcome? (Not just “build a feature” but “achieve this player experience”)

• Why does it matter? (How it ladders to the game vision)

• What are the success metrics? (Specific, measurable KPIs)

Real Example: The Procedural Worlds Team

Chris shared a concrete example from his work on procedurally generated games:

Team Mission: “Every new world seed feels fresh, readable, and performant. Traversal and discovery naturally surface points of interest, resources, and risks.”

The KPI: Biome novelty per minute/hour — How often does a player experience something genuinely new?

The Team:

• Procedural generation engineers

• World designers

• Environment artists

• Technical artists (for performance)

• Embedded QA

This team has everything they need to own procedural generation from concept to shipping. No handoffs to a separate “tools team,” “art department,” or “QA department.”

The result? They can iterate rapidly, validate in real-time, and actually ship features that work.

Compare this to the traditional model:

1. Design writes spec for procedural worlds

2. Engineering builds tech (maybe over-engineers)

3. Art creates assets (maybe they don’t perform)

4. QA tests (finds issues late)

5. Everyone iterates (slowly, painfully)

Traditional model timeline: 6-12 months of painful iteration

Outcome team timeline: 2-4 months with continuous validation

The Three Layers of Team Organization

Chris recommends structuring your studio in three layers:

Layer 1: Product Pillars (3-7 per game)

These are your major experiential areas. For a co-op shooter, this might be:

• Combat Systems

• Progression & Economy

• Social & Co-op Features

• Live Operations

• Technical Foundation

Layer 2: Outcome Teams (multiple per pillar)

Each pillar contains 2-5 outcome teams. For Combat Systems:

• Weapons & Gunplay Team

• Enemy AI & Behavior Team

• Abilities & Special Moves Team

Layer 3: Individuals (5-9 per team)

Each outcome team has 5-9 people with mixed disciplines, focused on specific outcomes.

“This gives clarity. Everyone knows what pillar they’re supporting, what outcome they’re driving, and who their teammates are.”

The Leadership Challenge: Why This Is Hard

Here’s where it gets uncomfortable. Chris is blunt about why studios resist this structure:

The Territorial Leader Problem

The scenario: A Lead Engineer who’s built their identity around managing a 30-person engineering department.

Shifting to outcome teams means their empire shrinks. They go from managing 30 people to maybe being on one team of 8 people, or leading 2-3 engineers across multiple teams.

“I’ve seen so many cases where someone’s identity is I’m the lead of this entire department. And then you tell them, actually, we’re going to distribute you across these teams. That is a very tough conversation.”

The hard truth: Some leaders won’t make the transition. They’ll resist, politick, or self-select out. And that’s okay—because holding onto territorial structures to preserve egos is how studios fail.

The “Craft Excellence” Pushback

Another common objection: “But who maintains our craft standards? Who ensures engineering excellence if engineers are distributed?”

Chris’s answer: Communities of practice.

Engineers still meet regularly to discuss patterns, standards, and technical debt. Artists still have art reviews. But day-to-day, they work embedded in outcome teams.

This is how Spotify, Netflix, and Google organize. Gaming isn’t special—we just think we are.

The Political Reality

The biggest barrier isn’t technical. It’s political.

“If you have a c-suite that’s divided on this or a board that’s divided, you’re not going to be able to make this change. You need unified leadership saying, ‘This is the direction we’re going.’”

This requires:

• Executive alignment

• Willingness to have tough conversations

• Accepting that some people will leave

• Committing to the long-term health of the studio over short-term comfort

What About Small Teams?

Question that came up: “What if you’re only 10 people total?”

Chris’s answer: Don’t overthink it.

At 10 people, you already have the benefits of a small team—transparency, fast communication, scrappiness. You don’t need to create sub-teams.

Instead:

• Ensure everyone understands the product vision

• Maybe split into two mini-teams focused on in-game vs. out-of-game

• Track progress against clear outcomes

• Keep the structure flat and collaborative

The principle matters more than the prescription. Apply the framework to your context, don’t follow it blindly.

The AA Co-Op Shooter Example

Chris provided a specific organizational structure for a hypothetical AA co-op shooter with 70 people:

Product Pillars (6 total):

1. Combat & Gameplay Systems (15 people, 2 teams)

   • Weapons & Abilities Team (8)

   • Enemy AI & Combat Flow Team (7)

2. Progression & Economy (10 people, 2 teams)

   • Player Progression & Rewards Team (5)

   • In-Game Economy & Shop Team (5)

3. Social & Co-Op Features (12 people, 2 teams)

   • Matchmaking & Lobbies Team (6)

   • Co-Op Mechanics & Communication Team (6)

4. Content & Live Ops (15 people, 2 teams)

   • Seasonal Content Team (8)

   • Events & Limited-Time Modes Team (7)

5. World & Narrative (10 people, 1 team)

   • Environment, Story, & Mission Design Team (10)

6. Technical Foundation (8 people, 1 team)

   • Platform, Infrastructure, & Tools Team (8)

Key insights from this structure:

• Combat is the priority → Gets the most people (15) and focus

• Technical foundation is lean → 8 people supporting 70 (strong engineering culture)

• Each team has clear outcomes → Not “build a feature” but “deliver this player experience”

• Natural collaboration points → Teams that need to work together are obvious

• Scales well → Can add teams within pillars as needed

Compare this to a traditional structure:

Traditional:

• 20 engineers (reporting to Lead Engineer)

• 15 artists (reporting to Art Director)

• 12 designers (reporting to Creative Director)

• 10 QA (reporting to QA Manager)

With this structure, every feature crosses all four departments. Handoffs everywhere. Misalignment guaranteed.

The Implementation Roadmap

So you’re convinced. How do you actually make this change? Chris offers practical advice:

Step 1: Get Executive Alignment

This is non-negotiable. If your C-suite isn’t unified, don’t bother. You’ll just create chaos.

• Present the case with data (friction points, missed deadlines, quality issues)

• Show examples from other industries

• Get commitment to support tough decisions

Step 2: Start with a Pilot

Don’t reorganize the entire studio overnight.

• Pick one team or one feature area

• Restructure it as an outcome team

• Run it for 3-6 months

• Measure results (velocity, quality, team satisfaction)

• Learn and iterate

Step 3: Communicate Relentlessly

People fear change. Over-communicate:

• Why you’re making this change

• What it means for individuals

• What stays the same (craft communities)

• What success looks like

Step 4: Prepare for Attrition

Some people won’t make it.

Territorial leaders who can’t adapt. Individual contributors who thrived in silos but struggle in cross-functional teams.

“This is where leadership has to have some fortitude and say, ‘We’re doing what’s right for the long-term health of the studio.’”

Step 5: Build New Rituals

Cross-functional teams need different rituals:

• Daily standups (across disciplines)

• Sprint planning with the whole team

• Retrospectives focused on outcomes, not just tasks

• Regular syncs between teams within a pillar

The Broader Industry Context

Here’s why this matters right now:

The gaming industry is in crisis. Studios are shutting down. Games are failing despite talented teams. Competition is brutal.

We’re operating in an attention economy where 10,000+ games launch per year. Players have infinite options. Your game has to be exceptional.

You can’t afford organizational friction.

Every week spent in alignment meetings instead of building is a week your competitor gains on you. Every miscommunication that leads to wasted assets is money you can’t get back.

“It’s a little bit egotistical to think that we shouldn’t be having introspection about how we operate. The way game development has been done for years is being blindly followed without adapting for the new environment.”

The studios that will survive are the ones that:

• Eliminate friction

• Move faster

• Iterate more effectively

• Ship quality on budget

And that starts with structure.

Key Takeaways

If you only remember five things from this conversation:

1. Structure is strategy. How you organize your teams determines your outcomes more than individual talent.

2. Go cross-functional. Teams of 5-9 people with mixed disciplines and clear outcomes beat specialized departments every time.

3. Minimize handoffs. Every handoff loses fidelity. Give teams everything they need to own their outcomes.

4. Leadership must commit. Half-measures fail. Territorial resistance will kill this. You need unified executive support.

5. Apply the principles, not the prescription. A 10-person indie team needs a different structure than a 200-person AAA studio. Think about your context.

The uncomfortable truth? Most studio failures aren’t about a lack of talented developers. They’re about putting talented developers in structures that create friction instead of flow.

And that’s on leadership to fix.

Chris Casanova is currently seeking his next production leadership role where he can help teams find their rhythm. If your studio is ready to think differently about structure, you can reach him on LinkedIn or at Chris.Casanova@outlook.com.

Read Chris’s full article: “Structure Teams to Win” on Medium

**Watch the full conversation with Chris Casanova here:**

He goes deeper into implementation challenges, handling resistant leadership, and specific questions about scaling these principles to different studio sizes.

About Chris Casanova: Chris has spent his career as a producer and product leader across Microsoft Xbox, Mojang (Minecraft), Relic Entertainment, Offworld Industries, Misfits Gaming Group, and Hypixel Studios. He specializes in helping game development teams eliminate friction, find their rhythm, and ship great games.

Welcome to the GM MAG Newsletter.

This week, we’re focused on the single most powerful and most expensive variable in game development: Leadership. It’s the “alpha” that can turn a franchise-killing disaster into a record-breaking launch, or a failing team into a high-performance unit, all with the same people.

We have a packed episode. In this week’s MAG, we cover:

• Macro: The Battlefield 6 redemption story. Selling over 7 million units in its debut, this launch proves blockbuster demand is back. We break down the “Vince Zampella effect” and the focused, humble AAA model that made it work.

• Alpha: A must-hear case study from UCLA Football. We explore how a mid-season leadership change—with the exact same players —flipped a 0-4 disaster into a 3-0 powerhouse by fixing the core strategy. How does this apply to your game team?

• GameDev: An excerpt from our chat with veteran producer Chris Casanova. He explains why your “craft-based” silos (art, eng, design) are killing your project and how to swap to “outcome-based” teams for better alignment and velocity.

→ Listen on Spotify, Apple (More Gamemakers content on YouTube)

Updates & Corrections:

Two updates/corrections from the last episode of MAG: The Economics of the EA Acquisition: Lessons in M&A for Game Executives | MAG #6

• EA Saudi PIF Acquisition Focus on Cost Cutting: In the last episode of MAG, I basically reinforced the consensus view from Bloomberg, Financial Times, Game Developer, and other publications that suggested that Silver Lake’s involvement and looking at operational efficiency by comps suggest we should expect “significant cost-cutting measures.” Based on a further discussion with Matthew Kanterman and Chris Petrovic on this topic, I have revised my view that: 1. While there may be some cost-cutting, it likely will not be significant as many think, despite Silver Lake’s involvement, and 2. The comparison of FCF/(# employees) of EA to others is not apples to apples.

  • Full details here: Exposing the EA-Saudi Deal: Why the $55B ‘LBO’ Breaks Every Private Equity Rule | Pixels & Profits

• Copy, Improve, Innovate Framework Credit: Although Zynga is often credited and often takes the credit for the concept of traditional/improve/new in game design/development, shout out to Phillip Black for finding the original source: Sid Meier.

  • Sid Meier in Ars Technica interview in 2014:

    • “Meier said Firaxis has adopted a “rule of thirds” for new Civilization entries: “one-third traditional gameplay, one-third is improved from the last version, and one-third is brand new.”

  • Pincus on “Proven. Better. New.”

1. Battlefield 6 Sells 7M+ at Launch (Reuters): EA’s flagship shooter moved more than seven million copies in its debut stretch, giving the newly private company a strong early signal on franchise health ahead of the holidays. Big for AAA pipeline confidence and Battlefield’s live-ops runway.

2. Nintendo ramps up Switch 2 production to 25 million units by March 2026 (Bloomberg): Nintendo increased Switch 2 production targets to 25 million units by March 2026, exceeding analyst estimates of 17.6 million. Early sales are tracking 77% ahead of the original Switch, making it Nintendo’s fastest-selling console since its June 5, 2025 launch.

3. Arc Raiders playtest reaches 185,000 concurrent players on Steam alone (PC Gamer): The extraction shooter Arc Raiders hit over 185,000 concurrent Steam players during its final pre-release playtest, ranking fourth on Steam’s most-played list. This signals strong anticipation for the October 30 launch, with actual player counts likely much higher across Epic, PlayStation 5, and Xbox Series X/S.

1️⃣ Macro | Battlefield 6: The Western AAA Redemption Story

After years of doom-posting about Western AAA being dead, Battlefield 6 just proved everyone wrong.

The Numbers That Matter

• 7+ million units sold in first 3 days (biggest franchise launch ever)

• 747K peak CCU on Steam (4x previous franchise record, EA’s biggest Steam launch)

• Went from 4.2M units (BF 2042) to 7M units (BF 6) at launch

• 172 million matches played opening weekend

• 15 million hours watched on streaming platforms

For context: Battlefield 2042 was considered franchise-killing. It dropped from 100K concurrent players to 5K within months. EA stock tanked. Players revolted.

So what changed?

The Turnaround Formula

Three and a half years of focused development broke down into:

1. Leadership That Matters
EA brought in Vince Zampella to oversee Battlefield. This guy co-founded Infinity Ward (Call of Duty), built Respawn (Apex Legends, Star Wars Jedi), and has a 20-year track record of shipping hits. When Vince signs on, quality follows.

2. Coalition Model With Clear Pillars
Rather than chaos across studios, each got specific outcomes:

• DICE: Core Battlefield DNA, multiplayer foundation

• Motive: Single-player campaign (addressing BF 2042’s biggest complaint)

• Criterion: Vehicle combat and systems (their specialty)

• Ripple: Battle Royale mode

Each studio focused on what they do best, with explicit accountability for their pillar.

3. Back to Basics
They stopped chasing trends and returned to Battlefield DNA:

• ✅ Four classes (assault, engineer, support, recon)

• ✅ 64 players (not trend-chasing 128)

• ✅ Single-player campaign

• ✅ Tactical squad play and massive destruction

• ❌ No more specialists

• ❌ No more broken launches

4. Battlefield Labs: Community in Development
They created an NDA community program that let real players test pre-alpha builds and provide feedback while there was still time to fix things. This wasn’t marketing—this was product development.

5. Delayed Until Ready
Originally slated for 2024, EA gave it another year. Rare for them, but it paid off massively.

Why This Matters for Western AAA

For years the narrative has been: “Western AAA is too bloated, too risk-averse, too broken. Chinese studios are eating our lunch.”

Battlefield 6 proves: Western AAA can still work if you execute well.

The lesson isn’t “bigger budgets” or “more people”—it’s:

• ✅ Give games time

• ✅ Bring in proven leadership

• ✅ Specialize production with clear outcomes

• ✅ Listen to your community

• ✅ Don’t chase trends—nail your core DNA

• ✅ Polish before shipping

Vince Zampella didn’t throw more money at this. He set vision and held teams accountable. That’s the anti-bloat model that works.

One More Thing: Vince Zampella is EA’s Secret Weapon

If you don’t know who Vince is, here’s the highlight reel:

Infinity Ward Era (2002-2010)

• Call of Duty (2003)

• Modern Warfare (2007) - redefined modern military shooters

• Modern Warfare 2 (2009) - 25M+ copies, bestselling game of its generation

Respawn Era (2010-Present)

• Titanfall franchise

• Apex Legends - 100M+ players, $2B+ lifetime revenue

• Star Wars Jedi: Fallen Order - 10M+ copies

• Star Wars Jedi: Survivor

Now Vince oversees all EA shooter franchises as Group GM:

• Battlefield

• Apex Legends

• Rumored Star Wars FPS project

• Star Wars Jedi series

Hot take: Vince Zampella might be the most valuable executive at EA—more valuable than CEO Andrew Wilson. EA is a hits-driven business, and Vince is a hits machine.

If I’m the Saudi PIF (EA’s new private owner), I’m doing whatever it takes to keep Vince happy.

2️⃣ Alpha | Leadership: The Ultimate Force Multiplier

We obsess over budgets, talent, IP, and market trends. But what if the biggest alpha is who’s leading and how they lead?

The Navy SEAL Story You Need to Know

In Jocko Willink’s Extreme Ownership, he recounts a story from SEAL Hell Week training:

• Boat Crew 6 consistently finished last

• Boat Crew 2 always won

• Instructors swapped the leaders

What happened? Boat Crew 6 went from last to first. Same people. Different leader.

The takeaway: Leadership can be a binary 0-to-1 outcome. The difference between total failure and massive success often isn’t resources—it’s leadership.

Case Study: UCLA Football’s Mid-Season Miracle

Now let’s talk about a real-world 2025 example that just happened.

The Setup:
UCLA football hired DeShaun Foster as head coach—UCLA alum, former NFL player, local hero. But he’d never been a head coach before.

The Disaster (First 4 Games):

• Record: 0-4 (zero wins, four losses)

• Lost to Utah 10-43 (33-point blowout)

• Lost to UNLV, New Mexico (sub-tier programs)

• Averaging 14 points per game on offense

• Giving up 31 points per game on defense

• Players looked lost, demoralized, no fight

What Was Wrong?

1. Wrong scheme for personnel: Foster ran a pro-style power running offense. But UCLA had a mobile QB (Nico Iamaleava from Tennessee), an undersized speed-focused roster, and a weak O-line. They were forcing a power scheme on a team built for speed and space.

2. Overcomplicated play-calling: Too many formations, pre-snap adjustments, shifts. Players were thinking instead of playing.

3. No accountability: Soft culture, no consequences for mistakes, no standards.

4. Wrong defensive philosophy: Bend-don’t-break soft coverage that gave up death by a thousand cuts.

The Mid-Season Shake-Up:
After 0-4, UCLA’s AD made a move you almost never see in college football: fired the head coach mid-season.

• Promoted Tim Skipper (defensive coordinator) to interim head coach

• Brought in Jerry Neuheisel as new offensive coordinator (former UCLA QB, spread offense expert)

Same roster. Same players. Same quarterback. Only thing that changed: leadership, scheme, philosophy.

The Turnaround (Next 3 Games):

• Record: 3-0

• Beat #7 Penn State 42-37 (top 10 upset)

• Beat Michigan State, Maryland

• Went from 14 PPG to 33 PPG on offense

• Went from 31 PPG to 22 PPG on defense

• Total yards jumped from 250 to 425 per game

• Players had energy, confidence, clean execution

  

What Changed?

1. Scheme Fit
Neuheisel implemented a spread tempo offense that matched the roster:

• RPO (run-pass option) plays for the mobile QB

• Faster tempo, simpler play-calling

• Utilized speed and space advantages

2. Brutal Accountability
Skipper instituted clear standards:

• Don’t execute in practice? You don’t play in games. Period.

• Miss assignments? Benched.

• Execute well? Publicly praised, more responsibility.

3. Crystal Clear Roles
Every player knew exactly what was expected:

• “You’re the starting linebacker—you’re the QB of the defense”

• “You’re backup corner—your job is special teams at 100%”

• No ambiguity. No confusion.

4. Restored Belief
Foster had lost the locker room. Skipper came in and said: “You guys are talented. We can beat anybody. I believe in you.”

When players believe, they play differently—faster, harder, with confidence.

The Alpha Insight

Same people. Different leadership. Different strategy. Infinite improvement.

This is as true for your game studio as it is for UCLA football or Navy SEAL boat crews.

Questions You Should Ask Right Now

If you’re a studio lead, answer these honestly:

1. Am I using my people well, or forcing them into roles that don’t fit their strengths?

2. Is my production model adapted to my team, or am I using a scheme that doesn’t work for us?

3. Am I focused on fundamentals, or chasing flash without nailing basics?

4. Am I adaptive, or rigidly sticking to a plan that’s clearly not working?

5. Do I have accountability and belief, or is my culture too soft?

The brutal reality: The games industry is in crisis. Layoffs everywhere. Studios closing. Budgets out of control. Games flopping.

Yes, the industry is hard. Yes, the environment is difficult.

But that’s exactly why leadership matters most right now.

You can’t fix bad leadership with more money or more people. You fix it with better leadership and better strategy.

UCLA proved a 0-4 team can fix everything in one week. Battlefield 6 proved a franchise-killing disaster can become a record-breaking success in one dev cycle.

Call to action:
Invest in your leaders. Train them. Hold them accountable. If they’re not performing, replace them.

If you ARE a leader: be adaptive, be aggressive, simplify, focus on fundamentals, make bold strategic pivots when necessary.

Because if you get leadership and strategy right, everything becomes possible.

3️⃣ GameDev | Structure Teams to Win: The Outcome-Based Model

For this section, we’re featuring insights from Chris Casanova, veteran producer with experience at Microsoft/Xbox, Mojang, Relic, and High Pixel Studios.

Chris wrote a fantastic article called “Structure Teams to Win” that addresses a root cause of failure in game development: bad organizational structure.

The Problem: Craft-Based Silos

Most studios organize by discipline:

• All designers in one team

• All engineers in another

• All artists in a separate group

This creates:

• ❌ Constant handoffs between teams

• ❌ Games of telephone that lose intent

• ❌ Delayed communication loops

• ❌ “Throwing specs over the fence” mentality

• ❌ Lack of ownership over outcomes

The Solution: Outcome-Based Teams

Instead of organizing by what people do, organize by what needs to be built.

Example: Procedural Worlds Team

Rather than having designers, engineers, and artists split across different silos working on “procedural generation,” you create ONE cross-functional team with a specific mission:

Mission: “Every new world seed feels fresh, readable, performant. Traversal and discovery naturally surface points of interest, resources, and risks.”

Team Composition:

• Designer (world generation logic)

• Engineers (procedural algorithms, performance)

• Artists (biome visual variety, readability)

• All focused on ONE outcome

Measurable Goal: “Biome novelty per hour” - players experience something new within X minutes.

Why This Works:

✅ Clarity: Everyone knows exactly what they’re building and why
✅ Ownership: One team owns the entire outcome, not pieces of it
✅ Speed: No handoffs, no waiting for other departments
✅ Specialization: Team builds deep expertise in their domain
✅ Accountability: Success or failure is clear

Key Principles

1. Keep Teams Immutable (Mostly)
Don’t constantly change the mission or composition. Let teams:

• Build excellence in their area

• Develop strong relationships

• Avoid constantly re-forming and re-storming

2. Arm Teams to Win
A football team of only quarterbacks would get destroyed. Similarly, your team needs the right mix of skills to execute their mission.

Ask: Can this team actually deliver on what we’re asking them to do?

3. Align to Product, Not Process
The team structure should reflect the product you want to build, not just the functional disciplines you hired.

The Sports Analogy

Imagine fielding a football team with only quarterbacks. You’d get wrecked.

Outcome-based teams are like having:

• QB (designer)

• O-line (engineers)

• Receivers (artists)

• Running backs (technical artists)

All working together toward ONE goal: moving the ball down the field.

Practical Takeaway

If you’re structuring (or restructuring) your studio:

1. Identify core outcomes your game needs to deliver

2. Build cross-functional teams around each outcome

3. Give each team clear ownership and measurable goals

4. Staff teams with the skills needed to fully execute

5. Keep teams stable so they can build mastery

This is a fundamental shift from “we have a design team, an engineering team, an art team” to “we have a combat team, a progression team, a procedural worlds team”—each fully equipped to deliver their outcome.

Wrapping Up

Three big ideas this week:

1. Battlefield 6 proves Western AAA can still dominate with focused execution, proven leadership, and respect for community + fundamentals

2. Leadership is the highest-leverage alpha you can find—same people, different leader, completely different outcomes

3. Outcome-based teams crush discipline silos because they own entire features end-to-end, not disconnected pieces

The full conversation with Chris Casanova drops soon as a special episode—we go deep on team size, psychological safety, and a complete org chart for an AA shooter. Keep an eye out for it.

Please like, share, and comment. It really helps us grow.

Catch you next week, GameMakers. 🎮

“If I got an unsolicited offer from a sovereign entity that has effectively unlimited money, would I push for more? I think so.”

That’s Matthew Kanterman, CFA, gaming industry analyst, raising the question almost no one else is asking about EA’s $55 billion sale to Saudi Arabia’s Public Investment Fund: Why did EA accept this price, right now?

The gaming industry has just witnessed its largest-ever LBO, with nearly everyone focusing on the obvious story—Saudi Arabia’s gaming ambitions, Vision 2030, and strategic synergies. But the more revealing insights come from flipping the question: What made Andrew Wilson and EA’s board choose to exit public markets at this specific moment, and why does calling this a “typical LBO” fundamentally misunderstand what’s actually happening?

I sat down with two people who know precisely how these deals work: Chris Petrovic, the architect behind Zynga’s massively successful acquisition strategy (now Chairman and CBO at FunPlus), and Matthew Kanterman, CFA, who specializes in dissecting gaming M&A and exposed an accounting myth that’s been distorting EA’s operational story for years. I totally got this one wrong as well!

What emerged wasn’t just an analysis of one deal—it’s a window into how sovereign wealth is rewriting the rules of gaming M&A, why traditional efficiency metrics have been measuring EA all wrong, and what happens when “patient capital” meets an industry built on quarterly earnings pressure.

The Real Question: Why Did EA Sell Now?

Most coverage obsesses over Saudi’s rationale. That’s the easy part—they’ve publicly stated gaming is central to Vision 2030, they want to own pieces of the attention economy, and EA’s sports portfolio aligns perfectly with their massive investments in live entertainment (WWE, UFC, global football).

But Petrovic, who’s negotiated enough mega-deals to know what happens behind closed doors, immediately zeroes in on the seller’s perspective:

“I think less has been discussed around it from the seller’s perspective, which is putting yourself in Andrew and the board’s shoes. Why would they choose to take themselves off the public markets where they have a great asset and a great currency against which they can do things like raise debt and continue to have a tool for compensating employees?”

The signals pointing to pressure:

Portfolio concentration risk: Over 50% of EA’s revenue flows from just four franchises. Sports remain dominant, but what comes after Battlefield launches? Where are the next major catalysts?

Chinese competition intensifies: With developers like Tencent and NetEase dominating shooters, can Apex Legends hold its position? The competitive dynamics are shifting dramatically.

Public market scrutiny: Perhaps EA’s leadership saw pending challenges they didn’t want to navigate under quarterly earnings pressure and analyst scrutiny.

Petrovic lays out the possibilities: “Perhaps there were some palatable headwinds they were noticing in their business. Perhaps there were some pending hostile takeovers that may have been in process. Maybe they were going to head off an eventual approach from other strategics that have always been rumored. Maybe they just didn’t want to deal with the scrutiny of being able to take riskier bets and think longer term.”

Coming Soon: Within 45 days of the announcement, EA must file a proxy statement detailing any conversations with other potential buyers. If Take-Two, Disney, or others were circling, we’ll know soon whether this was EA’s preferred “white knight” scenario to avoid a more hostile alternative.

The valuation question persists: At 17-18x forward EBITDA, EA sold at a discount to Activision’s 20x multiple when Microsoft acquired them. Different market conditions, yes—but Kanterman’s question lingers: what was happening inside EA that made this deal, at this price, the right answer?

🎧 Listen on Spotify, Apple Podcasts, or Anchor

Speakers:

• Matthew Kanterman. Director of Research at Blue River Financial Group.

• Joseph Kim. CEO at Lila Games.

• Chris Petrovic. Board Chairman and CBO at FunPlus.

The Accounting Myth That Changes the Efficiency Story

Here’s what many folks got completely wrong (including me in my last MAG newsletter) about—and why the “operational bloat” narrative needs to die.

The conventional wisdom: At $128K free cash flow per employee versus Activision’s $270K or Nintendo’s $390K, EA appears inefficient and overstaffed. Fire up the cost-cutting playbook, right?

The reality: “Prepare to be bored with accounting dorkiness,” Kanterman warns, before dropping the truth bomb that reframes everything:

“EA has been notoriously the most conservative of all the public companies. They just expense everything. So all the expenditures that go into making a game is just expense in the current period. Others like Take-Two, Activision, they’ve chosen to capitalize some of that stuff. So it doesn’t hit the P&L, it goes below free cash flow actually.”

Why this matters for everyone analyzing the deal:

• Different accounting rules create incomparable metrics: US GAAP vs. international IFRS standards handle software development costs differently

• Capitalization vs. expensing affects both timing and location: Where costs appear on financial statements varies by company

• Hardware revenue skews the numbers: Comparing Nintendo (which sells Switch) to pure software publishers is apples-to-oranges

• On a bookings basis—the comparable metric—EA performs competitively: The efficiency gap isn’t what surface numbers suggest

Kanterman continues: “When you’re comparing across geographies and even within the U.S. with Activision, Take-Two and EA, they all account for it differently. On a bookings basis it’s much more comparable and EA does compare pretty strongly.”

Translation: The case for aggressive cost-cutting isn’t as obvious as the per-employee metrics suggest. EA’s conservative accounting makes them look less efficient than they actually operate.

Why This Isn’t a Traditional LBO (And Why That Matters More Than You Think)

Call it an LBO if you must, but Petrovic is blunt about what this deal actually represents:

“I have a little bit of a different take on this in the sense that I don’t know that operational efficiency is going to be a front and center agenda item anywhere in the near to medium term.”

Wait—what? A $20 billion debt load and no focus on operational efficiency? That violates every private equity playbook rule. Which is exactly Petrovic’s point.

Silver Lake’s Real Role

“I wonder why they needed JP Morgan to float a loan when PIF could have set up a loan instrument as well. They have plenty of money to fund the whole thing. And my sense is that if anything gets squirrely around repayment, [PIF] will step in and just bridge the gap to protect their investment.”

The structure reveals the strategy:

• Silver Lake and Affinity Partners likely hold high single digits to low double-digit equity stakes each

• That’s perhaps 20-30% of equity, representing only ~10% of total enterprise value

• They came in for optics and expertise, not control

“Silver Lake came in for optics,” Petrovic argues. “This is not a typical LBO where they have a majority position they need to protect.”

No Proctology Exam Required

Forget the stereotypical PE playbook of day-one cost purges:

Traditional LBO playbook:

• Immediate 10-15% headcount reductions

• Aggressive portfolio rationalization within 90 days

• Quarterly debt coverage anxiety

• 5-7 year exit timeline driving every decision

The Saudi PIF reality:

• 10+ year investment horizon minimum

• No public earnings reports or quarterly scrutiny

• Benevolent benefactor willing to backstop debt if needed

• Investment in growth is prioritized over margin optimization

• Privacy to make long-term bets without explaining them to analysts

“You don’t need to worry about publicly stating your financials anymore,” Petrovic notes. “They’re not going to be tapping their wristwatches saying ‘when are we going to see the next 10% cuts?’ They want a plan for growth, they’re going to be willing to invest in it.”

The debt coverage reality: At roughly $2 billion in adjusted annual free cash flow (factoring in Battlefield’s contribution), EA covers the estimated $1.6 billion annual payment on $20 billion at 8% interest. And if concerns arise? Kanterman points out that PIF also owns Scopely—they could theoretically bundle cash flows or simply inject more equity.

“There’s a world where they could put part or all of Scopely’s cash flow in the box for the creditors,” Kanterman explains. “If the Saudis were concerned about debt coverage, they have a lot of levers to pull.”

The Mobile Question: Integration or Independence?

EA derives 15-18% of revenue from mobile (primarily FC Mobile and Sims Mobile). PIF owns Scopely, one of mobile’s most sophisticated operators known for aggressive UA strategies. The obvious question: massive synergies incoming?

The less obvious answer: probably not the way most people expect.

Why Full Integration Is Unlikely

Structural separation: EA and Scopely remain distinct entities under common ownership—not a merged company. Any collaboration requires independent business justification from both sides.

“Unlike Zynga and EA, this is going to have to be arm’s length dealings,” Petrovic explains. “These are two related entities that still operate somewhat independently because they’re not under common ownership in the on-paper sense.”

Anything Scopely does—whether taking over publishing, licensing IP, co-development—requires their own business case. Same for EA. They may have a pole position in negotiations, but this isn’t a top-down “you will collaborate” mandate.

The Best Practices Opportunity

Rather than full integration, expect knowledge transfer:

Where Scopely excels:

• Sophisticated UA optimization and scaling strategies

• Monetization framework development

• Live operations and player engagement

• Using global resources efficiently

Where EA mobile has struggled: According to industry sources, EA’s mobile UA function has been basic and not aggressive enough—the opposite of Scopely’s reputation.

“If they can drive efficiencies and best practices in UA to scale [FC Mobile and Sims Mobile], there’s huge synergies and that falls down pretty quickly,” Kanterman observes.

But he’s skeptical about the smaller titles: “You take a $5 million mobile game and make it $10 million, you’re probably losing money adding $5 million in revenue. You’re probably losing EBITDA to get that incremental revenue. So it’s not worth it.”

The opportunity: Optimize the few large existing mobile titles (FC Mobile, Sims Mobile) through Scopely’s proven frameworks. Don’t waste energy rescuing small struggling games.

The Take-Two Cautionary Tale

Nearly four years after Take-Two acquired Zynga for $12.7 billion, not a single mobile title based on Take-Two’s massive IP library (GTA, Red Dead, Bioshock, Borderlands) has launched through Zynga.

“This deal has been over four years old and not one hint of any Take-Two mobile IP coming to bear through Zynga,” Petrovic notes. “So you can maybe expect the same here, especially because there’s actually no public market expectation of that synergy having to happen within a certain period of time.”

Don’t hold your breath for rapid EA-Scopely IP co-development.

The Real Growth Drivers: Sports, Asia, and Transmedia

Sports: The Strategic Core

The synergies between EA FC (formerly FIFA) and Saudi’s sports investments practically write themselves:

PIF’s sports portfolio includes:

• Ownership stakes in TKO (WWE/UFC)

• Major investments in global football/soccer

• Saudi Pro League transformation ambitions

• Pursuit of the 2034 World Cup hosting

The transmedia vision: “I think it’d be very interesting if you start thinking about a world where you’re interacting with live events inside of FC, inside of the core video game, playing along with it and rethinking what it means to watch a soccer game,” Kanterman suggests.

“There’s a vision there that makes a lot of sense that can really make EA FC... not just the marquee video game for football soccer, but the marquee entertainment hub for football soccer.”

This could include:

• Live event integration with real matches

• Streaming rights innovation

• Interactive viewing experiences

• Blurring lines between gaming and sports consumption

Sports isn’t just EA’s most stable business—it’s the strategic centerpiece of the entire Saudi entertainment vision.

Asia: The China Opportunity (and Necessity)

EA has consistently underperformed in Asia relative to market potential. Meanwhile, Chinese developers now dominate certain categories globally—not just in China.

The competitive reality: “Enough writing is on the wall for Western developers to acknowledge that they just don’t have the muscle to develop as commercially effective, scalable games in certain categories as China,” Petrovic argues.

With Affinity Partners (Asia/China specialists) in the consortium and Saudi’s global ambitions, expect M&A targeting Chinese development capabilities:

Potential targets:

• Century Games (proven Roblox operator—Kanterman notes potential for “Sims for the next generation” on Roblox)

• Studios focused on using Chinese development talent for global markets like MiHoYo

• Proven operators who can either create global hits or level up EA’s existing studios… hmm, again like MiHoYo

“I wouldn’t be surprised if some combination of Scopely, EA, and their benefactors penetrate the China market with some kind of M&A to bring those capabilities under the banner,” Petrovic predicts.

The dual opportunity: Use Chinese development expertise for both cost efficiency (AAA budgets are escalating unsustainably) and competitive quality (Chinese shooters are beating Western ones).

What Could Go Wrong: The Execution Risk Is Real

Five years from now, this could be a strategic masterpiece or a cautionary tale. The risks aren’t trivial.

The Franchise Concentration Problem

With 50%+ revenue from four franchises, any major stumble cascades:

• Battlefield’s reception matters enormously

• Apex Legends faces fierce competition from Chinese shooters

• AAA development challenges (Dragon Age struggles) haven’t disappeared

• What comes after the current roadmap?

The Studio Rationalization Question

EA’s famously siloed structure—Vancouver (FC), Florida (Madden), Maxis (Sims), BioWare—may eventually face pressure.

Kanterman is direct: “Do I think the day this deal closes they’re going to shut down Studio X and Studio Y? No. But I do think they’re going to give these studios a chance. And I wouldn’t quite call it a fiefdom, but these are kind of their own entities. I do think there is going to be a period of ‘prove it or what do we do with this?’”

Studios that fail to align with Saudi Arabia’s broader vision (sports, transmedia, global entertainment) may face rationalization. BioWare, which has struggled with recent releases, represents this risk most acutely.

But Kanterman emphasizes: “Do I think it’ll happen on day one or day 60, day 90? No, I think it’s going to be a much longer process.”

The Financial Tightrope

Even with patient capital, the fundamentals matter. At $2 billion annual free cash flow covering $1.6 billion debt payments, there’s limited cushion without growth or efficiency gains.

“Obviously they’re astute investors and they’re not just going to keep funding something in perpetuity that’s going to bleed money,” Petrovic cautions. “There is a point at which time I’m sure that they’ll run out of patience. But I think that patience timeline is a lot longer than it would be if it was a typical private equity timeframe.”

Industry Implications: Who Controls Gaming’s Future?

EA’s privatization accelerates a concerning trend for the gaming industry’s public market presence.

The Shrinking Public Market

After this deal closes, Take-Two becomes the only remaining independent major publisher at scale in U.S. public markets.

The ripple effects:

For private companies:

• Fewer public comparables for valuation benchmarking

• Reduced strategic acquisition exit opportunities

• Less institutional capital interest in gaming startups

For the industry:

• Decreased investor visibility and understanding

• Less scrutiny of business practices (good and bad)

• Concentration of ownership among sovereign wealth funds and tech giants

“When you think about just what that means for investor interest at a high level in the sector, it’s not a positive,” Kanterman argues. “Having fewer large public players reduces the exit opportunities for strategic acquirers.”

The Geopolitical Question

Between Scopely, Niantic (Pokémon Go), and now EA, Saudi Arabia will control roughly 10% of the global gaming market. Add Chinese company dominance, and a question emerges:

“Who do you want owning this attention economy—the time and money spent in gaming?” Kanterman asks. “That’s an important question everyone needs to think about.”

American and European independent studios are shrinking, while sovereign wealth and Chinese conglomerates are expanding. The concentration of gaming ownership into non-Western hands deserves more discussion than it’s getting.

The Optimistic Counter-Narrative

Petrovic offers a different lens: “The more large scale companies that go private, in theory, the more that there is opportunities for them to lean into innovation and long term thinking.”

He sees cyclical opportunity: “We are in a natural cycle in our industry that perhaps is leading to contractions, consolidation. We could very well be on our way to the next upcycle in gaming once we’ve come out of this new innovation phase driven by privatization, enabled by technologies like AI, and driven by the ongoing size of gaming within entertainment.”

Next wave candidates: Discord (IPO expected), Roblox (growing), Epic Games (perpetually rumored), and new entrants could replenish the public gaming ecosystem within 3-5 years.

“I have no doubts that we’ll come out of this stronger,” Petrovic concludes. “We’re just going through growing pains and filling out this teenage body of ours.”

The 5-Year Outlook: Success on Different Terms

Here’s the fundamental insight that changes how to evaluate this deal:

Traditional metrics might be irrelevant.

“At the end of the day, it almost doesn’t matter how it turns out,” Petrovic argues. “EA is just a cog in the larger wheel. Even if it doesn’t [hit financial targets], it really won’t be called a failure because it’s feeding so many other things and upstream aspirations.”

Measuring Success Differently

Traditional view: Revenue growth, margin expansion, debt coverage, franchise health, return on investment

Strategic view:

• Contribution to Saudi Arabia’s entertainment and sports dominance

• Success in securing the 2034 World Cup and major sports events

• Development of Saudi Arabia as a global entertainment destination

• Control of attention economy touchpoints across 2+ billion players

• Reputation transformation on the global stage

“These are things at levels we’ve never seen before in terms of big picture,” Petrovic notes. “Us focusing on EA’s billion swing here or there is going to become so inconsequential to the larger picture that it really won’t matter from that perspective.”

Kanterman’s pragmatic view: “It also depends on how you measure success here. There’s the financial angle where like does $8 billion of revenue become $7 or $10? Obviously if it shrinks financially, they’re losing. But that doesn’t necessarily mean that they’ve lost.”

If FC integrates deeply with live football events, if EA helps secure the World Cup, if Saudi Pro League becomes a top-10 global league—the deal succeeds even if EA’s standalone financials plateau or decline modestly.

Key Takeaways

Rethink the efficiency narrative: EA’s conservative accounting makes direct peer comparisons misleading. The operational bloat thesis isn’t supported by comparable bookings-based metrics.

This isn’t about short-term cost-cutting: Expect long-term investment in growth rather than the typical LBO playbook. Silver Lake’s involvement appears primarily optical. Patient capital changes everything.

Mobile synergies will be limited: Expect best practices sharing between EA and Scopely for major titles (FC Mobile, Sims Mobile), but not the aggressive integration many anticipate. Arm’s-length dealings will likely prevail.

Sports is the strategic core: The real opportunity lies in making EA FC the central entertainment hub for global football, integrated with Saudi’s massive sports investments and transmedia vision.

Watch for China M&A: With Affinity Partners involved and Chinese developers dominating key categories, expect acquisitions targeting development capabilities for both quality and cost efficiency.

Studio rationalization will come—slowly: Entities like BioWare that can’t align with the broader Saudi vision face long-term risk, but not the typical day-90 purge of traditional PE deals.

Industry concentration accelerates: Take-Two becomes the lone major independent US publisher. The implications for startup exits, valuations, and who controls gaming’s attention economy deserve serious consideration.

Measure success differently: EA’s financial performance might matter less than its strategic contribution to Saudi Arabia’s broader entertainment and geopolitical goals. This is a 10+ year play, not a 5-year flip.

The EA acquisition isn’t just about one company changing hands. It’s about a fundamental shift in who controls gaming’s attention economy, how much patience capital will have for AAA development’s escalating costs, and whether privatization enables the bold innovation bets that public market constraints have suppressed.

The smart money isn’t asking whether EA will grow 10% or shrink 10%. They’re asking what happens when sovereign wealth deploys patient capital at unprecedented scale in gaming, and whether the traditional playbook for measuring success still applies.

We’re witnessing gaming’s first truly patient capital experiment at mega-scale. The rules have changed. We just don’t know yet if that’s good or bad.

About the Experts:

Matthew Kanterman, CFA is a gaming industry analyst specializing in M&A, financial analysis, and market strategy. LinkedIn

Chris Petrovic is currently Chairman and Chief Business Officer at FunPlus, and previously served as President of Publishing at Zynga, where he led the company’s transformative acquisition strategy. LinkedIn

What’s your take on the EA acquisition? Hit reply with your thoughts—I read every response.

It’s one of the most common frustrations in game development today: your traffic is growing, your DAU is climbing, but your ad revenue is stubbornly flat. You’re doing everything right, yet the numbers don’t scale. This isn’t just a feeling; it’s a reality for countless studios getting squeezed in an increasingly complex ad monetization landscape.

To get to the bottom of this, we sat down with veteran ad monetization consultant Kelly Kang, whose experience spans from casual puzzle games at Big Fish to social casino at Aristocrat. She provides a clear, no-nonsense framework for not just plugging the leaks in your revenue, but building a robust strategy from the ground up.

Here are the key takeaways from our conversation:

• Your Real Margin Isn’t What’s in Your Contract. The biggest pain point is shrinking margins, but not for the reason you think. Beyond the agreed-upon rev-share, hidden “tech fees” and other costs can mean you’re taking home less than 50% of the ad spend, even on a 70/30 deal. The focus shouldn’t be on negotiating percentages, but on the final net dollar amount you receive.

• Stop Obsessing Over ARPDAU and CPM. While important, these metrics don’t tell the whole story. Kelly advises focusing on deeper player experience metrics: Ad Engagement (are players opting in?), Ad Session Length (are ads driving users away or keeping them in the game?), and Impression Depth (are players seeing one ad or multiple ads per session?). These KPIs reveal the true health of your ad integration.

• For a New Game, Product Comes Before Tech. The most common mistake is treating ad monetization as an afterthought. Before you even think about choosing a mediation platform, you must first define your in-game ad formats and their impact on your game economy. Choosing your ad networks before you know whether you’re using rewarded video, offer walls, or interstitials is a recipe for disaster, as not all networks excel in every format.

• “Yelling at Your Partner” Doesn’t Work Anymore. The old-school method of calling your ad network rep to demand better performance is dead. With over 95% of deals now being programmatic, the key is to collaborate with partners to gain insights and establish fair, data-driven competition through your mediation platform. The machine has already won; your job is to provide it with the correct inputs.

This is one of the most practical, no-fluff conversations we’ve had on building a modern ad monetization stack. For the full deep-dive, watch our complete discussion with Kelly Kang below:

🎧 Listen on Spotify, Apple Podcasts, or Anchor

Speakers:

• Joseph Kim. CEO at Lila Games.

• Kelly Kang. Co-founder at Game Plan.

The Day One Audit: A 4-Point Checklist for Your Admon Stack

When Kelly takes on a new client, she runs a systematic audit that uncovers the most significant gaps and opportunities. It’s a framework any studio can use to assess its own strategy, focusing on four key areas.

1. Mediation: Header Bidding vs. The Waterfall

The default advice for lean teams is to start with in-app header bidding—it’s dramatically less operational work than a manual waterfall. However, the most significant insight here is that you shouldn’t treat it as a permanent, “set it and forget it” solution. As your game scales, you can begin A/B testing a hybrid model.

Kelly Kang: “Not all ad networks are better in in-app header bidding. Meaning if you change them to a manual waterfall...and give them the rate...they come in a lot higher than when you give them the freedom of to pick whichever user they can come in as.”

This is a critical nuance. Some networks perform better when given specific price points in a waterfall. In the full discussion, Kelly outlines the specific signals—like a dramatic DAU shift or a major advertiser budget in Q4—that tell you when it’s time to start experimenting with a hybrid approach.

2. Data: Measuring What Actually Matters

Are you tracking the right things? Moving beyond surface-level revenue metrics is crucial to determining whether your ads are helping or hurting your business. Kelly recommends tracking:

• Ad Engagement: What percentage of your players are actively choosing to watch an ad? If a new placement is launched, is it hitting a 70% engagement rate or a 20% rate?

• Ad Session Length: After an ad is shown, does the user’s time in-app increase or decrease? This directly answers whether your ads are supporting or killing retention.

• Impression Depth: Are users seeing just one ad, or are they reaching the third, fourth, or fifth ad placement you’ve designed? This metric shows how “sticky” your ad loop is.

3. Placements & Frequency: Personalizing the Experience

Not all players are the same, so why should their ad experience be the same? A critical optimization layer is segmenting players based on their behavior.

Kelly Kang: “Think of it as a heavy ad watchers or passive ad watchers...every time they have ads available they continue to watch an ad. That means there’s less likely they’re going to actually click on that advertising. So we want to give them more of the impression-based advertising rather than action-based.”

This simple segmentation—between “aggressive” ad viewers who watch everything and “passive” viewers who watch occasionally—allows you to serve different campaign types to maximize yield without alienating users. We dive deeper into segmenting by player level and reward scaling in the video.

4. Operational Efficiency: Stop Wasting Time

One of the biggest hidden costs is wasted time. Kelly often sees multiple people—a PM, a GM, and an ad monetization manager—all spending an hour a day manually pulling the same data. Creating a single, automated template report can save dozens of hours a month.

The other key is leveraging your partners correctly. Don’t spend five hours digging for an answer; ask your network partner for their data and insights first.

Building a Team and a Strategy from Scratch

For a new game, Kelly advocates a surprisingly lean approach.

Kelly Kang: “I do believe in too many cooks in the kitchen causes problem...You don’t need a big team right away for a new game. One dedicated person focusing on setup, mediation, just a basic optimization day-to-day is usually enough.”

For early-stage studios, she recommends starting with a fractional headcount (half a person) or an external consultant to establish the right foundation without over-committing resources.

In our discussion, Kelly provides a detailed breakdown of how to choose your first mediation platform (LevelPlay vs. Max) and your first 4-6 ad network partners, and how to scale the team based on the revenue each new headcount can generate, not just the workload.

What do you think? Is treating ad monetization as a core product feature from day one the biggest unlock, or are there other challenges we’re missing? Let us know in the comments, and be sure to watch the full video for even more deep-dive insights from Kelly.

TRANSCRIPT

The full discussion transcript with Kelly follows:

Read more

Hey Folks,

Coming at you this week:

• Macro: We dissect the massive $55 billion acquisition of Electronic Arts. I’ll take you beyond the headlines and offer an operator’s perspective. We’ll also deep dive into the likely buy-side and sell-side narratives that got this deal done.

• Alpha: We’ll explore the alpha lessons from Grow a Garden, a Roblox game made by a 16-year-old that quietly became the biggest game in the world for a time—secret interview with a games industry insider.

• Game Dev: I’ll share a powerful mental model called the “Alien’s Eye View” that will prompt you to examine whether your team’s actions truly align with its stated goals or intended purpose.

There’s a lot in the audio podcast that isn't covered in the newsletter; be sure to check it out.

→ Listen on Spotify, Apple (More Gamemakers content on YouTube)

1. Saudi Arabia-Led Consortium Acquires EA for $55 Billion (Axios): Electronic Arts announced a historic $55 billion acquisition by a consortium including Saudi Arabia’s Public Investment Fund, Silver Lake, and Jared Kushner’s Affinity Partners in the largest leveraged buyout ever. The deal, announced September 30, will take EA private with $20 billion in debt financing from JPMorgan Chase, raising concerns about potential studio closures and layoffs to service the debt.

2. Xbox Game Pass Ultimate Price Increases 50% to $30/Month (CNBC): Microsoft raised Xbox Game Pass Ultimate pricing from $20 to $30 per month effective October 1, adding Ubisoft+ Classics and Fortnite Crew while upgrading cloud gaming to 1440p. The restructuring eliminates the Game Pass Core and Standard tiers in favor of Essential and Premium options, prompting widespread user backlash and leading to a surge in subscription cancellations.

3. Ubisoft and Tencent Launch Vantage Studios (TweakTown): Ubisoft has established Vantage Studios, a new Tencent-backed subsidiary with a $1.2 billion investment, to oversee the development of the Assassin’s Creed, Far Cry, and Rainbow Six franchises. Tencent owns 25% of the new entity.

1. Macro: EA’s $55B Acquisition - In-Depth Analysis from an Operator’s Perspective

This acquisition marks a somber ending for a company that began with a noble mission. In 1983, founder Trip Hawkins created Electronic Arts to empower a new class of entrepreneur: the “Software Artist.” That vision of championing creators feels worlds away from the EA of today—a scaled publisher of big IP, more focused on commercialism than artistry. In that light, the acquisition feels like a natural, if uninspired, fit.

Trip’s original vision for Electronic Arts:

So, why did the largest leveraged buyout in history happen now? It comes down to a collision of strategic imperatives and compelling financial narratives.

The Banker’s Pitch: Two Sides of a $55B Coin

Every major deal is built on two stories. Here’s the potential pitch that got this done.

• The Buy-Side Story (Why the Saudi PIF Bought EA)

  • Strategic Fit: The deal is a cornerstone of Saudi Vision 2030, the country's plan to diversify its economy away from oil by investing in culture, sports, and entertainment. Gaming is a key pillar of the modern “attention economy,” and EA provides a dominant position.

  • Massive Upside “Potential”: The pitch from the bankers likely centered on four pillars of growth for the underperforming giant:

    • Operational Efficiency: EA generates just $128k in free cash flow per employee, compared to Activision’s $270k and Nintendo’s $390k. With operational specialists from Silver Lake involved, expect significant cost-cutting measures.

    • Mobile Growth: EA’s mobile revenue is stagnant at ~16% of total bookings. The plan is likely to inject liveops DNA from other PIF investments (like Scopely) and 2-3x this business.

    • Asia Expansion: The APAC region accounts for 50% of the global games market, yet it generates less than 15% of EA’s revenue. Fixing this historical weakness is likely a core part of the thesis.

    • Transmedia & AI: EA’s IP (Mass Effect, Dragon Age) is significantly underleveraged in film and TV. AI is also framed as a call option for massive future cost reduction and revenue growth.

• The Sell-Side Story (Why EA’s Board Sold)

  • Stalled Growth & Franchise Fatigue: The on-the-ground reality at EA is challenging. The company posted -1% top-line growth in fiscal ‘25, and live services growth has stalled at 1%. Key franchises are showing their age, with Dragon Age: The Veilguard missing sales targets significantly after nine years in development.

  • Concentration Risk: Over 50% of EA’s revenue comes from just four franchises. The biggest, Ultimate Team (~$2.2B), relies on loot box mechanics that face serious regulatory risk.

  • The Competitive Threat: China is producing massive, high-quality original IPs (Genshin Impact, Wuthering Waves) that are leaving Western development behind. EA’s historical advantages in retail distribution are now meaningless. From this perspective, EA looks more like a dinosaur than an insurgent, making a sale at a 25% premium a smart move.

Deal Mechanics & Industry Implications

• High Leverage, High Risk: The deal involves roughly $20 billion in debt. This requires ~$1.8B in annual debt service, which is roughly equal to EA’s average free cash flow. This represents a high level of leverage (9.8x debt-to-EBITDA, compared to an LBO average of 5.2x). While Saudi’s deep pockets can cover any shortfalls, execution is not guaranteed.

• Industry Impact:

  • More Big Deals: The banks made nearly a billion dollars in fees here. You can be sure they are already pitching similar LBOs for companies like Ubisoft and CD Projekt.

  • Information Blackout: With fewer public companies, we lose valuable business data from public disclosures.

  • Saudi’s Push to Be #1: The PIF seems intent on being the top player in gaming. Expect them to push their portfolio to create #1 hits, which will have significant ripple effects on the competitive landscape.

How will EA do under new ownership? Well, that’s a big question. I’m skeptical, but you can do your own analysis. Here’s a hypothetical example illustrating how you can draw your own conclusions by focusing on the impact of the various theses on FCF:

^^ This is obviously a conceptual example.

I delved deeply into all of these issues and more in the podcast. Be sure to check it out!

2. Alpha: Lessons from the Biggest Game in the World—Grow A Garden

Grow Garden hit 22.3 million concurrent players on Roblox—more than Fortnite’s all-time peak. The secret? It’s just Farmville with a modern coat of paint.

This week, we have a Q&A with a secret game industry insider who has specific expertise in Roblox games and Farmville.

Here’s what I asked:

1. From your perspective, how did Grow a Garden become so successful?

2. To some degree, Grow a Garden is just a modern-day FarmVille. What are the similarities between the two games?

3. What are the differences between FarmVille and Grow a Garden that enabled the game to achieve so much success on Roblox? What are the platform-specific drivers?

4. Grow a Garden was built by a young dev… but then a deal was struck with a team. What did that deal look like, and what did that team do to help drive the game to so much success?

5. How do you view the opportunity of looking at older games and reviving them for a modern audience on a platform like Roblox, similar to Grow a Garden?

Lesson 1: The Innovation Framework That Still Works

Our secret gaming insider reveals the formula popularized by Mark Pincus/Zynga:

The Zynga Framework:

• 1/3 Proven mechanics (farming gameplay)

• 1/3 Better execution (3D, real-time multiplayer)

• 1/3 New elements (viral TikTok-worthy plants, PvP stealing)

Lesson 2: What Farmville Got Wrong That Grow Garden Fixed

The Core Loop Evolution:

• Farmville: Plow → Plant → Harvest (PPH)

• Grow Garden: Plant → Harvest (simplified)

• Result: Lower friction, faster dopamine hits

The Audience Adaptation:

• Removed crop withering (killed Farmville’s casual audience)

• Added stealing mechanics (Gen Z loves competitive elements)

• Made plants “over the top” for social sharing

• Real-time multiplayer vs asynchronous Facebook requests

Lesson 3: The Publisher Playbook

BMW Lux built it in 3 days, but Splitting Point Studios scaled it:

The Growth Hacker Model:

• Scout for diamonds in the rough

• Invest/acquire promising games

• Professionalize the operation (UI/UX, LiveOps, monetization)

• Move fast, pivot faster

• Current team: ~30 people managing multiple games

Key insight: We’re seeing 9-figure deals for single Roblox games. Dave Baszucki wants billion-dollar Roblox studios, and it’s going to happen sooner than you think.

Lesson 4: The Broader Opportunity

Don’t just port old games—adapt and refresh:

• Zynga couldn’t even port Farmville successfully to mobile

• Supercell won with Hay Day by rebuilding from scratch

• The winners understand platform-native design

The formula for success:

1. Find proven mechanics from 10+ years ago

2. Adapt for modern platforms and audiences

3. Add viral/social elements native to the platform

4. Iterate fast based on player behavior

Make sure to check out the podcast for a lot of really great insights!

3. GameDev: Actions vs. Intent, The Alien’s Eye View

For this week’s GameDev lesson, I want to share a powerful mental model that exposes the gap between what your team says it values and what its actions prove it values.

Chris Williamson’s concept of “The Alien’s Eye View” exposes the brutal truth about what we say and what we actually do:

“Imagine an alien watching you work. It knows nothing about your motivations or intentions—it can only see what you do. What would it think you value?”

• You might say the biggest priority is fixing the core loop.

• But the alien, watching your team spend 80% of its time on a new vanity cosmetic system, would conclude that cosmetics are what you truly value.

The insight is profound: Action eats intention for breakfast. We often say our biggest problem is X, but our weekly tasks are focused on Y and Z.

How to Align Action with Intent

At my studio, we’ve restructured our weekly product meetings to combat this drift and force alignment:

1. Monday Sprint Review: We kick off the week by verifying that our planned tasks align with our top-priority goals for the sprint.

2. Midweek Check-in: We discuss progress and brainstorm solutions, always through the lens of our primary goal.

3. End-of-Week Retro: We assess what we accomplished and ask the hard question: “Did our actions actually move the needle against our stated intention?”

This simple structure forces you to confront the Alien’s Eye View. It ensures your team is vector-aligned—with all its energy pointed in the same direction—rather than vector-neutral, where effort is scattered and velocity is lost. I highly recommend integrating this thinking into your sprint planning and retros. It’s one of the simplest ways to maximize your team’s effectiveness.

Most game developers are leaving money on the table—lots of it. Only 18% of the world speaks English, which means if your game is English-only, you’re missing 4 out of 5 potential players.

I recently sat down with Alexander Murauski, CEO of Alconost, a localization company that’s been helping game studios go global for over 15 years. Alex shared insights from localizing hundreds of games and revealed why most studios are doing localization completely wrong.

Key Takeaways:

• Start localization on Day 1, not as an afterthought - Building with 2 languages from the beginning solves 80% of future localization headaches

• The “Base 10” languages cover 80% of the world’s population - Chinese, Spanish, English, Hindi, Portuguese, French, Arabic, Japanese, German, and Russian

• AI reduces localization costs by ~30%, not 90% - Human review is still critical for quality

• Localization without marketing presence is wasted money - You need more than just translated text to succeed in new markets

• Use a proper localization platform from the start - Google Sheets breaks down after 10,000 lines and creates versioning nightmares

🎧 Listen on Spotify, Apple Podcasts, or Anchor

Speakers:

• Joseph Kim. CEO at Lila Games.

• Alexander Murauski. CEO at Alconost.

The Economics of Going Global

Why Localization Is Your Best Growth Hack

Alexander opened with a compelling Mr. Beast case study. By simply adding subtitles and voiceovers to his YouTube videos in multiple languages, the creator massively expanded his global audience. The same principle applies to games, but the impact is even more dramatic.

Consider these market realities:

• App Store visibility: Apple and Google won’t feature your game in local stores without localization. No Japanese translation? No Japanese App Store featuring. Period.

• Market expectations: In Korea, Japan, and China, 90% of users simply won’t accept non-localized products

• Organic growth multiplier: Localized games see significantly higher organic downloads from app stores

The math is simple: translating into Spanish alone gives you access to 550 million speakers. Chinese opens up over 1 billion potential players. Hindi adds another 600 million. These aren’t just numbers—they’re untapped revenue streams.

Choosing Your Target Languages: Strategy Over Spray-and-Pray

The Two Approaches to Language Selection

Alexander outlined two distinct strategies for choosing which languages to localize:

1. The Sophisticated Approach (for funded studios):

• Analyze unit economics for each market

• Use tools like App Magic or Sensor Tower to study similar games’ performance by country

• Calculate customer acquisition costs vs. ARPU by region

• Consider marketing capabilities and local presence

• Test markets with cheap auto-translations first, then invest in quality if ROI is positive

2. The “Base 10” Approach (for indies and smaller studios): Start with these language groups that haven’t changed in 20 years:

• FIGS: French, Italian, German, Spanish (covers Europe)

• CJK: Chinese, Japanese, Korean (covers Asia’s premium markets)

• Additional high-impact languages: Portuguese, Arabic, Hindi, Russian

The Hidden Cost Factor Most Studios Miss

Here’s what Alex emphasized that most studios overlook: localization alone won’t guarantee success. You need:

• Customer support in that language

• Community management capabilities

• Marketing presence (or at least partnerships)

• Understanding of local payment methods and pricing

Without these elements, even perfect localization might not deliver ROI. As Alex noted, “Japan is a high-paying country, but advertising there is very expensive. Indonesia has lower ARPU but much cheaper advertising costs. Sometimes Indonesia is the better choice.”

The Technical Foundation: Start Right or Pay Later

Day One Localization: The Game-Changer

Alex’s most emphatic advice: “Localization should be started on day one.”

Starting with two languages (even if one is just a test language) from day one forces you to:

• Build proper internationalization into your architecture

• Implement language switching from the start

• Extract strings into separate files instead of hard-coding

• Consider currency and payment processor variations

• Think about font support for different character sets

This approach solves what Alex estimates as “80% of your future localization challenges” before they become expensive problems.

Common Technical Pitfalls to Avoid

The Font Trap: Western developers using custom pixel fonts often discover they don’t support Japanese/Chinese characters. Asian developers face the reverse problem when localizing to Latin-based languages.

The German Problem: German text is typically 30% longer than English. Fixed-width UI elements that look perfect in English overflow catastrophically in German.

The Arabic Curveball: Right-to-left languages don’t just reverse text—they flip entire UI layouts. In Arabic Windows, the Start button is on the right.

The Context Crisis: “Cancel” can mean vastly different things depending on where it appears. Without context, translators guess wrong, breaking user experience.

Implementation: Tools and Workflows That Scale

Why Spreadsheets Are a Trap

Many studios start with Google Sheets or Excel for localization. According to Alex, this breaks down fast:

• Performance degrades after 10,000 lines

• No version control for multiple translators

• Zero context for where strings appear

• No structure for managing glossaries

• Manual processes that don’t scale

The Professional Stack

Instead, Alex recommends using proper localization platforms from the start:

Key Features to Look For:

• Git integration for automatic syncing

• Context screenshots for translators

• Translation memory to ensure consistency

• Glossary management

• Support for continuous delivery

• Multiple file format support

Recommended Platforms:

• Crowdin (Alex’s primary choice)

• Lokalise

• Phrase

• Smartcat

These platforms typically integrate with your build pipeline, enabling true continuous localization where translators can work in parallel with development.

The AI Revolution in Localization

How AI Actually Works in Practice

Studios are increasingly asking about AI translation, but Alex reveals the reality is more nuanced than “just use ChatGPT”:

What Works:

• AI can reduce costs by approximately 30% (not 90% as some expect)

• Different AI models excel at different languages (Mistral for French, DeepSeek for Chinese, DeepL for Polish)

• AI + human review delivers the best cost/quality ratio

What Doesn’t Work:

• Raw AI without human review (inconsistent terminology, hallucinations)

• Generic prompting without glossaries (key terms translated differently each time)

• Expecting AI to handle sensitive content (many models refuse war/violence terminology)

• Large glossaries that exceed context windows

The Optimal AI Workflow

Alex’s agency offers three tiers:

1. Premium: Human translator + human editor

2. Hybrid: AI translation + human post-editing (30% cost reduction)

3. Raw: Unedited machine translation (only for user-generated content)

The key insight: agencies like Alconost add value by knowing which AI to use for each language pair and content type, managing the technical pipeline, and taking responsibility for quality.

Working with Localization Partners

When to Use an Agency vs. Going Solo

Community/Crowdsourcing (like Minecraft):

• Pro: Free

• Con: Rarely achieves 100% completion, requires heavy management

Direct Freelancer Hiring:

• Pro: Direct control

• Con: Quality assessment challenges, heavy management overhead

Agency Partnership:

• Pro: Managed service, quality guarantees, established workflows

• Con: Higher cost than DIY

Evaluating Localization Agencies

Alex’s criteria for choosing a good agency:

Transparency: They should share translator credentials and create shared communication channels (like Slack) between your team and theirs.

Quality Verification: Request test translations and have native speakers on your team evaluate them if possible.

Technical Capability: They should guide you away from spreadsheets toward professional platforms and continuous delivery.

Long-term Thinking: The best agencies assign dedicated translators who stay with your game for years, becoming domain experts.

The Bottom Line: A 15-Year Case Study

Alex shared his favorite client story: a farm simulator that’s been running for 15 years. They started small with just two languages, gradually expanded to exotic markets, and now have translators who’ve been with the game for over a decade. These translators aren’t just vendors—they’re extensions of the development team who know the game inside out.

The lesson? Localization isn’t a one-time checkbox. It’s an ongoing investment in global growth that compounds over time. Start early, use the right tools, and build localization into your development DNA rather than treating it as an afterthought.

As Alex puts it: “Localization gives you scale.” In a world where 82% of potential players can’t access English-only games, can you afford not to localize?

Want to learn more about game localization? Connect with Alexander Murauski on LinkedIn or email him at am@alconost.com

Discussion Transcript

**Joseph Kim** *00:00:06*

Hi, Alex, and welcome. I’m glad to have you here. Today we’re going to be talking about game localization, a topic that many people probably put off for too long.

I wanted to start with an interview I saw with Mr. Beast, where he talked about the massive impact localizing his YouTube channels into different languages had on his growth and success. That brings us to the economic impact of localization. In today’s climate of efficiency and ROI, people want to know what that impact looks like.

Could you characterize how you think about it? For game developers, what does localizing into one or several languages really mean for the bottom line?

**Alex Murauski** *00:01:02*

That’s a great question. I’m familiar with the Mr. Beast case because my kids are big fans. They used to watch him in their native language, and now they watch him in English. What he did was brilliant. By subtitling and adding voiceovers to his video clips, he gained a huge audience.

If you look at the numbers, only 18% of the world’s population speaks English. This means four out of five people won’t use a game or watch a video if it’s only in English. In countries like Korea, Japan, and China, 90% of people will not accept products that aren’t localized. They find it uncomfortable and simply don’t understand it.

By localizing, you gain access to a much larger audience of players around the world. For example, translating into Spanish gives you access to 550 million speakers, which is more than the population of the United States.

The Chinese language gives you access to more than 1 billion people, and Hindi adds another 600 million potential players who can buy your product. That is the economic effect of localization.

Some game developers see localization as a growth hack because it drives more traffic from the app stores. If you want to get more organic players from the App Store, Google Play, or Steam, you need to localize. The App Store, for instance, won’t feature you in a local store if your game isn’t localized for that region. They check if your game is localized into Japanese, and if it’s not, you won’t be accepted. You simply can’t enter that market. Not localizing limits your organic opportunity to attract more users.

For smaller games and apps, it’s not that expensive to localize into the top 10 languages, which cover 80% of the world’s population. If you localize into Chinese, Spanish, English, Hindi, Portuguese, French, Arabic, Japanese, German, and Russian, you cover a vast majority of the people on Earth.

It’s worth localizing at least the initial user-facing elements, like the game description, screenshots, and the first few levels. Localization gives you scale.

**Joseph Kim** *00:05:10*

So, you view localization primarily as a way to broaden reach, which in turn broadens the audience that can monetize in your game. You mentioned a few key languages that can dramatically expand reach beyond the 18% of English speakers.

How do you and your clients typically decide which languages to prioritize? What kind of analysis goes into that decision, and where do you draw the line between adding a few languages versus fifteen?

**Alex Murauski** *00:06:10*

Simply localizing isn’t enough to be successful in a new region. It really starts with your marketing capabilities. For example, if you want to localize for Japan, even with a successful game, you need customer and community support. Without that, localization won’t be very effective.

The first thing to consider is how you will drive traffic to your game in that new market. If the cost of player acquisition is too high, localization might not provide a positive return on investment. Sometimes it’s better to localize for Indonesia instead of Japan. Japan is a high-paying country, but advertising there is very expensive. In Indonesia, you have a large population that may pay less per user, but the advertising costs are much lower.

The language selection comes after you determine what’s possible from a marketing perspective. The ideal scenario is to first plan your promotion and marketing, and then localize. If you already have a presence in Japan or Korea—marketing people, advertising opportunities, local partners—then localization becomes a clear necessity.

A more basic scenario is to translate into the top 10 languages. For Europe, this is often abbreviated as FIGS: French, Italian, German, and Spanish. Then there’s CJK: Chinese, Japanese, and Korean. You can also add languages like Arabic, which covers a large population but requires extra development work for its right-to-left layout.

So the decision comes down to two things: your marketing capabilities in a given region, and then selecting the most popular languages among internet users. You can also look at the GDP of different countries to assess their paying potential.

**Joseph Kim** *00:09:55*

Alex, let me pause you for a second. There’s a bit to unpack in your comments.

First, it’s important for our audience to understand that localization isn’t just in-game. You’re talking about all the localization touchpoints. There’s the in-game content, but also the app description for the App Store, marketing materials, and community channels like Discord or your website. The first point is that developers need to consider multiple localization touchpoints.

Second, regarding the original question about how to decide which languages to localize for, I heard two things. First, you mentioned understanding the attractiveness of a region. For example, even if Japan has an extremely high ARPU, you need to consider the potential ROI.

Personally, when I look at attractive geographies, I use a service like App Magic to check the data for similar games. I’ll do a country-by-country split and analyze those markets. You mentioned just going with the “base 10 languages.” Do you recommend that to your clients as a default, or do you suggest they do an analysis like mine, looking at data from similar games to determine the attractiveness of specific countries?

**Alex Murauski** *00:12:03*

Regarding those 10 languages, that advice has been around for 20 years. Nothing has changed. Every game localizer would say to just go with those 10 languages for maximum impact.

However, if you’re a sophisticated game developer who cares about unit economics, or if you have investors looking for a return on marketing and advertising investment, you need to assess conversion across different channels. In that case, you would use services like App Magic or, for games and app stores, Sensor Tower, App Annie, or Similarweb.

You dig into the numbers. But that’s really the task of a chief financial officer, not an indie game developer.

**Joseph Kim** *00:13:11*

There’s one other point I wanted to bring up. You suggested that before localizing, you should also consider the required investment beyond just translation. You mentioned needing a marketing team in India for Hindi, or in Japan for Japanese.

So if you do localize for one of these countries, what do you recommend in terms of an on-the-ground presence for marketing or localization?

**Alex Murauski** *00:14:02*

I would definitely recommend having people responsible for marketing in each region. That’s a must.

This applies to larger studios. When they ship their game, they need to work with influencers and handle advertising in different markets. This requires a large marketing department or a dedicated marketing agency.

For small indie developers, the resources and opportunities are very limited. They can usually only publish their game on Steam, select all the languages Steam supports, and provide localized content like screenshots and game descriptions. That’s pretty much everything available to them. They just spray and pray.

**Joseph Kim** *00:15:24*

That sounds like most of the games industry.

**Alex Murauski** *00:15:30*

Yes. We work with many different game development studios, and 90% of them take the “localize to every language we can” approach to cover everything, everywhere. Sometimes it works, because people on the app store download more and it pays off.

But some companies are driven by finance, investors, and return on marketing investment. They test the waters first. For example, to test a market like Indonesia, they might auto-translate some content with Google or ChatGPT and publish it. If they see a positive return on investment, they will then redo the translation with higher quality.

They might also add a representative for community support or do more to get local coverage and influencers to stream the game. They start by stepping on very thin ice, testing the waters by localizing just a few assets to see what happens.

There are also A/B testing platforms like SplitMetrics. You can test how downloads or installs are affected by having a localized store page versus a non-localized one. Before you invest in full game localization, you can invest a small amount in this kind of testing to see how it might work.

**Joseph Kim** *00:17:46*

I would consider myself not deep enough on the topic of localization, and it seems there’s more to it than I had originally considered. I’d like to ask you a few more questions about that.

I can tell you how we’re doing it, which is probably not ideal. We tend to delay localization until the back end of the development cycle. After focusing on all the core game development issues, we integrate localization towards the end. It’s not quite an afterthought, but it’s close.

As an expert who knows the best practices, what is the best timing and process to maximize efficiency and reduce time and cost, instead of treating it as a last-minute task before launch?

**Alex Murauski** *00:19:06*

Yes. This is my best advice for any developer, whether it’s a game, an app, or a website: localization should be started on day one.

**Joseph Kim** *00:19:22*

Okay.

**Alex Murauski** *00:19:23*

If you have any ambition of localizing your game—whether into two, three, or even 30 languages—you should start with two languages from day one. One of them can even be your second language or an artificial one.

Starting with two languages immediately forces your game to be technically internationalized and ready for 20 languages, not just two. It compels you to think about how languages will be switched, so you implement a language switcher. You learn to use the right SDKs and avoid hard-coding strings into the code, instead extracting them into separate files. Supporting two files in parallel is manageable.

This process establishes good development hygiene from the very beginning. You learn to live with localization from day one, considering things like date, time, and number formatting. The SDKs and the programming language you’re using will have internationalization guidelines to help you. You’ll also think ahead about currencies and implementing different payment processors for various markets.

You’ll also have to think about fonts, which can be tricky. A common mistake occurs when a game is developed by Japanese or Chinese developers who use hieroglyphs. The fonts that support them are larger than those used in the Latin-based world. But the real problem arises when a Western developer tries to localize their game into Japanese. They may have a custom, immersive pixel font that only contains Latin letters. When they switch to Japanese, the font doesn’t support the vast number of new symbols. It’s rare for a custom font to support everything.

Another example is localizing to Arabic, where the layout is right-to-left. In an Arabic version of Windows, the “Start” button is on the right, not the left. Just being aware of these issues from the start, by working with two languages, solves 80% of your future localization challenges.

Adding more languages later becomes as simple as adding more files to your localizable assets folder. At that point, managing even two languages manually will start to feel tedious. Adding a single line of dialogue means inserting text into two separate files. This annoyance will naturally lead you, as an engineer, to a solution like a localization platform.

A localization platform integrates with your GitHub repository, takes your source language resource file, and allows it to be duplicated and translated into any number of languages automatically. It merges all the changes for you. You only have to work on the source file in GitHub; you make a commit, and you get a pull request from the localization platform with all the translations. The pain of managing localization is gone. It just works.

If you set up continuous delivery and nightly builds, you can integrate localization directly into your build pipeline. Your game gets localized almost automatically. With machine translation, it can be instantly localized into any number of languages. If you add human translators, they can work on the localization platform and have your updates translated by the next day. You get your localized build in almost no time.

This is how it should work. You don’t postpone localization; you make it continuous. My best advice is to start with two languages, and this will resolve almost everything else.

**Joseph Kim** *00:26:24*

So, your argument is that it’s better to think about localization from day one from the perspective of scalability. If you consider it from the beginning, you don’t have to go back and refactor code or redo art assets later.

For instance, you shouldn’t bake text into images, and you have to consider how language will be presented in the UI. Thinking ahead makes the entire process much more scalable and easier to manage in the future.

You mentioned fonts, and I remember one of the biggest UI problems for us in the past was German. We had a fixed amount of space, and the German text would always overflow.

Besides not planning for scalability from the beginning, what are some other common problems and mistakes you’ve seen when meeting with different studios? What can we avoid?

**Alex Murauski** *00:28:06*

There are two types of errors: technical and translation. A technical error, for instance, is when German text overflows because it doesn’t fit the size of a button. This can usually be fixed by the developers themselves, as they can see that something is visually wrong without needing to know German. No language proficiency is required for that kind of QA.

The trickiest mistakes in localization come from mistranslations, which often happen when a translation is done without context. For example, a button captioned “cancel” can be translated in many different ways and have different meanings depending on the language and context. Is it a verb, or are you confirming a cancellation? Without context, a translator can easily get it wrong.

The best way to solve this is by structuring the resources for translation. Instead of just an Excel file with strings, you could use a structured format like a JSON file. The structure could be something like: game screen one, save dialog, save button, with the caption “Save.” This structure provides context for the string. It tells both the translator and the developer that this specific “Save” is used here and not somewhere else in the project.

Another solution is providing screenshots. Some localization platforms can automatically find screenshots that match the strings being translated using AI or machine vision. This allows the translator to see the context, because while they always play the game, a universe can be too big to navigate to every single screen. We hunt for these mistranslations during the quality assurance stage, after the game is localized and we have a test build.

Finally, term consistency is critical. If characters in the game have names and reference each other, those names must be translated consistently, or it breaks everything. The same goes for artifacts or location names. If you have an axe that is later referred to as a hammer, it creates confusion.

This is why having a glossary—a book of every term and item description in the game—is a must. It helps ensure consistency throughout the story. The glossary and a style guide are common practice, and when we start localizing a game, the first thing we require from a client is to build a glossary. If they don’t have one, we build it for them.

**Joseph Kim** *00:33:11*

You mentioned a glossary and a style guide. I understand what a glossary is, but how does a style guide differ?

**Alex Murauski** *00:33:21*

A style guide provides context for the translator by describing the game itself. For example, it might specify how to address the player—formally or informally. This is very important in languages that have different forms of “you.”

The style guide also covers things like translating humor. Should it be recreated for the new culture, or translated literally with a reference? It provides context about the game that isn’t found in the text itself.

**Joseph Kim** *00:34:31*

Shifting to practical implementation, I see many small game studios use an Excel or Google spreadsheet. They’ll have English in column A, Spanish in column B, Japanese in column C, and so on.

You’ve identified several common problems with this approach. First, you mentioned a lack of structure and suggested a JSON file might be better to show where the text actually goes. Second, you talked about context. It would be helpful to see a screenshot of where a specific text string is used. And third, you mentioned the context of style, like whether the tone should be formal or informal.

With those points in mind, what is the best practical implementation? How can smaller studios avoid the common problems of the Google Sheet approach, and what is a better alternative?

**Alex Murauski** *00:36:44*

Google Sheets is a convenient and obvious way to manage localization. You can have tabs or columns for different languages. However, the main problem is performance. When a spreadsheet has more than 10,000 lines in one language, it can become very slow or even fail to open.

They are also easy to damage. If someone deletes something, it can cause problems, and while there is version history, sharing it with multiple translators risks corruption. For a very small project at the beginning, it might work. But the bigger issue is that every professional localization platform requires a key.

**Joseph Kim** *00:38:00*

What do you mean by “the key”?

**Alex Murauski** *00:38:02*

The common standard for localization resources is to store them as key-value pairs. The key is the name of the string, and the value is the string itself.

**Alex Murauski** *00:38:15*

Using key-value pairs is a common and structured approach. If you have a key, you already have some structure. For instance, you can use dots within a key to provide context that a plain string in a spreadsheet lacks. With a key, the developer knows the context of the string.

In a spreadsheet with no keys, the developer doesn’t know where that line of text appears in the code. This is unmanageable, though it might be okay for a very small game.

There’s also a terrible format: PO files, which are often used for website translation. In PO files, the source text itself acts as the key. This means if you fix a simple typo in the source text, the key changes, and all of your existing translations for that string disappear because the app can no longer find them.

Using spreadsheets can be convenient for calculating translation volume or sharing with freelance translators, but it only works for a very small amount of text. For any real project, your files must have a key-value structure. The best approach is to use a localization platform that organizes all your files in one place, so you don’t have to manage them manually.

**Joseph Kim** *00:40:13*

What is an example of a localization platform?

**Alex Murauski** *00:40:17*

There are plenty of them on the market. Some of the best are Crowdin and Tolg. There’s also Phrase and Smartcat. A simple Google search for “localization platform” will give you many options. We mostly use Crowdin in our work.

What developers like about these platforms is that they support every file format. You can even use your spreadsheets, but the spreadsheet just becomes a container for your localized resources. The actual localization work happens on the platform itself. The platform provides tools like a glossary, style guide, translation memory—which is very important—and integrations with various machine translation engines and modern LLMs. A localization platform allows you to organize your localization pipeline properly.

This leads to my second piece of advice: if you’re localizing for even two languages, use a localization platform from the very beginning. It doesn’t cost a fortune, but it organizes the process and makes it professional. We use Git repositories to store our code and professional IDEs to develop. We should use professional tools for localization as well, rather than just manually tossing strings into spreadsheets.

**Joseph Kim** *00:42:22*

Got it. Localization isn’t something a typical game studio does on a continuous basis, so working with a partner vendor like yours makes sense.

When choosing a partner, how should game studios evaluate different localization vendors? What metrics—like cost, turnaround time, or quality—should they consider when looking at a company like yours?

**Alex Murauski** *00:43:13*

Game development companies have several options for localization. The first thing developers often consider is hiring freelancers to do the job for free. This typically involves crowdsourcing, where members of a game’s community volunteer to translate in exchange for being credited. Minecraft is a famous example, with 80% of its translation done by the community. This is often the first approach indie developers consider.

However, community translation has its limits. Translation is real work, and for a large game, it’s unlikely that volunteers will complete 100% of the project without payment. The enthusiasm often isn’t enough to cover everything. This approach also requires management, including an editor and tools for managing glossaries and translation memory. It’s a tricky process, though some games are successfully translated this way.

The next option is to hire and pay freelance translators directly. This requires a localization manager on the developer’s team, a role that might be filled by a marketing manager, a content manager, or even one of the developers. The challenge here is finding reliable, cost-effective translators who deliver high-quality work. It’s also difficult to assess quality if you don’t speak the target language, so you might not discover a poor translation until users complain.

The best option is to work with a localization agency that handles everything. An agency has a pool of experienced game translators for dozens of languages, all with a proven track record.

When choosing an agency, I recommend prioritizing transparency. A good agency will tell you who the translators are, perhaps by providing blind CVs or their names. This makes the linguist part of your team, ensuring consistency for years to come. We have cases where the same translator has worked on a game for a decade. They know the game inside and out and work as needed, whether it’s one hour a week or one hour a month.

The agency provides both the pool of translators and the localization platform. If you don’t have a platform, the agency will recommend one and help you build the right process around it. You also get a dedicated localization manager at the agency who serves as your single point of contact. This person takes full responsibility for the project, which is crucial because it frees up the game developer’s internal resources.

So, how do you choose the right agency? First, as I said, look for transparency. They should be willing to share information about their translators. We often create shared Slack channels with clients, linguists, and project managers to speed up communication and problem-solving.

Second, ask for a test translation. This is especially useful if you have someone in-house or a loyal user who can assess the quality. Many companies come to us with 20 languages and no prior relationship, so a test translation is a great way to build trust and ensure quality. Translation style can be preferential, so we allow clients to choose from different linguists to find the best fit. If a game studio has a native speaker, like a community manager, they can assess these test translations and help build the perfect team for their project.

A good agency already has these teams ready. The key is that it shouldn’t be a black box. It must be transparent and technically equipped to support you.

**Joseph Kim** *00:50:53*

Can you give me an example of what it means for an agency to be “technically equipped?”

**Alex Murauski** *00:51:00*

For example, if you come to an agency with only a spreadsheet, they should advise you to use a proper localization platform. Even if your internal infrastructure is built around spreadsheets, the translation work itself should happen on a professional platform, which they can often provide.

This professionalizes the entire process. The platform manages glossaries, translation memories, and automated quality assurance checks. It enables structured pipelines, such as having one person translate and a second approve, or having a machine translate followed by human approval.

It’s impossible to manage this level of professional work by simply sharing spreadsheets among 20 people. That’s what I mean by being technically equipped.

**Joseph Kim** *00:52:15*

Okay, let me summarize what I understood. Your first point is that developers have several options. One is free, crowdsourced translation, but not every game has the popularity of Minecraft to make that work. A second option is hiring freelancers, but that comes with significant management overhead, variable quality, and a lack of awareness of best practices.

When it comes to choosing an agency, you recommend evaluating them on a few key criteria. The first is transparency—getting high visibility into who the translators are and what the communication processes look like.

Second is quality, which can be verified by requesting a test translation. This ensures the agency can meet your quality expectations.

Third is technical support—ensuring the agency has the right tools and infrastructure to handle localization professionally. Did I miss anything, or is that a good characterization of your points?

**Alex Murauski** *00:53:46*

On the technical side, the continuous delivery cycle is very important. The real pain starts when just two new lines are introduced into the game, and the agency needs to be able to translate them by the next day. Two lines in 20 languages means 20 people will be involved. Without a localization platform and proper management, this small change can break the entire communication and delivery pipeline. This is why continuous support is crucial.

Another important thing is quality assurance. There are several quality gates, starting with automated checks on the translated content. Then, a second linguist reviews the materials. Finally, there is localization quality testing, where linguists test the game in their regions, with their language enabled, to check all the screens and report issues. A quality localization agency will always discuss the full localization cycle, including testing.

**Joseph Kim** *00:55:52*

I have a few follow-up questions. We’ve been talking about in-game localization, but you also mentioned other touchpoints. It isn’t just in-game content; you have creative assets for marketing, community content, and ASO app store descriptions.

For your customers, how much of your business is in-game versus those other assets? And what do you generally do for those other localization touchpoints?

**Alex Murauski** *00:56:37*

For game development, 99% of our work is localizing in-game content. The best clients are games that live for 10 or 15 years and have seasonal events, like Halloween or New Year’s parties, to keep the content fresh.

When we work with apps or SaaS businesses, it can be more complicated. We handle both product and product marketing localization. This means translating the product itself, as well as related materials like help systems, reference guides, blog posts, advertising materials, and case studies.

This is rare for games, however. For games, the non-game content is usually part of the user journey. Developers typically have a small website or a simple landing page that’s easy to translate. It’s mostly app store descriptions, and that’s about it.

Influencer streaming is generally not localized, though some creators are starting to use AI for that. Overall, the user journey content outside of the game itself is very limited.

**Joseph Kim** *00:59:24*

I have one more follow-up question regarding the different localization solutions, and I think this will lead us into a bigger discussion about AI. You mentioned a few options. Crowdsourcing for free is rare. Some studios use freelancers. Many studios use agencies.

But an increasingly popular option, especially for smaller studios, seems to be AI. For example, you could have a spreadsheet with English text and use a service like ChatGPT or Claude to fill in the translations for other languages.

Can you talk about the use of AI in localization? Do you use AI? And what would you say to a studio that thinks they can just handle it all with AI instead of an agency?

**Alex Murauski** *01:00:54*

That’s a very good question. We definitely use AI. And AI isn’t just for small indie developers; the big studios with huge amounts of content are also using it.

In fact, they’re the most interested because they have the resources to choose between different LLMs and train them to produce better translations for their specific content. The biggest developers are the most interested because they have large budgets they want to cut.

**Alex Murauski** *01:01:48*

Clients were the first to come to us and ask if we could translate everything with ChatGPT. It seemed like it might work, but not in the way they imagined.

How it really works now is that you can translate with ChatGPT directly within a localization platform. You turn on automated translation, and you’ll get your output from ChatGPT, Claude, Gemini, or any other LLM, as they are all supported. It will translate well for some languages, but there are some tricks.

Not every LLM translates well into every language. For French, Mistral performs better. For Chinese, DeepSE is obviously superior. For most European languages, ChatGPT is okay, but for Polish, DeepL is still the best, even though it’s a traditional translation engine and not an LLM.

Selecting the proper engine is a key part of our service. Knowing the client’s content type and language pairs allows us to choose the best AI for the job. But AI doesn’t produce 100% quality on its own. It must be prompted properly, and the glossary, style guide, and a human reviewer are all critical. The reviewer checks all the translations. You could use basic machine translation, but then no one is responsible for the outcome. As an agency, we are responsible for the quality, so we must have that human check.

We negotiate with the client, explaining which AI is best for their specific content and language pair, and show them the test metrics to back it up. We agree to use AI with a human reviewer, which results in a price that is about 30% less. It’s not 90% or 100% less, because the reviewer is still reading and editing, which is not a huge time difference from translating from scratch. Even in the past, with translation memory, a human still had to review every text chunk to give it a seal of quality. What we sell is responsibility and quality. We guarantee our localization is good, and the client is paying for that insurance.

We are transparent about using AI when we negotiate. When a client asks for a price, we usually present three options. The first is a professional human linguist, plus another editor for review, depending on the budget. The second option is Machine Translation Post-Editing (MTPE). We agree to use a specific machine translation engine plus a human reviewer to cut costs for certain languages.

The third option is raw, unedited machine translation. This is only suitable if the client has the right content type for it, such as user-generated content in a chat or Airbnb comments. Nobody reviews those with a human; they are just machine-translated. The client chooses from these three options and always knows if we are using machine translation.

Some game developers rely 100% on ChatGPT and program translations through the OpenAI API. We were initially very afraid of this scenario as a business, but we’ve seen that it’s hard to automate translation effectively. You get unpredictable results.

Without a glossary, your translations will be inconsistent, with the same word translated in different ways depending on the LLM’s creativity settings. The models also hallucinate, inventing things that aren’t in the original text. Fixing this costs money. If you have a large glossary, it won’t fit into the LLM’s context window. You have to implement a system to retrieve parts of the glossary and insert them into the prompt. Furthermore, if you need to translate sensitive terminology about war, bombs, or chemicals, ChatGPT is often prohibited from discussing those topics and will refuse to provide a translation.

**Joseph Kim** *01:10:02*

Let me recap. It sounds like while some game studios use AI directly, the advantage of working with an agency like yours is that you also leverage AI, but you know which models are best and how to optimize them. You pass those cost savings—around 30%—on to your customers.

The value you add is in quality control, since AI isn’t perfect and can hallucinate. You also provide best practices and ensure the AI is used optimally. Is that a fair characterization?

**Alex Murauski** *01:10:54*

Yes. You get quality assurance for the localization, testing of the localized game build, and the technical setup, which includes choosing the right model.

You don’t have to write code to use LLMs for translation. You can’t just copy and paste a large JSON file into the ChatGPT window—it will get corrupted, and doing it for 20 languages is tedious manual work. The alternative is using the API, which requires development. You would essentially be reinventing the wheel. A localization provider handles all of this, and that platform access is basically free.

**Joseph Kim** *01:12:19*

To help our audience understand what it’s like to work with an agency, could you walk us through a case study? Please take us end-to-end, from the very start of a client engagement to the finish. What does that process look like?

**Alex Murauski** *01:12:48*

Every client is different. Let’s talk about one of my favorite clients, a studio that has been with us for over 15 years.

They started very small with a farm simulator game. You plant seeds, gather fruit, make juice, and sell it to grow your economy. It’s an evergreen game with characters that has kept users returning daily for 15 years.

They started with two languages for the US market, and it was a hit. At the time, it was a web-only game. We met them at a game development conference, which is a common way to meet new clients. We started by using spreadsheets. From the very beginning, a key issue was that some of their existing translations weren’t done by native speakers.

They had a large user base and active community managers. To go global, they decided to add four languages: French, Italian, German, and Spanish. This was when localization platforms were just beginning to appear. Our main value proposition was to organize an automated delivery pipeline for them.

We built a mixed team where their community and content managers worked alongside our localization managers in a shared Slack channel. They all collaborated within the localization platform, looking at the same resources. From that platform, their developers can now export string resources in different formats for iOS, Android, and web apps. It serves as the central repository for all their localizable content.

We know their monthly release cycle. Their writers have a long-standing process of creating content in Excel spreadsheets. Those spreadsheets are then uploaded to the localization platform, where they are translated, checked, and returned to the client.

As they needed more languages, including some exotic ones, we added more linguists to the team. About two years ago, they wanted to adopt AI-driven translations. We worked with them to build an AI workflow and pipeline. This happened at just the right time, as this type of AI-involved pipeline was becoming the new industry standard.

We followed our client’s lead, whether it was adding an exotic language, performing additional testing, or helping their community managers integrate teams. We created a system where their in-house translators and our translators could check each other’s work. We’ve worked with them for a lifetime, and our linguists have become an extension of their development team. They are players who know the game inside and out; they are the keepers of its localization. That’s why it’s my favorite case.

But the common journey for a new client can be quite different. It usually starts with a price inquiry. We get leads every day from studios asking how much it would cost to localize a game with a specific word count—say, 2,000 words or even 2 million words—into ten languages. When a client asks for a simple price quote, it often tells us they are already mature in their localization process. They know about quality, platforms, and setups; they just need the price.

As I mentioned, we always offer three options. We start by interviewing the client to understand what they really need. Every game has different content types, so we ask them to prioritize what needs the highest quality translation and what can be done with less. From there, we build a localization roadmap with a clear timeline. We create a Gantt chart showing all the stages so the client can see exactly how the project will be implemented.

Regarding pricing, it’s volume-based for text translation and hourly-based for localization testing. The more content a game has, the bigger the budget will be. Prices vary a lot, from six to eighteen cents per word, depending on the language. However, no one gives a final price without seeing the content, the game itself, and knowing which linguists will be involved.

When we create an estimate, we already have specific linguists in mind. If a huge game with a million words requires five linguists and months of work, we can often negotiate better rates with them, which translates into a better price for the client.

The process always starts with an interview to understand the client’s needs, followed by a rough estimate of price and time. Sometimes, a client will request test translations to select their preferred linguists. We provide samples from three to five high-quality linguists with different styles, allowing the client to choose the best fit. This is how we form the team.

Once the team is formed, we begin. We get the localization kit from the client—which could be anything from Excel files and JSON files to an existing project on a localization platform. We can work with anything.

**Joseph Kim** *01:24:16*

All right, great. That’s all the questions I have about your business. I thought maybe we could end with a little more of your personal story. We’re all on this earth trying to do interesting things and create value in the world.

Could you talk about how you got into this business and how you founded your company?

**Alex Murauski** *01:24:44*

My background is in mathematics and software development, and I still do some development at Alkonost. My passion is product development.

Over 20 years ago, I built a product and started selling it worldwide. It was obvious to me that the software had to be in English to sell well, so I translated it into languages I knew. I had several friends in a software developer community who were doing the same thing, but their software wasn’t localized. They asked me to help them translate it into English, and I did.

Then, I recommended my girlfriend, who translated software into German, to my friends. They started ordering from her, and I would simply pass the money from them to her. Everything was great until one of the clients said the translation quality was poor and demanded a refund.

My girlfriend was a native German speaker but not a professional translator. I wasn’t selling translations; I was just helping my friends. But because they were sending money through me, they considered it a service. That’s when I understood that if you take money for something, you bear responsibility for the quality. That moment, it became serious. I realized if I deliver something for payment, it must be high quality, so I found a better translator.

It started as a little business game for me, not too serious, because I had my day job as a software developer and revenue from my own software. This was a side hustle. But then I wrote an article in a local newspaper about my experience, and suddenly, translators started asking me to help them find work.

I told them I could give them translation work, but I would charge 50% of what the client pays. That was my margin, and that’s how it started. I announced in the community that I had translators for four languages who could help with software translation. I organized all the technical aspects—this was before localization platforms, so we dealt with XML files, .ini files, and all sorts of things that were easy to break.

I managed the technical side and communication with clients, and my translators did their job. It began with me translating software myself, then adding some friends, and it gained traction. More languages and more orders came in. I started doing some marketing and sales and hired managers to handle it all.

Last year, Alkonost had more than 100 full-time employees, who are our core team of localization managers. We also have 1,500 translators all over the world. That’s how the company grew from that small start.

**Joseph Kim** *01:29:38*

At what point did you decide to go all in? I assume with that much scale, you decided you shouldn’t be a software developer anymore and should focus on this business. What was your thought process when you decided to make that switch?

**Alex Murauski** *01:29:57*

It wasn’t money that made me decide. I had a very good salary at my day job as a software developer, and I also had revenue from selling my own software. But I understood that I had to build a company. With a day job, you’re alone, and you have this glass ceiling on how much you can earn.

Selling my own software as a solo developer was also limiting. It’s easier now with AI-assisted coding, but back then, developing products was very hard work. It didn’t provide enough money to hire someone, and I didn’t trust anyone with my source code. I was worried they would just go and sell my software on their own.

Somehow, I understood that I needed to be a company, not an individual. Building a company is a completely different thing than solo software development. It was a childhood dream of mine to build a company with a name and see it grow.

The realization came when I was negotiating with translators and with developers whose software was much more successful than mine. I was helping them localize, and I found myself surrounded by people who were creating things. I understood that a company was being born right at that moment. One of my translators became a manager, and it became the two of us, plus some translators. Then we invited more people, and it grew organically without any outside investment.

When I decided to commit to my own company, I quit my job with its very good salary. I had no money, was living in my parents’ house, and was even taking money from my mother to pay salaries.

While developing Alkonost, I was also involved in different startups, always trying to build something new. This DNA of doing something else is still with us in the company today. We are always inventing and developing products inside the company, not just focusing on localization.

For example, we use GPT, Gemini, and other LLMs to assess translation quality. We built a SaaS service that can assess the quality of translations using modern AI, and it’s available for free right now. We also have our Nitro translation service, where you can top up your balance, send a few lines of text, and have them translated into 20 languages by native human translators. We developed that system on our own.

This development culture helps us understand how developers think. When a client comes to us, we understand how they would organize their files for localization. This understanding is our key differentiator. We’re not a translation agency that translates everything, like marketing documents. We’re a localization agency that knows how developers think, how they work, and what they need. That is the main reason Alkonost exists and has been successful.

**Joseph Kim** *01:35:14*

I have one last question. It sounds like you’re developing an AI evaluation product for AI-powered translations. When you think about the future of Alkonost, say five years from now, where do you see the company heading?

Do you think you’ll have more of a product focus rather than a service focus, or will it be a blend? Are there other areas you plan to explore? I’d love to understand how you think about the future of your company.

**Alex Murauski** *01:36:11*

The future of Alkonost is absolutely in localization and languages. That is our main expertise.

**Joseph Kim** *01:36:19*

Sure.

**Alex Murauski** *01:36:21*

But it’s not only about the service; it’s also about the technical side and creating products around localization. For our technical stack, we started the Alkonost MT project, a lab where we experiment with AI and deliver different localization tools. This helps us stay on the cutting edge. With every new version of AI that comes out, we implement it into our technology stack.

I see Alkonost as a localization services vendor. As I like to say, we are like a tree, growing organically when there is sun and water. This is common for localization companies worldwide. We’re in a very competitive global market, with at least 100 direct competitors.

We will continue to grow organically, occasionally disrupting the market with new tools, platforms, or platform features. I believe that is the most organic path. We don’t have super big ambitions. I’ve been here for 20 years and have had lots of ambitions, but what has always worked is Alkonost doing localization.

As an entrepreneur, if it could be separated from the company, there might be some side projects. But those would be different businesses or different companies. I have plenty of other ideas, but when it comes to Alkonost, it will be about localization forever.

**Joseph Kim** *01:38:51*

Got it. Specialization and depth—that sounds great. Alex, that’s all the questions I have. Thank you so much for going into depth about your business and providing more context on localization. I don’t think people consider it enough, so I really appreciate the information you’ve shared today.

If people want to get in touch with you, how should they do that? Should they visit your website or find you on social media?

**Alex Murauski** *01:39:32*

The best way is LinkedIn. You can also email me at am@alkonost.com.

**Joseph Kim** *01:39:50*

Great. To our audience, thank you for your attention. We’ll catch you next time. Thanks, everyone, and thank you, Alex.

**Alex Murauski** *01:40:01*

Thank you, Joseph. Thank you for having me.

In today’s newsletter, we’re diving deep into three critical topics that are shaping the industry right now.

• Macro: We’ll break down how Delta Force on mobile managed to outperform the legendary PUBG by a staggering 75% on recent revenue peaks. How did they do it, and what does it mean for the shooter genre?

• Alpha: An intriguing report from Investgame claims a “Take-Two Mafia” is succeeding in mobile. We’ll examine this thesis, show why I disagree, and uncover where I believe the real alpha is hiding.

• Game Dev: I’ll share a lesson that’s very close to my heart—why recruiting is often broken and how applying a product management framework from Tony Fadell’s book Build can create a massive competitive advantage.

Let’s get into it. 👇👇👇

→ Listen on Spotify, Apple (More Gamemakers content on YouTube)

1. EA in Advanced Talks for $50 Billion Buyout (PocketGamer.biz): Electronic Arts is reportedly in advanced discussions with investors, including Saudi Arabia’s Public Investment Fund (PIF), for a potential buyout valuing the company at $50 billion. If completed, it would be one of the largest acquisitions in video game industry history.

2. The Take-Two Diaspora: Why Alumni Win Big in Mobile, Not AAA (investgame, Anton Gorodtesky): A report by InvestGame and GDEV suggests startups from Take-Two alumni have secured $3.7 billion in deals since 2020. The analysis shows this success is concentrated in the mobile sector, led by veterans from subsidiaries like Peak Games, contrasting with Take-Two’s AAA focus.

3. Delta Force Overtakes PUBG Mobile in Historic Revenue Milestone (AppMagic): Delta Force achieved an unprecedented 75% revenue outperformance over PUBG Mobile on September 20, generating $6.48M versus PUBG’s $3.68M, marking the first time any mobile shooter has significantly challenged PUBG’s dominance.

1. Macro: How Did Delta Force Outperform a Titan?

We just saw a truly unprecedented event in the mobile shooter space. Our partners at AppMagic flagged that Delta Force had a monster run, outperforming PUBG Mobile by 75% on certain days. While Garena Free Fire has occasionally topped PUBG, it has never been with this magnitude or sustained over several days.

This raises some huge questions. Is a new king being crowned? And how did a relatively new game pull this off? Let’s look at the data.

First, some context on the giant they’re challenging.

• PUBG Mobile: Launched in April 2018, this game is one of the most successful of all time. It has generated over $9.2 billion in lifetime net revenue and boasts over 1.2 billion downloads. It’s a cash machine, averaging about $3.4 million in daily net revenue.

Now for the challenger.

• Delta Force on Mobile: Launched in September 2023, with a global release in April 2024. To date, it has generated $224 million in lifetime net revenue from 47 million downloads, averaging around $1.1 million daily since its global launch.

The Surprising Story in the Data

When you compare their launches, PUBG had dramatically higher downloads and daily active users. That’s not a surprise, as hitting massive scale was easier back in 2018.

However, the revenue story is shocking. Despite a massive gap in users, Delta Force generated significantly higher revenue at launch than PUBG did. This suggests its per-user monetization is on another level, likely driven by superior design and LiveOps.

The real explosion happened around September 20th. Delta Force’s average daily revenue of $1.1 million spiked to over $6.5 million on September 20th—a 6x increase.

So, what caused this massive surge?

• Esports Integration: The spike was driven by the game’s first-ever esports world championship, the Delta Force Invitational, which had a $500k prize pool.

• Sophisticated LiveOps: The event was supported by a “Pick’em Challenge” letting players wager on matches, a massive server-vs-server competition, and a dedicated tournament store. This shows how much LiveOps and esports integration have evolved into powerful monetization engines.

Key Takeaways:

1. The China Factor is Huge: Both games earn the majority of their revenue from China, but the skew for Delta Force is extreme. About 97% of its lifetime revenue ($217M of $224M) comes from China. For Western studios, this highlights both the immense opportunity and the challenge of being effectively locked out of the market without a local partner.

2. LiveOps & Esports Drive Revenue: In today’s attention-constrained economy, generating more revenue from existing players is critical. The scale and complexity of the Delta Force event is a masterclass in how to drive monetization spikes far beyond baseline levels.

3. Watch the “Splinternet”: We’re already seeing a digital world fractured along national lines, especially with China. Recent regulatory moves, like the forced sale of TikTok’s U.S. business, show a real risk of regulation dramatically altering the competitive landscape. This is something every global gaming company needs to keep on its radar.

2. ALPHA: The Real Mobile Gaming Mafia Isn’t Take-Two—It’s Peak Games

A recent report from market intelligence firm Investgame made a bold claim: “when [Take-Two’s] veterans launch their own studios, a surprising pattern emerges: the most significant wins come not from replicating AAA glory, but from mobile studios.”

Investgame’s thesis suggests that Take-Two confers “mafia” effects to their alumni that leads to winning in mobile vs. AAA.

This immediately struck me as odd. Take-Two is a publisher with a portfolio of studios that, by all accounts, operate as independent city-states. How could a holding company imprint a unified “mafia effect” on its alumni?

While I applaud the effort and line of investigation, I believe the report’s conclusion is a stretch. But in digging into it, we can find some real alpha.

The Flaw in the Thesis

The report’s logic seems to be built on the massive success of one company: Dream Games. The analysis suffers from a few critical issues:

• The Publisher vs. Studio Problem: Game-making DNA—the culture, processes, and talent that drive success—resides in the studio, not the publisher. It’s hard to see how a “Take-Two effect” could exist across its independent studios.

• The Single Data Point Problem: The conclusion is almost entirely driven by the outlier success of Dream Games.

• The Timeline Problem: Dream Games was founded in 2019 by alumni from Peak Games. Zynga acquired Peak Games in 2020. Take-Two acquired Zynga in 2022. The organizational overlap never happened. The success of Dream Games is a Peak Games story, not a Take-Two story.

If you remove the Dream Games and other Peak Games alumni data, the thesis flips. The organic Take-Two alumni network actually favors console and PC development.

Finding the Real Alpha

So, I believe the report’s conclusion is wrong, but the line of thinking is valuable. Here are the key lessons:

1. The Peak Games Mafia is Real: There absolutely appears to be a “mafia effect” at play here, but it originates from Peak Games. This is the real alpha. We should be studying Peak to understand what in their DNA enabled such repeatable success. Was it their people, their practices, their culture?

2. There Are No Infallible Experts: We often look to market intelligence firms or legendary studios as if they have all the answers. But everyone makes mistakes—I make them constantly. In an industry as dynamic as gaming, our goal should be to question assumptions and seek the truth together.

JK Note: I’ve offered to collaborate with Investgame on a future study of Peak.

3. Game Dev: Your Most Important Product is Your Recruiting Pipeline

I believe the success of any game or company comes down to one thing: the people. The difference between a great team member and a mediocre one isn’t 10% or 20%—it’s 10x or more. Steve Jobs famously said that Steve Wozniak was a 30x engineer. In the age of AI, that gap is only going to widen.

And yet, for years, recruiting at my own company wasn’t good enough. I failed to manage and direct it properly, and our processes suffered. We’ve only recently begun to dramatically improve through a number of initiatives, and it’s already making a huge difference.

Today, I want to share one concept that has been top-of-mind for us: The Candidate Journey and Touchpoints.

This is my adaptation of a framework from Tony Fadell’s phenomenal book, Build. He talks about mapping the entire “Customer Journey”—from the moment a customer first hears about your product to the day it’s no longer in their life.

We’ve applied this by treating our recruiting candidates as our customers. We are now systematically mapping and optimizing every single touchpoint in their journey with us.

• How do they first hear about our company?

• What does the initial outreach email look and feel like?

• What is the experience on our website’s career page?

• Are our written tests and evaluations clear, respectful, and effective?

• What is the tone and quality of every communication they receive?

For a long time, we were doing a poor job at this. We’re not perfect yet, but now we have a framework to focus our efforts.

If you believe that people are the most critical ingredient for success, I urge you to obsess over your candidate touchpoints. In today’s competitive market, attracting and hiring great people is one of the most powerful sources of advantage you can build.

Another tweak to our new MAG (macro-alpha-gamedev) format today.

Before starting this series, I recorded a video podcast that got very little traction and in which someone told me: “That was the most cringe video I’ve ever seen.” 😂

Oddly, since shifting to audio podcast and adopting the M-A-G format, shockingly the Substack podcast listens have wildly over performed from my expectations; especially MAG #2… my current theory is it’s because I covered successful Western game developers and the success of Hollow Knight Silksong which was trending.

Having said all that. I’m staying foolish. Imma go full cringe on you guys again and we are now going back to video podcast.

Why?

I want to publish back onto my hated platform Youtube and I’ll also try to start putting this on X (which I also hate) where I get 0 to 1 likes on anything I publish there. Lol.

One other change, I will now focus on 1 topic per section. So 1 macro, 1 alpha, and 1 gamedev topic.

In today’s episode:

• Macro: Deep dive on Arena Breakout: Infinite - from mobile to PC, from pay to win to friendly PC monetization

• Alpha: Destiny: Rising marks the “end of alpha” for the former Big Western IP + China game dev model. Read why

• Game Dev: Why Secrets and Stupid Ideas Win in competitive markets

→ Listen on Spotify, YouTube, Apple

Macro | All About Arena Breakout: Infinite

Top 3 Most Significant Gaming News From Last Week:

1. People Can Fly Announces Major Layoffs After Publisher Cancels Two Projects (TheHRDigest): The studio behind Outriders laid off staff after an unnamed publisher (speculated to be Square Enix) canceled Project Gemini and Project Bifrost without providing clarity on project continuation. This highlights increasing publisher-developer tensions and communication breakdowns.

2. Anonymous Game Developer on New $100K H-1B Visa Fee (Nikhil Malankar): An anonymous Indian game developer discussed the impact of the new US H-1B visa rules after a recently announced $100K fee caused widespread panic. The developer expressed concern that this change could significantly diminish the US's appeal for foreign talent in the competitive gaming industry.

3. Arena Breakout Infinite is out on Steam, and it's no longer pay-to-win (PCGamesN): Arena Breakout: Infinite, a free-to-play, hyper-realistic extraction shooter, has launched on Steam. The new version of the game removes the controversial "pay-to-win" currency, and adds five maps, a solo mode, new weapons, and a trophy room.

Arena Breakout | Will It Break Out?

Arena Breakout Infinite – Case Study:

• Arena Breakout Infinite, a PC extraction shooter from Tencent’s More Fun (Magic Cube) Studios, debuted on Steam after its popular mobile version. Notably, this PC release shed aggressive “pay-to-win” monetization.

  

• Despite being dismissed as a “Chinese Tarkov clone,” its launch shows PC-market adaptation:

  • Peak week one CCU (Concurrent Users): ~56k

  • Design pivots: Monetization shifted from power/currency buys to cosmetic/quality-of-life, solo mode playable, onboarding streamlined.

• Tencent’s broader involvement: their Interactive Entertainment Group, with massive stakes in Riot, Supercell, Ubisoft, etc., proves the weight behind the launch.

• Key industry lesson:

  • Western PC audiences vehemently reject monetization schemes that work on Chinese mobile games—direct pay-to-win mechanics cannot just be ported. Consider Arena Breakout’s mobile/China revenue ($142M of $188M global) vs. US ($12.5M).

  • Cultural sensitivity and adaptability in game design, monetization, and community engagement is crucial for cross-market success.

• Extraction Shooter Genre Assessment:

  • Early PC results for Arena Breakout and secondary hits (e.g., Hunt: Showdown at 5k-20k CCU) may point to good, not breakthrough, performance. Most “extraction” games underperform versus expectations outside China. Upcoming titles like Bungie's Marathon, Embark’s Arc Raiders deserve close watch for potential genre-defining innovation.

^^ Keep an eye out for this game. Looks hella good!

Alpha | The End of Western IP + Chinese Dev Model

The massive success of the Western IP + Chinese Dev collaboration model marked a new alpha back in 2018 to achieve massive global success on mobile.

Western IP + Chinese Developer Collaboration – Rethinking the Model:

• The IP partnership model (e.g., Activision’s Call of Duty Mobile with Tencent TiMi) proved incredibly successful, combining world-class franchise power with rapid, resourceful, and extremely skilled Chinese development.

• This model was also predicated on the “secret” that at large scale, non-linear user acquisition effects kick in causing greater organic inflows and reduced CPI. Watch the discussion on this topic below:

• Destiny Rising, a big-budget game developed by NetEase with Bungie, signals the possible end of this era—revenue charts and relationship data point to a fade-out in effectiveness. Remember there are massive dev costs and royalties involved with these deals.

• Root Causes for Collaboration Breakdown:

  • Economics:

    • Deals benefited the IP owner (Western side), with Chinese developers shouldering all development costs, heavy IP oversight, and protracted approvals—seen increasingly as unfair/sour deals on the Chinese side.

    • With hit homegrown IPs (e.g., Genshin Impact, Black Myth: Wukong), Chinese studios are reevaluating whether licensing Western IP justifies high risk and expense.

  • Cultural Misalignment:

    • The “996” Chinese work culture (9am–9pm, six days/week) breeds resentment when Western teams—often perceived to work less—hold significant oversight and share few risks. Asymmetrical commitment and frustrated sacrifices erode morale and collaboration.

    • Human perspective: Western teams often overlook the real, personal cost on the Chinese side. The result? Attrition, bitterness, and a collapse of trust that undermines the model.

• The Strategic Takeaway:

  • This may be the end. Studios/investors should expect fewer big-budget US/China IP partnerships—focus will shift towards native IP development and more symmetrical cross-border deals… if any at all.

  • The industry should anticipate acceleration in “homegrown” Chinese hits and potentially tough transitions for IP owners hoping to repeat the success of past collaborations.

Game Dev | Secrets & Stupid Ideas

Embracing ‘Secrets’ as Differentiators:

• Inspired by Peter Thiel’s Zero to One, the “secret” is an important truth about the world that most people don’t see or agree with. Success in hyper-competitive categories nearly always requires you to bet on these secrets—ideas outside conventional wisdom.

• As a startup, trying to find success in a highly competitive space where there are nothing but dead bodies, the actual stupid thing to do, is to do what everyone else has already tried. Conventional thinking in this context won’t work.

• Why ‘Stupid’ Ideas Sometimes Win:

  • Stupid ideas are often contrarian by nature—misunderstood, mocked, or rejected at first blush. But humility is recognizing that your “common sense” can be flawed.

  • Case study: Early skepticism around eBay’s user trust model and Uber’s rideshare concept proved wrong—these ‘ridiculous’ ideas delivered billions in value.

  • Marvel IP pitch analogy: If Spider-Man and Captain America were pitched as original concepts, few would have invested. The seemingly ‘stupid’ idea can actually be a massive opportunity.

  

• Key Principle for Teams and Leaders:

  • Don’t necessarily dismiss the strange, unpopular, or highly risky ideas. True product innovation in today’s games industry demands you “think different,” channel humility, and move forward despite disagreement with your team.

• Actionable Wisdom:

  • Leaders should encourage the exploration of wild ideas, support team members willing to risk looking foolish, and remain open to outcomes no one can predict.

Happy MAG-nificent Monday!

• 3-Macro: 3 stories impacting the competitive environment in which we operate

  • Hike shuts down after India RMG ban

  • Why GenAI is failing

  • Amazon and Netflix announce Gaming+Streaming bundle

• 2-Alpha: 2 potential sources of Finding Alpha

  • Clash Royale hits 3 year all-time highs… how?

  • The Illegible Margin (by Taylor Pearson)

• 1-Game Dev: 1 lesson learned from getting my ass kicked in game dev

  • The Problem with Product Validation

This week one other slight tweak. Instead of doing text summaries of the Alpha and Game Dev podcast content, I just transcribed the text below. Which do you prefer?

Check it out! 👇👇👇

Brought to you by:

Stash: The only full-suite D2C platform built for game studios ready to win. Stash goes beyond payments with loyalty, incentives, and conversion tools that turn transactions into growth. With Stash, you don’t just go D2C—you grow D2C.

3. Macro

Top 3 Most Significant Gaming News From Last Week:

1. Hike Shuts Down After India Bans Real-Money Gaming (TechCrunch): Former unicorn startup Hike is shutting down completely after a new Indian law banned all real-money online games. The regulation fatally undermined the business model for its gaming platform, Rush, making its primary market unviable and prompting the 13-year-old company's founder to cease all operations.

• Naavik: India’s Post-RMG Era

• Konvoy: India’s $23B Ban: A Warning for Investors

2. Why Generative AI Projects Fail (Computerworld): Generative AI projects often fail not because of technological limitations, but due to fundamental issues with strategy and implementation. Key reasons include a failure to identify a clear business problem, leading to solutions without real-world value. Many projects also falter by starting with the technology itself rather than the user's needs, resulting in a poor user experience

• Marcus on AI: Generative AI’s crippling and widespread failure to induce robust models of the world

• Unity Podcast: Confessions of an AI Maximalist with Joe Kim

• JK’s First Post on AIxGaming in 2022: Mr. Beast, Tesla AI Day, and the Future of Game Development: Small Vector Creation is Coming!

3. Amazon and Netflix Strike Major Ad Platform Partnership (The Hollywood Reporter): Amazon and Netflix have partnered to bring Netflix's ad-supported tier inventory to Amazon's DSP (demand side platform), starting Q4 2025 across 11 countries. Amazon now can programmatically buy advertising across all major U.S. streaming services through its DSP, despite owning the competing Prime Video platform.

2. Finding Alpha

Alpha Insight #1: How Did Clash Royale 5X Revenue?

For our first Alpha insight, we are looking at the recent revenue explosion from Clash Royale last month. You might think it's an old, declining game from March 2016, but Supercell has done it again. Just like Brawl Stars last year, Clash Royale is now having its moment.

Let's start with the headline number because it's a banger. According to AppMagic data, Clash Royale pulled in over $68.1 million in revenue in August. This is the kind of revenue the game hasn't seen since the summer of 2017. We're talking about a nine-year-old game hitting numbers from its peak glory days. For example, back in December 2017, they made a little over $80 million. For more recent context, Clash Royale was making just $12 to $15 million per month throughout late 2024 and early 2025. We're looking at a 4-5x increase since the beginning of this year. These aren't just small improvements; this is a complete renaissance.

In fact, Clash Royale out-earned both Brawl Stars and Clash of Clans combined in August. I also want to point out some additional data points I found from SQURS.com.

In terms of new accounts per year, Clash Royale was on a steep decline from 33 million in 2021 to 16 million in 2023 and just under 13 million in 2024. But now, in 2025, the turnaround is fully underway. SQURS is showing roughly 12.6 million new accounts so far, which is about the same as all of 2024, with three and a half months left to go, including Christmas. So 2025 will dramatically outperform 2024.

SQURS is also showing 56 million MAU in August, up 63% since April. Total accounts have increased by 4 to 10 million, but active users have increased by 30 million, which means significant reactivation.

So, what happened? What did they do? Oddly enough, the primary changes seem to be along the lines of product, live ops, and growth. Let's look at each of these areas.

Starting with product, they made a number of changes beginning in April, including the elimination of chest timers. They removed chest keys, killed banner tokens, simplified the season shop, expanded the battle pass from 60 to 90 tiers, and merged daily rewards with battle rewards, shifting to instant gratification. The June 2025 update overhauled progression: Trophy Road stretched to 10,000 trophies, ranked mode got seasonal resets, Masteries got nerfed, and resource tweaks were introduced to help casual players.

Then, in July, they introduced what I think may be one of the biggest factors: a new major mode called Merge Tactics. As you can guess, it's a four-player, turn-based auto-battler mode. Just like Riot added Teamfight Tactics to League, Supercell added Merge Tactics to Clash Royale.

On the LiveOps side, they've gone berserk. The Clash Royale LiveOps calendar is jam-packed. They're rotating everything, bringing back old favorites like 2v2 mode, reskinning old formats, and just throwing a constant stream of one-time events at players. It's hard to get bored. On top of that, they expanded the meta with Card Evolutions. This seems to be a major driver on the LiveOps side. It's a new, deep progression system that creates a whole new power chase for veteran players. This feature forces everyone, even maxed-out whales, back on the progression treadmill.

In the final area, growth, this is a big one. I was speaking to a UA expert at a game dev lunch I hosted, and he confirmed that the word on the street is that Supercell has changed its stance on performance marketing. The theory is that they are now way more aggressive with UA than they historically have been. For a decade, the narrative around Supercell marketing was rumored to be just making a great game and running brand campaigns. What I'm now hearing is that they are pumping a ton of cash into targeted campaigns, especially after the June update, with downloads jumping 20-30% month-over-month in key markets. There are rumors about heavy ad spends on TikTok, Instagram, and YouTube.

On top of all that, they got a lot of free promotion from creators. Influencers like Jynxi and Ken had streams and meme edits that went viral, and even Mr. Beast jumped on at one point. All these changes in product, live ops, and growth are leading to a virtuous cycle of increased engagement, ARPU, and spend on growth, which leads to more attention and viral promotion from influencers.

There you have it. Supercell does it again, this time with a full slate of initiatives that kicked some ass. Take notes, everyone. Supercell, with Brawl Stars last year and Clash Royale this year, is teaching a masterclass on reviving legacy franchise games.

AppMagic Clash Royale LiveOps Data

^^ Free for a limited time

Alpha Insight #2: Understanding the Illegible Margin

Alpha Insight number two is about the concept of the "illegible margin," which was first documented in a book called Seeing Like a State by James C. Scott.

I became aware of it a few years ago when someone sent me an essay by Taylor Pearson summarizing the book. Last Friday, I was talking to my buddy Corey Klein from Fateless Games and getting feedback on a new product validation framework I had designed when he suggested I read that same essay. The first time, I think I just skimmed it, but this time I took the time to read through it, and now I get the point.

Let me start by reading the first few sentences from that Taylor Pearson essay:

Margin, in a strict business sense, is the difference between how much you can sell a product for and how much it costs to produce.

In a more general interpretation, margin refers to the edge of something, the amount which something falls short of or surpasses another item.

You can build margin into your day by leaving a gap between your scheduled meetings and tasks so that you have room to be flexible. You can have margin between how much money you earn and how much money you spend, so that if something unexpected comes up, you can afford it.

But not all margin is easy to see, and finding previously unseen margin is incredibly valuable.

The huge growth in mobile usage over the past decade has shown that those little pockets of margin—between meetings or commuting during the day—add up to a lot. Smartphones, less than a decade old, are surpassing PCs on their way to 5 billion users.

There are a few nuanced interpretations of this concept, but let's start with the basic idea. The book Seeing Like a State is all about how governments and big systems try to make the world legible for control and measurement. But what they end up doing is screwing over the messy, valuable, complex stuff that's hard to measure. The book suggests that states try to simplify everything into grids and metrics to see it better and control it. It's like the Peter Drucker saying, "You can't manage what you can't measure."

The specific example given is 18th-century German foresters who turned wild forests into uniformly designed and planted pine farms, structured in a highly organized way so it was easier to do timber counts. The short-term win was that more logs were sold, and it was easy to calculate the number of trees. The first-order effects were achieved, and it was easier to measure. But then came unintended second- and third-order effects: the forests started dying from pests and soil rot because the state ignored the illegible part of the forest—things like biodiversity, underbrush for animals, or mixed trees for natural balance. A bunch of complicated factors associated with the natural way forests grow didn't happen in these highly structured forests. While they made measurement easier and more predictable, they effectively killed the forests.

A second lesson from the book is about what Scott calls "high modernists"—people highly focused on top-down, authoritarian rule with a blind faith in structured, measured planning. As an example, the book mentions building Brasília in the 1950s as a perfect grid city to modernize Brazil. The end result was a sterile ghost town where nobody wanted to live. Compare that to more organic cities like Brooklyn that evolved in a bottom-up, messy way, with shops under apartments, mixing life with business. That was much more successful.

One last lesson from the book is the concept of value found in the illegible margin. Taylor Pearson talks about the little pockets of margin that mobile phone users have, as I read earlier—small bits of time between meetings or commuting. This all adds up. He suggests it was difficult to see the massive growth in mobile usage because these small bits of time are difficult to measure. They are, in effect, illegible. But it's here, in these illegible margins, where the real value hides.

Let's recap the lessons. First, we should be careful about too much top-down, authoritarian planning. Some processes can have detrimental second- and third-order effects by pushing a specific KPI too much. We may achieve a North Star metric and win a battle but ultimately lose the war. To mitigate this, the Microsoft Outlook team for mobile is famous for trying to achieve specific KPI targets but then ensuring they also compare any big feature change against key health indicators like engagement to ensure there weren't any negative impacts they couldn't account for ahead of time.

The second lesson is about the dynamic between top-down, pre-planned structures and organic iteration. This is a common duality that parallels others, like the two archetypes of creating art: conceptual innovators versus experimental innovators, as described in the book Old Masters and Young Geniuses. I depart a bit from the book in that I would not suggest everything needs to be bottom-up and organic; you should understand the duality on both sides and draw from both.

The final lesson goes back to finding alpha. Where do we get our competitive advantage? According to Seeing Like a State, you should explore the illegible margins and find the great value hidden underneath the surface, hidden by things that aren't necessarily measured. I'll come back to this concept in the future, but for now, I highly recommend you at least read the essay. Think about this concept more deeply. It's a very cool concept with great lessons to learn if you want to be successful and find that alpha in your game development studio.

1. Game Dev: The Problem with Product Validation

Last section, Game Development. For some reason, I'm always running out of time here. I have a hell of a lot of work, and I'm not sleeping much, but I'm doing this for you. Well, it's also good for me to spend time reflecting on macro, competitive advantage, and game development. If you appreciate the content, please give a like, share, or comment because it really helps.

For today's game development lesson, I wanted to talk about an issue we have been struggling with at our studio, one I've been thinking about a lot: product validation. Let me start with what I believe to be a typical philosophy by many game studios, executives, and investors when it comes to product validation—some call it player validation or market validation. What I'm talking about is trying to determine whether the game you are working on will be successful. This generally involves either qualitative or quantitative data analysis to give you an indication of whether your game design, if properly executed, can lead to success.

I actually believe most people are not making good use of their time or resources in product validation. From my own experience, I have seen a number of very time-consuming, expensive product validation initiatives that were highly intellectual but, in my opinion, just yielded fancy numbers, a fancy-looking report, and a lot of theory that didn't impact the product at all. In the worst cases, fancy product validation consulting firms have executed costly projects that actually hurt the product.

In my career, having worked studio-side and led a number one top-grossing game with zero product market research, and having been involved in dozens of market validation studies, I have to say there is value to concept testing. But a lot of the initiatives I've seen around feature validation, personas, or player motivations—at least the personas that were not data-driven—had extremely limited to no value. In fact, speaking to a few extremely smart and experienced game development leaders in the past few weeks, one who is the head of R&D at a famous studio told me something similar. With respect to player motivation analysis, he also has not really seen anything demonstrably valuable over his career. Another famous gaming leader who created one of the biggest games in the world mentioned something that stuck with me: does talking to and getting data back from a bunch of random people actually give you more signal, or does it just give you noise?

With that context, let me tie this back to what we discussed in Alpha, from the book Seeing Like a State. Do you see the connection? Executives, investors, product managers, and studio leads all want certainty. They want an answer to whether a game will be successful as early as possible. In short, they want to measure success. The fundamental question they are trying to answer is, "Will this game be successful?" But if we think about the illegible margin, another question we should ask is, "Is this question even measurable, or are we yet again trying to measure the illegible?"

It's a real concern because the other issue with product validation is that it is extremely complicated. That head of R&D I spoke to did suggest it can be valuable, but it is extremely complex to execute. Yet, when I speak to so many gaming executives or investors, I hear time and time again, "Did you do the product/market/player validation?" The problem is that they see this as a binary checklist item. "Oh, you did it? Great." or "Oh no, you need to do it." But it's not binary. It's not about whether you did it; the bigger question is how you did it.

The prevailing narrative is, "You have to do the player validation as quickly as possible. Just do it." "Okay, we did it. Check." "Great, they did the player validation. They're fine." But the vast majority of these people just don't have a clue what they are talking about. They have a high-level conceptual idea, but the fact is, there are a million ways of doing this, and probably 999,999 bad ways. Honestly, I'm not sure if I've ever done it in a good way or even seen it done well, and I've been in this industry for a long time. I will caveat that with concept testing; there's a lot of good stuff there. And building personas out of data through clustering can be useful. But during the game development phase, a lot of validation is just extremely difficult to do.

To make this more real, let me describe a scenario where this kind of player validation could go wrong. Imagine you're working on a big-budget game with a lot of risk. Executives are worried about the gameplay, so you hire a consultant to do player validation. You want to take the game in a MOBA direction, or maybe an RTS direction. This consulting firm, some of whom have never made a game, will then do a broad-based survey with a lot of players. They'll ask questions about concept art and gameplay models like a MOBA, an RTS, and city building. They ask players what they want.

The thing is, many players, unlike experienced game designers, can't conceptualize high-concept things. So they might just say, "We love MOBAs, we love RTS, we love city building. You should just put it all together!" This actually happened. And it was a wrong approach. Players are not game designers. Depending on how you frame the question and execute these surveys, you can get all manner of feedback. The execution counts. In this particular case, they launched a Frankenstein game that made no sense because "the players said they wanted a MOBA and an RTS and city building." It was a disaster.

The point I'm trying to make is that this is not an exceptional case. I have seen so many bad validation studies. There is so much nuance and complexity that you can easily go wrong. It can lead you to a point where you're not just wasting time but actually getting more noise than signal and going in the wrong direction.

So what do you do? I'm not saying all validation is bad. There are likely good ways of doing it, but it's very difficult to execute properly. Let's characterize the kinds of validation possible by stage. Since I'm mobile-centric, I'll talk in that context: pre-production, production, soft launch, and hard launch.

During the pre-production phase, I have seen some types of validation that are useful, which is concept testing—testing a high-level game idea. For example, testing a game concept where you combine an IP with a new genre. While there are bad ways of doing this, there's some maturity here, and it can give you good signal. The hyper-casual guys are masters of this. They test video creatives of a game they haven't made yet, use a fake app store, and look for high IPM (impressions per mille) and CVR (conversion rate). This can give you a rough idea of whether players would react positively and can lead to good estimates for CPI. It can also be done for relative testing, like comparing a World of Warcraft 4X, a Pokémon 4X, or a Naruto 4X. You create the creatives and test them. This kind of high-level concept testing can work.

Then in production—ideally, in pre-production, you'd also have a sense of the features. This is where we get into trickier territory when we try to use validation to drive feature design. This can take the form of frameworks around player motivation and personas, hypothesizing the kinds of players who would want your game and prioritizing features for them. To be fair, there are likely folks who have done this well. I know Riot has thought about this a lot. But I don't think the majority of the industry has this figured out, and I include myself in the clueless majority. So I am not in that successful minority, if there is one. Sometimes I think this question might be part of that illegible margin.

Another way this is done is through things like customer satisfaction surveys, where you ask players what they think about certain features and score them. But again, this comes down to execution. I have done concept tests where putting a gradient on one character versus another makes a difference. The quality of what you're showing off—all these minute details make this a very complex problem.

I forgot to talk about another important concept: two primary schools of thought around the approach to player validation. I call it the "pyramid player validation model," with top-of-pyramid and bottom-of-pyramid approaches.

The top-of-the-pyramid model is the opinion that "we are the tastemakers, we are the gods of design. We create something that we like, and because we like it, the world will like it." In this scenario, you don't need a lot of external validation. It's more of a Steve Jobs, Elon Musk, or Rick Rubin approach. Believe in yourself; you are the tastemaker. The only issue is, do you have great tastemakers on your team? A prototypical example of this approach, at least historically, is Blizzard.

Then there is the bottom-of-the-pyramid validation, which is about external validation by players. The philosophy is that you shouldn't make a game for yourself because you may not be the target audience. Instead, you should be great at getting feedback from your target players and have them help drive the design. Riot, for example, is famous for being "player-first." When you look at the success of League of Legends, it was through a lot of external player feedback that they were able to take what was a janky Warcraft 3 mod and incrementally improve it to become one of the biggest games in the world.

Which approach is right? I believe you need both, depending on the stage of the game. Start with your own design of what you think would be good, and then as you get more data and players, shift toward a more data-informed approach to iteration.

I'm out of time, so I apologize for not structuring this discussion better. There's a lot more to say, so maybe we can revisit this. But as a quick recap:

1. Too many execs and investors view product validation in too simple, binary a way. It's not a checklist; how you do it is material to the value you get.

2. There's only so much you can validate, and some things you're looking for are in that illegible margin. Be careful of what you try to force to be measured or be predictive about when it may not even be possible.

3. Some level of product validation is possible and valuable, but it's got to be simple. Many of the frameworks I've seen were too complex. It also depends on the stage of development, and as with so many things, the devil is in the details. Execution matters, the details matter, how you do it matters, and who does it matters. Be careful.

I'm going to cut this one short again. Thank you for tuning in. If you like this content, find it valuable, or have disagreements, please comment. Otherwise, please like, share, and subscribe. It all helps. I hope to catch you all next week. Bye.

Hello GameMakers!

Well, looks like we have a winning format. With that said, I am doing one more tweak to the format. Every Monday I will cover:

• Macro: 3 stories impacting the competitive environment in which we operate

• Alpha: 2 potential sources of Finding Alpha

• Game Dev: 1 lesson learned from getting my ass kicked in game dev

Let me know what you think! 👇👇👇

3. Macro

Top 3 Most Significant Gaming News From Last Week:

1. Dreamhaven Games Hit by Layoffs (GamesIndustry.biz): Dreamhaven, the studio founded by former Blizzard Entertainment co-founder Mike Morhaime, confirmed it has undergone an unspecified number of layoffs. Despite the cuts, the CEO expressed optimism about the studio's future and the projects currently in development.

2. Hollow Knight: Silksong Sees Massive Launch Success (IGN): Following its highly anticipated release, Hollow Knight: Silksong immediately surged on Steam, achieving over 453,000 concurrent players within hours of launch. The immense demand, which briefly caused issues on various digital storefronts, placed the game as the third most-played title on Steam and positioned it as potentially one of the biggest indie game launches ever.

3. Pete Hines slams Xbox Game Pass as "worth jack s***" without proper compensation (Kotaku): Former Bethesda VP Pete Hines delivered a scathing critique of subscription services, arguing Game Pass creates an unsustainable ecosystem that doesn't properly value developers, Kotaku +2 reigniting heated industry debates about subscription models. His comments highlight growing tensions between platform holders and content creators as the industry grapples with profitability challenges.

2. Finding Alpha

Alpha Insight #1: The TikTok Music Arbitrage

Magic Tiles 3, a mobile rhythm game from Vietnamese developer Amanotes, just demonstrated a masterclass in platform-native marketing. After years of steady but unremarkable performance, the game suddenly exploded to become the #1 rhythm game by downloads in August 2025, with revenue hitting all-time highs.

Their secret weapon? Ditching expensive licensed music for TikTok's viral "brain rot" songs.

The Strategic Pivot:

• Abandoned traditional music licensing (which can cost millions at scale)

• Embraced TikTok's Commercial Music Library of 1M+ royalty-free tracks

• Featured viral sensations like "Passo Bem Solto" (115-116 BPM Brazilian funk) and "Sigma Boy" (82.7M TikTok posts)

The Results:

• 30% better return on ad spend (ROAS)

• 41% higher user lifetime value

• Zero licensing complications or DMCA risks

• Algorithm boost from TikTok promoting content with trending sounds

Key Takeaways for Developers:

• Platform-native content consistently outperforms generic assets

• Audio is essential on TikTok - 88% of users expect sound, and audio-enabled content is 73% more likely to stop scrollers

• Sometimes the "cheaper" option (royalty-free music) is actually the strategically superior choice

• First-mover advantage is real - Omanotes capitalized while competitors stuck to traditional licensing

Alpha Insight #2: The Great Western Game Development Divergence

While industry headlines scream about studio closures and AAA failures, a quiet revolution is happening. Small Western teams - sometimes just one person - are generating revenue that rivals major studio releases.

The data tells a story that contradicts the doom-and-gloom narrative:

Lethal Company

• Developer: 1 person (Zeekers)

• Budget: Near zero (personal expenses + asset store)

• Results: 10M+ units sold, $70M net revenue after Steam's cut

• Peak concurrent: 240,000 players

• Marketing spend: $0 (pure viral/streamer growth)

Project Zomboid

• Developer: ~10 people (The Indie Stone)

• Development model: 13 years of early access

• Results: 10-20M units sold, hundreds of millions in lifetime revenue

• Philosophy: "The Drip" - steady-state sales funding everything

• User reviews: 94% positive after 13 years

No Man's Sky

• Team size: 15 at launch, ~40-50 today

• Business model: Buy once, infinite free updates

• 2022 profit: £40.5M

• Transformation: From industry's most hated to most beloved

• Strategy: Chose legacy over short-term profit extraction

Phasmophobia

• Started with: 1 developer

• Current team: 5 people

• Results: 25M+ copies, $350-500M revenue

• Peak concurrent: 450,000 players

Schedule I (Lethal Company)

• Solo developer using AI tools for visuals

• 8M+ units sold

• Peak concurrent: 450,000 players

• Marketing: $0 (TikTok viral)

The Pattern Behind the Success

These aren't outliers anymore - they represent a fundamental shift in what works:

1. Authenticity Over Polish

• Players forgive janky graphics for genuine gameplay

• Community interaction matters more than AAA production values

• Systems transparency builds trust

2. Lean Operations Enable Risk-Taking

• 60% profit margins (see Tibia MMO) vs AAA's razor-thin margins

• No corporate overhead means faster iteration

• Direct player feedback loops

3. Platform-Native Growth

• TikTok and streamer virality replaces traditional marketing

• Proximity chat in Lethal Company created more marketing value than any trailer

• Social features designed for content creation

4. Time Horizons Matter

• Project Zomboid: 13 years of development

• No Man's Sky: 8 years of free updates

• Building for a much longer time horizon

Strategic Implications

For AAA Publishers:

• The $300M blockbuster model is broken

• Consider autonomous small teams with $20 products

• Look at Helldivers 2: AA budget, AAA impact

For Indie Developers:

• Tools have never been more accessible

• Audiences actively seek alternatives to AAA

• Serve niches with passion, don't chase trends

• Consider longer development cycles with community involvement

For Investors:

• ROI potential in lean teams dwarfs bloated studios

• Look for authentic voices, not polished pitches

• Platform-native growth strategies are the new normal

The alpha isn't in competing with AAA on their terms. It's in recognizing that the rules have fundamentally changed. While corporate studios optimize for quarterly earnings, a new generation of developers is optimizing for sustainable communities and long-term value creation. The market is rewarding them accordingly.

1. Game Dev Lessons: On Decision Making

Behind every failed game launch and studio closure lies a trail of poor decisions. After analyzing patterns across the industry, a clear picture emerges: the way game studios make decisions is fundamentally broken, and it's costing us billions.

The Macro Problem: Who Gets to Green-Light a Game?

There's a toxic relationship between budget size and decision-maker competence in Western AAA development. As budgets scale from $1M to $300M+, decision-making authority moves up the corporate ladder - and further away from people who actually understand games.

The cruel irony? The people controlling the largest budgets often don’t actually know games or their audience well. They demand data to justify every decision, but the most successful games of the past decade - from Minecraft to Lethal Company - would have failed any data-driven approval process.

Could anyone have predicted the successes outlined by the small indies in the Alpha section through traditional greenlight processes?

The answer is no.

The Micro Problem: The "Blind Men and the Elephant"

At the studio level, critical decisions about features, hiring, and design are frequently made "on the fly" without rigorous contemplation. This is especially dangerous for what Jeff Bezos calls "one-way door decisions"—costly choices that are difficult or impossible to revert.

This flaw is compounded by departmental silos. An artist approaches a problem with one objective, an engineer another, and a product manager a third. Like the parable of the blind men and the elephant, each person accurately describes their piece of the puzzle while completely misinterpreting the whole.

The solution requires a more structured approach: using logic trees, first-principles thinking, and decision analysis to map out all associated issues, knock-on effects, and risks before committing. In an industry filled with brilliant creatives, the failure to implement basic critical thinking frameworks leads to half-assed decisions and, ultimately, failed projects.

The path forward requires more than just smaller budgets; it demands a shift in philosophy, empowering creators and re-instilling analytical rigor into the development process.