Seeing the Chessboard

On the Competitiveness of UGC Games Platforms

Andrew Cedotal

This essay is about UGC games platforms:1 How well they’re competing in the broader landscape of games platforms (meh), and how they can do better (by rewarding developers for driving inter-platform competitiveness).

If you enjoy what you read, you can sign up to receive more like it here:

Please enable JavaScript in your browser to complete this form.

Part I: Introduction

Platforms, Platforms, Platforms, Platforms

Q: What’s a platform?

A: A platform is an entity that lowers the costs of production and/or consumption in a given market.

More importantly–plumb near everything is a platform. You are reading this on a device that is a platform, using an operating system that is a platform, using a browser that is a platform, to access a content hosting site that is a platform, to read a piece written in the English language (which is itself a platform).

You send your text messages, consume your social media, order your rideshare rides, buy health insurance, watch makeup tutorials, shop for groceries, find your vacation rental, book your restaurant reservations, and set up unexpectedly-terrible online dates…on platforms.

As the global economy grows, more segments of more markets adopt platforms. As the wave of personal computing continues, a half-century on, to wash over the human economic landscape, more segments of more markets that economically “should” have been platforms all along flip over into adopting platforms.

UGC Games Platforms

Platforms seem to have finally2 come for multiplayer games.3 Maybe? Roblox and Epic Games are in the market, along with a trail of other platforms (and companies trending in the direction of being true app platforms)4 who currently have less traction.

But the category seems stuck in low-growth mode.5 Hm. Maybe it’s just reached its natural boundaries? Maybe this particular wave of the grand computing tsunami has inundated the entire valley?

Or maybe something can be done to raise the water level and flood some more territory. Maybe current platforms are managing themselves in suboptimal ways that are blocking market adoption.

Spoiler alert: It’s the “suboptimal ways” one.

Why Care (If You Run a Games Platform)

For those who run games platforms for a living:

You are in an arena. There is a sword in the middle of the arena. Are you running to grab it?

You can see an inscription glinting on the blade from where you’re standing. It says: “Supply-side capital intensification.”6 Odd! What could that mean?

Several other people are running for the sword. Are you running for the sword?

You should probably be running for the sword.

You might be happy with your growth rate for now. Your shareholders might be happy with your growth rate for now.

But platform wars tend to be winner-take-all. If you believe you are in a category with great opportunity ahead of you, one worth spending precious years of your life building, then you should assume that other people are going to figure that out as well. If they convert the supply and demand sides of the market more effectively, by implementing a better strategy than yours…well…

Why Care (Entrepreneurs, Executives, Investors)

Now, for those entrepreneurs, executives, and investors who live at the expanding edge of the computing application layer:

Being good at making strategies is like being good at making anything else. One must find instructive examples, examine them to see what makes them work in their specific case, and logically induce rules about what works in general. Then, one can move through the world more safely and mold it more skillfully.

This essay is an examination of strategies. Let’s get some reps in.

Special Thanks

Special thanks go to Jon Bruner, Gareth Falkingham, Emilia Lazer-Walker, Greg Marra, Ava Pierce, and Nicole Villeneuve, who provided review and discussion on drafts of this essay.

Extra-special thanks go to Lars Doucet, whose excellent essay “So You Want to Compete with Roblox”–the best piece of analysis I am aware of on UGC games platforms–served as a direct inspiration.

Now, let’s move on to the current competitive performance of UGC games platforms.

Part II: Current Competitive Performance of UGC Games Platforms

We’ve established that a UGC games platform is an entity that lowers the costs of production and/or consumption in the market for multiplayer video games.7

So how is the category doing?

UGC Games Platforms are Doing…Okay

Epic Games hasn’t released any hard data about the business performance of UEFN, but we can ask: How is Roblox doing?8

20% yearly growth over the past few years isn’t bad. It beats the stock market! But that looks an awful lot like an S-curve. And an S-curve is often the sign of a saturating market segment, with no further market segments in the pipeline.

How did they get here? Let’s start with how these platforms got anywhere in the first place, since every platform has to solve a chicken-and-egg problem.

Roblox found a clever solution: They lowered the floor for participation on both the supply and demand sides. Their supply side is primarily developers who haven’t the engineering skill to create a multiplayer backend, nor the money and management skill to hire someone else to do it.9 Their demand side is primarily adolescents who can’t use other gaming platforms because they’re too young to use consumer OSes to browse and use apps, or because they can’t afford to buy software or hardware to play on other gaming platforms, or because they’re on low-powered devices that won’t run current-generation games, or because their parents will only unlock a walled garden for them. It turns out that when you put those two together, you get (roughly) a thin talent-distribution head of 17-year-olds making games for the entire population of 13-year-olds, and that this pairing is worth a few billion dollars a year in revenue.10

Epic Games solved the chicken-and-egg problem by routing in the supply side from the second-most-popular game engine (Unreal), and the demand side from one of the most popular multiplayer games (Fortnite). This has put them in a substantially similar position as Roblox, on the supply side since platforms tend to pull in low-quality sellers first,11 and on the demand side since Fortnite’s audience is quite young.

Okay so the beachhead is adolescents on both sides of the market. Why the S-curve? Why is it proving hard to move off the beachhead?

When You Lower the Floor, You Get the Floor

If you’re at the bottom of a ladder, it’s easy to move up if no one is above you on the ladder. People laughed, laughed, laughed back in Web 1.0 at eBay. “Beanie Babies LOL” was the approximate reaction, and those reactions were wrong, but why? They were wrong because there were very few12 pre-existing online auction platforms above eBay on the ladder; it was blue ocean all the way up from stuffed collectibles to Maseratis.

If you’re at the bottom of a ladder and someone is above you on the ladder, though, they’re going to kick you in the face if you try to climb their rungs.

Roblox and UEFN have lowered the floor for who can participate in the multiplayer games market. They added a new rung to the bottom of the ladder; bravo! But guess who else is on the ladder above them now, on both the supply and demand sides of the market? Every other game platform in existence.13 And they’ve had quite a long time since Pong to sharpen their cleats.

Roblox and UEFN Have Lower-Quality Content than Competitor Platforms

The S-curve above is an empirical observation about the slowing of demand-side growth. Demand sides of content platforms entering mature markets14 tend to fail to grow when content quality isn’t high enough on the supply side.15 And sure enough, Roblox’s and UEFN’s content (by average content-minute consumed) is not competitive quality-wise with the best content on any major games platform.

This is admittedly a subjective statement. However, I offer the following points in support:

  • It’s the most plausible explanation for the slowing S-curve. It’s not like Roblox or UEFN have come anywhere near saturating the population of humans on the planet, or devices that will run their software.
  • It’s what one would expect if you entered a market from the bottom. You’d expect to be at the bottom.
  • Can you find anyone who will argue that current UGC games platforms’ are content-quality-competitive who doesn’t work at Roblox or Epic Games?16

Why Is the Content Quality So Low?

Because developers on these platforms are not successfully performing capital intensification.

Content gets better through capital intensification, which is a fancy way of saying “the more money a developer spends, assuming they spend it correctly, the better the content will be.” This is a general correlation. There will always be weird outliers in any content market where a few scrappy indies with zero financial capital make a huge hit,17 or where an enormous production budget gets set on fire with little to show for it.18 However, generally, over time, the creators creating content in a given media type will figure out how to deploy more and more capital per minute of user experience until they reach a point of saturation dictated by the nature of their medium,19 because the Darwinian pressures of competing against peers within their medium (and against previously-existing media) dictate that they do so. 

A given medium will evolve as far as it can along the continuum from Asteroids to Cyberpunk: 2077, from three seconds of a guy sneezing to Avengers: Age of Ultron, from an illiterate bard singing by a campfire about Troy to Harry Potter and the Cursed Child.

How Are UGC Games Platforms Doing at Capital Intensification?

Any reasonable person must conclude, upon an honest evaluation of the category’s currently-most-successful developers,20 that said developers are either unable or unwilling to capital-intensify.

  • Brookhaven has had only minor content updates during its 4 years live.
  • Blox Fruits owes its success not to product quality, but to appropriating substantially all of its IP from the immensely popular anime/manga franchise One Piece (a theme we’ll return to in Part IV).
  • The studio behind Adopt Me! has shipped no updates substantially changing gameplay in 3 years, and appears to be making no serious attempts at a follow-up game despite having collapsed to less than 10% of its peak playerbase. This is despite hiring a team of ~80.

Each of these games has earned, after platform fees, at least $50M and possibly as much as $100M, but all of that money is either:

  • Leaving the platform (as in the first two cases). It doesn’t really matter whether it’s sitting in a bank account, being invested in index funds, or being spent on a pile of golden Teslas–it’s leaving.
  • Being reinvested into the platform with low enough management skill such that it’s effectively being wasted (as in the Adopt Me! case)

Whatever the cause, we’re 0/3 for successful capital-intensification among the (probable) three highest-earning developers.

So how can UGC games platforms make capital intensification happen? We’ll answer that question in the next part by taking a globetrotting economic-development tour to South Korea, the Philippines, Venezuela, and Colombia.

Part III: Platform Competitiveness is Economic Competitiveness 

In the last section, we discussed the current competitive performance of UGC games platforms against other games platforms. Summary: They seem to be hitting a growth wall, because their games aren’t competitive with those of other platforms, because their developers aren’t successfully performing capital intensification.

So what can a UGC games platform do to solve the capital-intensification problem? To find the answer, we’re going to look at an entity that managed to go from poverty and non-competitiveness to of wealth and competitiveness: the nation of South Korea.

How Poor Countries Get Rich: South Korea

Let’s say you’re in charge of the South Korean government in 1953. You’re a country with:

  • Lots of globally-poor people
  • No Fortune 500 companies

You would like to, at some point in the future, have a country with:

  • Lots of globally-rich people
  • Many Fortune 500 companies. A Samsung, an LG, and a Hyundai would be nice.

You also have no natural resources. What should you do?

Well, your domestic market is limited in size–you only have so many people, and they’re broke–so you need to figure out how to sell to the rest of the world. Fortunately, Joe Studwell’s excellent book How Asia Works outlines the steps to becoming globally competitive:

  1. Redistribute land ownership so that individual farmers (who make up 90% or more of your population) own their lands instead of renting them from landlords. People work harder when they’re working for themselves!21
  2. Set up government-controlled investment vehicles offering terms good enough to attract the farmers’ newfound money away from flashy, short-term consumer goods (like, say, nice clothes).
  3. Funnel that investment money to sectors you deem highest-long-term-ROI.
  4. “Encourage”22 large firms in those sectors to take loans from these investment vehicles.
  5. Ensure that the loans contain provisions that require the receiving firms to achieve higher export targets (a percentage of revenue from sales outside of the native country) each year. Remember, your domestic market is small and broke; you have to go sell to rich foreigners to get rich yourself!
  6. Ensure that the penalty for failing to hit export targets results in the government gaining some ability to “encourage” a sale of the business.
  7. As some firms inevitably fail to meet export targets–selling in a highly-competitive global market is hard work–step in and force them to merge with (and, importantly, become managed by) firms that are meeting export targets.
  8. Repeat until whole sectors of the economy are controlled by the most export-competitive firms.

And of course, this plan23 24 worked: South Korea is, today, globally wealthy. Capital intensification unlocked global competitiveness. Instead of the median economic actor in these economies being an uneducated farmer with pre-Industrial-Revolution tools trying to sell what pops out of the ground, it’s now a factory manager drawing on hundreds of millions of dollars of physical capital in the form of plant machinery and transportation infrastructure and a universe of contract manufacturers, or an investment-banking VP drawing on hundreds of millions of dollars in human capital in the form of analysts and law firms and proprietary deal-valuation frameworks. And the factory manager and investment-banking VP are obviously going to command a higher price in global markets for what they offer.

How Poor Countries Stay Poor: The Philippines

Now let’s say you’re the Philippines in 1953. You’re a country with:

  • Lots of globally-poor people
  • No Fortune 500 companies

You would like to, at some point in the future, have a country with:

  • Lots of globally-rich people
  • Many Fortune 500 companies

You should do exactly what South Korea is doing, right?

But let’s say you think all of that economic-development-policy stuff is bunk, and that highly-structured government attempts to guide investment are silly. So you don’t take serious measures to make sure that investment capital goes anywhere in particular. This approach has the benefit of being way easier than having to sit down and devise a system of formal-yet-indirect mechanisms to get other humans in the mass to do what you want, and as a bonus you don’t have to pore over company P&Ls every year (boring!) in order to figure out who’s export-competitive.

Wait, why is the Philippines’ GDP per capita today only 9% that of South Korea? And why does South Korea have Samsung, LG, and Hyundai, whereas you’re stuck with your top-5 companies being things like a funeral home chain?

How Resource-Rich Countries Unexpectedly Stay Poor: Venezuela and Colombia

Let’s say you’re Venezuela in 1953. You’re a country with:

  • Lots of globally-poor people
  • No Fortune 500 companies

You would like to, at some point in the future, have a country with:

  • Lots of globally-rich people
  • Many Fortune 500 companies

Unlike South Korea or the Philippines, you are sitting on a literal sea of oil. What should you do?

This sounds suspiciously like easy mode. You should pump the oil out of the ground and sell it, right?

Wait, why is Venezuela’s GDP per capita today only 6% that of South Korea? And why does South Korea have Samsung, LG, and Hyundai, whereas you’re stuck with your top-5 companies being things like a regional ice-cream manufacturer?

Now let’s say you’re Colombia in 1953. You’re a country with:

  • Lots of globally-poor people
  • No Fortune 500 companies

You would like to, at some point in the future, have a country with:

  • Lots of globally-rich people
  • Many Fortune 500 companies

Like Venezuela, you have lots of valuable natural resources. Unlike Venezuela, though, those resources are coca leaves. Their derivative products command a high price–higher even than oil!–but are illegal on the international market. What should you do?

Well, you should grow the coca leaves and sell them, right? 

Wait, why is Colombia’s GDP per capita today only 26% that of South Korea? And why are your top-5 “companies” all cartels that threaten to turn you into a global pariah state?

The problem in both Venezuela’s and Colombia’s cases is the resource curse. Reliance on oil (or coca leaves, or diamonds, or any easily-extractable, high-inherent-value resource) is bad for a national economy. It’s easy to make a moderate amount of money in a global sense by extracting it and selling it through low-value-add processes, and the moderate amount of money earnable from doing this discourages people from doing the really, really hard work of creating globally competitive firms like Samsung, LG, and Hyundai.25 It’s obviously even worse when the resource in question is illegal, and you get sanctioned over it.

Trust Human Nature, Except Don’t

What are the foundational insights into human nature that emerge from our tale of four countries here? Roughly:

  • The human economic spirit is indomitable. If people are given a path to unimaginable prosperity and a place of honor for their country among the community of nations, though it may require many decades of grueling sacrifice, they will grit their teeth and accomplish honest-to-god miracles.
  • Except the human economic spirit is incredibly lazy. If you let people take any path easier than the “grit their teeth” path, they will take it and screw up all your plans, because humans are shortsighted.

Any firm, being run by fallible, fallible humans, will prefer, in descending order, to:

  • A. Do nothing, sitting on a beach and collecting checks
  • B. Extract high-value, low-value-add resources like oil (if available)
  • C. Produce goods and services to serve a low-expectations domestic market
  • D. Produce goods and services to serve a highly-competitive global market

Unless the government deliberately puts firms in a position where they will lose their businesses unless they do a hard thing, they will do an easy thing instead, and the long-term project of economic development will fail.

Are Roblox and UEFN Poor Countries?

Yes.26

Let’s revisit our discussion of the transition from a country full of non-globally-competitive firms to one full of globally-competitive firms: The one about going from an economy of uneducated farmers to an economy of capital-intensified factory managers and investment-banking VPs.

Today, the median economic actor on UGC games platforms is a 20-year-old, working alone, who may or may not know how to set up an LLC. The median economic actor on the other major games platforms is a studio general manager drawing on hundreds of millions of dollars of human capital in the form of producers, and data scientists, and art directors, and user-acquisition agencies, and QA teams, and a finance office, and tools-engineering teams, and recruiting agencies, and audio people…

UGC games platforms may not be looking to go from a nation of farmers to a nation of Samsungs, LGs, and Hyundais. But they are looking to go from a nation of solo coders to a nation of Supercells, Cryteks, and Capcoms!

UGC games platforms aren’t literally countries, though,27 so in the next part we’ll examine how to adapt these policies to solve their specific capital-intensification problem.

Part IV: Rewarding Developers for Inter-Platform Competitiveness

We’ve discussed the policies that allow the economies of whole countries to go from poor to rich, from non-globally-competitive to competitive. We’ve also discussed multiple ways to fail at that transition and stay poor.

Now we’ll talk about how to adapt those policies to UGC games platforms.

What’s Our Desired Outcome, Again?

Let’s review here the preference stack of developers from Part III:

  • A. Do nothing, sitting on a beach and collecting checks
  • B. Extract high-value, low-value-add resources like oil (if available)
  • C. Produce goods and services to serve a low-expectations domestic market
  • D. Produce goods and services to serve a highly-competitive global market

Interestingly, each of the top-three developers in the category that we discussed in Part II is stopping at one of A through C:

  • Brookhaven is (roughly) doing A: Nothing.
  • Blox Fruits is doing B: Extracting a high-value, low-value-add resource named “anime IP.”
  • Adopt Me! is doing C: Capturing domestic demand with substantially no chance of pulling in players from other platforms.

Our goal is to create a set of policies such that developers who want to keep their businesses have no choice but to dig all the way down the stack to D and get globally-competitive.

How Things Work Now

Here’s how UGC games platforms currently allocate revenue to developers.

Roblox does the following:

  1. Players play games they discover through the platform’s search, discovery, and social layers.
  2. Players can buy game-native items (or, if you will, in-app purchases) or platform-native items (items that can be used across the whole platform).
  3. The platform passes developers a cut of items purchased in their games. The cut differs for game-native vs. platform-native items.
  4. The platform also allocates a separate, platform-wide “engagement payout” pool to developers in proportion with player engagement, for some definition of player engagement.

UEFN’s process is currently much simpler:28

  1. Players already on the platform play games they discover through the platform’s search, discovery, and social layers.
  2. The platform allocates a separate, platform-wide “engagement payout” pool to developers in proportion with player engagement, for some definition of player engagement.

If you were paying attention in Part III regarding national economic development, you’re probably starting to see some of the problems here, but let’s go through them one by one.

Fixing the Supply Side

We’ll talk about policies on the supply side (the game developers) first, since there’s a lot more low-hanging fruit. Remember: For the (likely) three highest-earning developers on the platform, the vast majority of capital is either being sent off-platform or misspent.

If You Want Global Competitiveness, Reward Global Competitiveness

Current UGC games platform policies reward developers equally for taking spend away from other games on the platform as for taking spend away from games on other platforms. This rewards domestic (intra-platform) competitiveness the same as global (inter-platform) competitiveness. The former is easier than the latter, so developers will choose it if it’s available.

The problem is that intra-platform competition doesn’t grow the platform, because it increases product quality/global competitiveness verrrry slowwwwly. Yes, if a new game launches onto a UGC games platform, it needs to be better than the available domestic competition to attract players, but it can be ~10% better, instead of the ~300%29 better it would need to be to pull in players from other platforms. You could just wait for that 1.1X multiplier to compound over many intra-platform competitive cycles and maybe you’d eventually you’d get to 3.0X. Maybe. But that would take a long time. 

It also assumes that the other platforms are standing still in terms of platform quality, which they are not. The studio general manager on other platforms, drawing on hundreds of millions of dollars of human capital in the form of producers, data scientists, art directors, etc., is having their machine do everything imaginable to improve their games, from polishing the first-time user experience to calculating whether hiring celebrities is ROI-positive. It’s not a race UGC games platforms are likely to win.

How can we take the easy path away from developers? 

Let’s consider the following approach:

  1. Be abundantly clear, at C-level and down through the entire org, about the specific form of developer performance the platform wants to reward: Global competitiveness; that is, the ability of games to pull in new paying players from off of the platform.
  2. Design and implement a metric that measures the ability of games to pull in new paying players in from off of the platform. 
  3. Pull every policy lever available to come as close as possible to allocating developer revenue according to this metric.
  4. Let Darwin sort ‘em out.

The metric in question, for Game X, looks something like:

Total spend in all games on the platform by players who entered the platform for the first time through Game X30 (and are not part of the current, captive demand-side market who the platform would have gotten for free anyway), for each of those players’ first year31 on the platform.

Now, this metric isn’t a complete solution. It has several problems, but almost anything is better than the current situation of “the domestic capital is mostly just flying out the barn door,” so let’s try to work through those issues.

Making Rewarding Global Competitiveness Feel Fair to Developers

“A player spent money in my game and someone else got a cut of it!” doesn’t feel great, even if it’s the optimal policy for everyone in the long run.

But note that developer revenue for a given game is roughly equal to:

Average Revenue Per User (ARPU) * (1 – Platform tax) * Game Traffic

And that any of these are fair game to tweak in the service of molding the distribution of actual developer revenue to the distribution of “successfully converted paying off-platform players.”

It would indeed feel unfair, perhaps intolerably unfair,32 if developers made money from games they didn’t own. And it would feel unfair, perhaps intolerably unfair, if the platform arbitrarily changed the platform tax for each developer according to opaque rules.

But that leaves us with game traffic as a free variable. Is game traffic highly, algorithmically controllable by a platform?

Heck yes it is! The three legs of the discovery stool are search, recommendation, and social. Media platforms of any type have medium-to-high levels of algorithmic control over all of these. Search ranking is an algorithm; recommendations are an algorithm. “Your friends are doing X”-style social notifications aren’t an algorithm the platform controls directly, but this isn’t a problem, since a platform has algorithmic control over how the three legs are weighted against one another in terms of allocating user attention, and it can turn down the importance of the leg it doesn’t control. The VP33 responsible for traffic gets a quarterly goal of decreasing the average distance between a developer’s global-competitiveness share and revenue share.

The point being that platforms, within the set of inputs to developer revenue, have enough control over the opaque inputs to make the outputs basically whatever they want.

Making Global-Competitiveness Rewards Transparent Enough to Create a Feedback Loop

Okay, but doesn’t using opaque inputs to meet your revenue-allocation goals create a problem where developers can’t figure out what behavior they’re supposed to exhibit? Don’t we need transparency in order to create a fitness feedback loop?

Maybe we don’t need transparency. Can you think of a flagship platform from a technology megacorp that is famous for constantly manipulating obscure traffic-allocation rules behind the scenes?

It’s Google search. For a quarter-century now, Google has been tweaking their SEO calculations inside of a mysterious black box. Despite the box being oh so very black, the Web has evolved strongly around what’s going on in that box, because when large amounts of money are at stake, humans in the mass tend to do a pretty good job of inferring what’s happening inside of complicated black boxes and adapting their behavior.34

Even if a reward-behavior feedback loop doesn’t get going, there is value even in executing a single iteration of the loop. Even if developers don’t change their behavior based on your reallocation, the platform is still in a better position after performing the initial reallocation of revenue to developers who are fitter in a global-competitiveness competition.

Determining Whether a Platform Would Have Gotten a Player “For Free” or Not

A naive system of rewarding developers for pulling in any new players to the platform is going to result in a situation of no more benefit to the platform than the current one. Because the playerbases of the two current major platforms are on the bottom of the product-quality-tolerance spectrum, this would just replace the battle for low-expectation intra-platform traffic (which, again, creates a verrrry slowwww global-competitiveness feedback loop) with a battle for low-expectation about-to-age-into-the-platform traffic (which has the exact same problem).

Nevertheless, the platforms have to try. Fortunately, they have several levers to pull here:

  • Tightening up the process by which users verify their age
  • Using third-party services to statistically determine the likely age of players, regardless of what they say in the verification process35
  • Statistically predicting the age of low-age-confidence players by comparing their on-platform behavior to that of a high-age-confidence population

Handling the Hard Tradeoff of Rewarding Global Competitiveness with Rewarding In-App Conversion

One good thing about Roblox’s current revenue-allocation policy is that it rewards developers for the fitness criterion of being able to convince players to push a button labeled “Buy Digital Thing With Real Money.” This is an important criterion to reward! UEFN is significantly further behind Roblox on this front, because they haven’t figured out that they need to force developers to build the muscle of converting players to in-app purchases.36

But the Good Thing of rewarding the developer muscle of inter-platform competitiveness trades off, dollar for dollar, against the Good Thing of rewarding the developer muscle of in-app purchase conversion. Hm.

There’s likely no perfect solution to this one. The value of a dollar is a dollar!

However, it does point us to the fact that every dollar the platform spends on platform R&D trades off against a dollar rewarding developers for doing hard things the platform needs them to do. And it’s possible that that sort of IRR evaluation is perhaps not, currently, being performed as rigorously as it could be.

The Resource Curse Revisited (or, Anime IP is Oil)

All right; let’s assume that we’ve figured out how to allocate revenue to developers according to their global competitiveness. We now have to deal with the fact that UGC games platforms are trending Venezuela. This author’s estimate as of Q1 2023 was that 40% of platform revenue on Roblox was flowing through games making unauthorized use of anime IP; recent discussions with leading publishers on UGC games platforms has reinforced confidence in this estimate.

Anime IP, like oil, creates a resource curse. If it’s available to developers as an option, the developer base is going to get stuck at:

  • B. Extract high-value, low-value-add resources like oil (if available)

Instead of making it down to:

  • D. Produce goods and services to serve a highly-competitive global market

But of course, using IP you don’t own is illegal. Like coca leaves, it creates a state of latent hostility and potential sanction with the rest of the world. So UGC games platforms are actually trending Colombia, not Venezuela. Anime IP is drugs!37 38

So UGC games platforms should try to figure out how to legitimately bring anime IP onto their platform, if that’s what players want? Right? Well…

Remember that even oil, which is legal, creates the resource curse. If someone manages to land a licensing deal, platforms certainly shouldn’t try to stop it from happening. But they shouldn’t spend precious leadership cycles trying to make it happen.

Keep the Capital In the System

Okay, so we’ve managed to stack platform incentives such that developers, will, in general, be pushed all the way down to D in the preference stack of:

  • A. Do nothing, sitting on a beach and collecting checks
  • B. Extract high-value, low-value-add resources like oil (if available)
  • C. Produce goods and services to serve a low-expectations domestic market
  • D. Produce goods and services to serve a highly-competitive global market

Surprise! There’s actually another option hiding in here which doesn’t apply as strongly to countries, but definitely applies to online platforms:

  • A. Do nothing, sitting on a beach and collecting checks
  • B. Extract high-value, low-value-add resources like oil (if available)
  • B-and-a-half. Take the money and reinvest it anywhere else in the world except on your platform
  • C. Produce goods and services to serve a low-expectations domestic market
  • D. Produce goods and services to serve a highly-competitive global market

Countries deal with this (sort of, sometimes, with some level of effectiveness) through capital controls. If it’s hard (or forbidden) to send your money to other countries, you have no choice but to reinvest it domestically. But once you cash out money from an online platform, you can do anything you want with it. UGC games platforms can’t do capital controls. Can they?

They can’t,39 but they can do the opposite: Development rebates.

  1. The platform sets a list of approved development costs.
  2. Developers send the platform their books each year.
  3. The platform reimburses the developer for the approved development costs up to some multiple of their net-of-platform-fee earnings.

Let’s be frank: This idea has a lot of problems. The platform is going to have a very, very difficult time correlating “a studio hired an engineer” with “a studio hired a non-value-destructive engineer,” or even “a studio said they hired an engineer but actually that ‘engineer’ doesn’t exist and the studio head appropriated the money.” It will lead to all the perverse expense-categorization incentives that pop up with this kind of program. It has to be designed such that developers can’t buy expensive items in their own games to multiply the spend into earnings.

But the important point is that this may be less bad than what is happening now from the platform’s perspective, which is that there is no thumb at all on the scale between re-investing money back into the platform vs. using it to buy Treasury bills, or cryptocurrency, or a pile of golden Teslas.

It would also have the salutary effect of rewarding studios who level up their general-management ability; you have to have a P&L to be able to submit a P&L. And as we’ll see in Part V, even if platforms solve all of their financial re-investment problems flawlessly, there’s still a general-management/intellectual-capital problem lurking on the horizon.

Fixing the Demand Side

Discourage Spending on Flashy Consumer Goods

Remember from our discussion of South Korea that it’s not enough to make sure that the most inter-platform-competitive companies get access to the most financial capital from whatever pool is available. You also have to increase the pool of available capital by incentivizing individuals with money to route it to places where it can be invested in high long-term-ROI uses (like government funds that lend to Hyundai to build factories, or the bank accounts of game developers), and not into flashy, short-term goods like luxury clothes.40

Fortunately, Roblox doesn’t have to worry about its demand side routing all of its money into flashy, short-term goods like luxury clothes, does it?

Ah, crap.

Well, it’s not so bad, right? That item is sold by the platform itself, so at least that money is being invested into platform capabilities rather than going into the pockets of say, a maker of digital hats whose capital-intensification ceiling is too low to increase the global competitiveness of the platform. Right?

Ah…double crap.

To be fair, this isn’t the biggest deal in the world from a platform P&L perspective. I can’t imagine that more than 1% of the total cash flowing through the platform is being paid out to digital hatmakers instead of to game developers. Problems arise, though, when 1% of total cashflow gets misrouted here, and 1% gets misrouted there, and pretty soon you’re talking about real money.

Does this mean that UEFN should reconsider allowing players to bring their Fortnite cosmetic items into UEFN games? Yes, it does.

What Not to Do: Try to Pick Winners Manually

Don’t do it! Picking consumer-media winners is extremely difficult. Hollywood has it in its DNA (at least some parts do, sometimes). Reed Hastings could build it into his organization over a half-decade when dealing with a media type whose user engagement is highly non-trivial to analyze, and yet still much easier to analyze than games. I do not recommend any of the current UGC platforms or platforms-in-waiting attempt this.

Plus, trying to pick winners manually goes against one of the major advantages of being the platform that occupies the extreme low end of a market,41 which is the ability to enable the businesses of absolute-nobody weirdos that editorial gatekeepers living in coastal cultural bubbles would never for one moment take a chance on, but that paying masses actually want.

We’re Not Out of the Woods Yet

Okay, let’s assume that we’ve solved all of the problems of capital availability and allocation on a UGC games platform. We’re still not done in making sure that the right kind of capital intensification happens; that we get a platform full of more-competitive content instead of a platform full of Too Human, Cats (2019), and Supertrain.

Part V: Money Can’t Buy You Competitiveness (At Least Not By Itself)

We’ve now discussed what UGC games platforms need to do in order to:

  • Stop financial capital from escaping their platforms and becoming, say, piles of golden Teslas
  • Route that capital to developers who, measurably, are able to pull new paying players into the platform

Even if platforms manage to achieve both of those goals, that’s not the end of the strategic problem. Money—financial capital—can help solve a lot of problems, but to actually solve them requires one more thing: Intellectual capital.42

As the Adopt Me! example above shows, it’s entirely possible for a studio to re-invest a large amount of money into a platform, but to execute poorly enough that they build the equivalent of a bridge to nowhere.

How can a platform, systematically, discourage this and encourage its opposite?

How to Get Intellectual Capital

Let’s go back to our games-studio general manager on, say, iOS or Google Play or Steam or Playstation or what have you. They know how to draw on hundreds of millions of dollars of human capital in the form of producers, and data scientists, and art directors, and so on. They weren’t born knowing how to do this; the know-how was built up inside of organizations over many decades, and transferred to them mostly under the guidance of other people while actually operating the business. This skill that’s being learned is intellectual capital.

Now let’s turn to our 20-year-old solo UGC games-platform-developer who suddenly wakes up with $50 million in their personal bank account. Where do they start? Who do they call? What’s a cashflow statement? How do you interview a QA director? What’s a paid marketing channel? What does a realistic production plan look like? How many artists does it take to maintain a content pipeline of X cosmetic items a month? What’s D30 new-user retention? How do I negotiate an employment contract? What are taxes? What’s the best meeting and reporting structure to keep all of this together?

Unless they have an extremely strong base of business experience among their close relatives, they’re lucky if they don’t get swindled by the first ethically-flexible set of people they network into. It’s easy to understand why many of them just check out and go sit on a metaphorical beach.

Is there any way to acquire intellectual capital besides through the guidance of other people while actually operating the business? Well, people can acquire it on their own, purely through experimentation. But this is extremely slow. Life is short, and craft long, opportunity fleeting, experimentations perilous, and judgment difficult.

Fortunately, there’s a faster way: Most of the countries that grew their economies at a blistering pace from Meiji-era Japan onward (including South Korea) hired outside advisors to help them move up the experience curve much more quickly. But this model doesn’t apply to the games industry, a rough-and-tumble frontier enterprise where companies are hewn out of bare rock with bare hands by unsophisticated firm founders who directly manage their companies from foundation until eternity, right?

Try again:

  • Atari was started by a former carnival barker and a junior electrical engineer. It reached its greatest success after being taken over by Warner Communications, because it turns out that the people needed to sell video games to a mass market were boring record-distribution executives who knew how to get brightly-colored boxes onto store shelves.
  • Blizzard was started by several friends who were such scrappy outsiders that they met while working at a clipart company.43 The company found its way to success under the ownership, and financial and marketing guidance, of Davidson Associates, who were so incredibly square that they are the people who made Math Blaster!.
  • Naughty Dog was started by two friends, who remained a small outfit making low-budget games until Kaz Hirai called them up during the development of the original Playstation and, along with his team at Sony, mentored them into producing Crash Bandicoot and becoming a major games studio.

So UGC games platforms should be welcoming credentialed outsiders with open arms, right? Or at least not making it difficult for them to do business on the platform?

Substantially everyone I’ve talked to involved in UGC games platforms (at least those with more than a decade of experience elsewhere) has described Roblox as being unexplainably difficult to work with compared to other platforms. Management is unreachable even by companies with significant credibility and financial backing; basic financial reports for games are unnecessarily difficult to obtain; essential developer-ecosystem transactions like the buying and selling of games are interfered with procedurally and legally. Hopefully UEFN will do better on this front, given Epic’s corporate DNA of selling complex, mission-critical software to demanding customers.

Overwatering Your Plants Wastes Water, and Also Kills the Plants

Capital is like water, and companies are like plants. If you plant a plant and give it zero gallons of water in the first month, it’s going to die. But if you plant a plant and give it 10 gallons of water every day in the first month, it’s also going to die.

If you give a one-or-two-person company run by people with no professional experience—but that has made something that people want to pay for—enough money to hire a few friends, you might get magic. If you give them $50 million off the bat with no oversight–effectively the situation of our current top-3 UGC developers–you’re going to get a dead plant. Or, as the proverb goes, “Give people enough and they can do anything, but give them too much and they can do nothing.”

In Part IV, we discussed the extreme flexibility that media platforms have in routing traffic, and thus revenue, preferentially according to whatever business goals they deem worthwhile. One of those goals should be to flatten the head of the revenue distribution. Being able to state higher numbers for the highest-earning developers has theoretical PR value in attracting new developers to the platform; the product-quality observations from Part II show that it does not have any practical value in solving the inter-platform competitiveness problem. Better to give less water to more plants!

Implement Personalized Game Recommendations, So New Users Don’t Bounce Off the Platform

Plenty of media platforms are able to support many completely non-overlapping groups of users on the demand side. The Youtube users watching people in hoodies saying “Yyyheeellloooo gaaaaamers!” are largely not the same people watching crafting tutorials. Platforms are able to do this because they offer personalized media recommendations.

It’s 2024. Recommendation engines are a solved problem. Amazon will just sell you one. And yet today, if you open several different Roblox or UEFN accounts, and play only games from Genre A on the first, only games from Genre B on the second, and then only games from Genre C on the third, all three accounts will be recommended the exact same games to play.44

The problem this creates from the perspective of pulling in intellectual capital is that, at any given time, the next rung of the competitiveness ladder (which is also the capital-intensification ladder, which is also the intellectual-capital ladder) is constantly asking itself, “Should I tip over into developing for this platform?” If the answer is “Well, you could, but even if you’re able to pull a new kind of player–older, higher-spending, whatever–into your games, they will bounce out because the platform is mostly showing them games irrelevant to them,” then that means the answer is “No.”

The UGC games platforms-in-waiting are starting to figure this one out, so the actual platforms had better get on it.

Conclusion: Maybe, Maybe, Asymmetric Advantage is Possible

Let’s say that one of the UGC games platforms currently in existence, or one not currently on the radar, manages to solve all the problems above. This merely gives the category a chance to break into new market segments; nothing is guaranteed.

Disruption theory got a bad rap in the 2010s, when it was thoroughly debased by low-information actors using it as a suitcase word. However, it’s a useful framework for understanding how companies at the bottom of a given market grow upwards, so we’re going to exhume it, blow off the dust, and hope that an evil spirit doesn’t emerge to begin a thousand-year reign of darkness.

Classical disruption requires a new product to exhibit some quality that causes a new or existing market to prefer it, and that is not replicable by incumbents for some reason (e.g. a differing technological base, unwillingness to cannibalize existing businesses, or business-model path dependence). It is possible that UGC games platforms have some asymmetric, disruptive advantage over other games platforms that will allow them to eat into those platforms’ share of the attention economy from below. Possibilities include:

Not every platform gets this chance. Consider Etsy, with other, better-capitalized platforms forever squatting above it on richer ladder rungs and protected by their network effects. Etsy can talk about craft-fair culture all it wants, and be as lax as it wants on sellers making laser-cut coasters with NFL team logos that they definitely don’t have the rights to,47 but neither of those is an asymmetric, disruptive performance axis that will flip a massive market over into their corner.

Maybe, maybe, maybe UGC games platforms have some sort of inherent advantage that will allow them to raise the water level and flood the next floodplain.

But the category has to learn to walk with regards to competitiveness dynamics if it wants to find out if it can run.

Remember, if you enjoyed what you read, you can sign up to receive more like it here:

Please enable JavaScript in your browser to complete this form.
  1. Although most of it applies to all app platforms, regardless of whether they are gaming-exclusive! Hi there, iOS, Google Play, Playstation, Nintendo, Shopify plugin ecosystem, etc.!
  2. In a consequential way. Sure, there were sites hosting a bunch of .io games before, even some with a unified social layer, but those were too small to be chessboard-worthy.
  3. For some reason, the consumer-software sector has collectively decided that multiplayer games platforms should be called “UGC games platforms” even though UGC isn’t their distinguishing feature. On the one hand, terminological inaccuracy is just the pits. On the other hand, why fight the tide? Much time was spent in the Seeing the Chessboard ideation palapa attempting to come up with a better term, but we stopped after Server-Integrated Game-Multiplayer-App platforms (SIGMAs), Platforms for Serverless Multiplayer Apps (PLASMAs), and Multiplayer-App-Store-Having Server Platforms (MASHSPLATs), and decided to just go with the crowd on this one. We’ll call them “UGC games platforms.”
  4. Including Rec Room, VRChat, Manticore/Core, and a cloud of smaller companies.
  5. Evidence coming in Part II.
  6. Sounds complicated and academic, right? We’ll get to what it means in Part III.
  7. Again, UGC games platforms don’t actually strictly have to allow UGC, but we’re going with the industry-consensus category name. Terminological drift is a fact of life; “RPG” remains a useful term despite its drift from the original meaning of “a game in which one plays a social or literary role.”
  8. This graph is DAU. While we’re taking as a given here that the goal is to maximize enterprise value, which more nearly means maximizing revenue than DAU, it’s arguable whether DAU or revenue is more indicative of a saturating consumer market. So we’re going to run with DAU for this analysis.
  9. This worked not just because Roblox and UEFN have adopted platform-as-a-service technical models, but because their development tools abstract away even more server-side concerns than is strictly necessary.
  10. This was only not obvious in foresight, it’s still not obvious to a lot of people in hindsight. Plenty of analysts’ attempts to explain Roblox and UEFN as businesses still, today, go something like “Ahh…Minecraft? Minecraft Minecraft STEM? Minecraft metaverse?” rather than the much clearer “Wow, turns out late-adolescent basement coders making games for adolescents is a significant market.” Roblox doesn’t get enough credit for identifying this market, which one must assume they pathed into consciously given highly non-trivial efforts such as becoming the only large COPPA-compliant platform on the Internet and leaning heavily into gift cards as a spending channel.
  11. On which see Akerlof’s “The Market for Lemons: Quality Uncertainty and the Market Mechanism.”
  12. I’m not willing to say “none.” God knows what was happening on Minitel.
  13. And, given the crowded terrain of the attention economy, most streaming platforms and social-media platforms as well. A cross-comparison of different platforms by media type is beyond the scope of this essay, although we may return to it at a later time.
  14. If a market isn’t mature, then the bottleneck might be on the demand side, where people who have never even considered using the platform category need to be activated through direct and brand marketing. But the games market is now mature in that sense. Iwata be praised, most people on the planet who will be using games platforms are already using them.
  15. Or when category variety isn’t high enough, but “You can find a low-quality clone of any multiplayer game genre on Roblox” is close enough to true, and likely to be close enough to true on UEFN in a few years, that we’re going to assume it’s not the problem axis here.
  16. The reader is invited to play the first 5 games they see on Roblox or UEFN “Most Played” and form their own opinion.
  17. Vampire Survivors! The Blair Witch Project! It’s Always Sunny in Philadelphia!
  18. Too Human! Cats (2019)! Supertrain!
  19. It’s possible to spend $300M making a Star War; it’s much harder to spend $300M making a novel.
  20. Which happen to be the top 3 Roblox developers, as Roblox is currently leading the market. As with many aspects of business-performance evaluation, we’ll have to wait a few years until the trendline becomes apparent for UEFN.
  21. This is not leading to a point that Roblox or UEFN should reduce the platform tax it collects on player spend before passing the money on to developers. If you have a captive developer base—which, as we saw in Part II, these platforms do—and need to invest a lot of capital in platform-level capabilities to get your flywheel going, then you’re completely justified in taxing them extremely heavily. After all, you’re investing to grow the platform as a whole, which means your developers will make more money in the future. It is leading to a point, however, made in Part IV, that Roblox and UEFN should modify or abandon a fairly significant revenue stream!
  22. It helps if you’re a military dictator who can lock CEOs in hotel rooms until they sign the paperwork you give them. Fortunately, platforms don’t have that level of control over developers. Right?
  23. As Studwell points out, this approach goes back at least to Meiji-era Japan, and was also implemented with variations by Singapore, Taiwan, and Hong Kong.
  24. Which we are not going to apply step-by-step to UGC games platforms once we get to making the comparison, although we are going to adopt large parts of it.
  25. There are also other bad dynamics, like the fact that it’s structurally easier for resource-extraction firms to get into a cycle of corruption with government overseers than manufacturing or services firms, but we’ll leave those aside for now.
  26. This actually understates the case; app platforms start even further behind our example of South Korea in 1953, since as pointed out in Part II, they always have to solve the chicken-and-egg/immigration problem on the demand side.
  27. Although several of them have GDPs larger than most countries, so maybe the question of their economic-development policies should be receiving more attention?
  28. Although they’re changing policies very rapidly, so who knows what it will be next quarter?
  29. Again as with the product quality evaluation of UGC games platforms in general, these are subjective pseudo-numbers, but sufficient for the current discussion.
  30. Yes, this is basically just referral programs/affiliate marketing!
  31. Thus solving the “What if developers decide to go sit on the beach and collect checks?” problem.
  32. Roblox and UEFN can do basically whatever they want to their current, captive market on both the supply and demand side, but as they start trying to walk up the supply-side competitiveness ladder, they do need to start worrying about higher-competitiveness tiers of developers choosing other platforms.
  33. Or SVP, or C-level exec. Once a UGC games platform solves the chicken-and-egg problem and isn’t default-dead, inter-platform competitiveness is the most important thing besides keeping the servers up.
  34. You do end up with a lot of meaningless pigeon ritual dancing on top of the actual adaptive behavior, but at least you get some adaptive behavior.
  35. I am not a lawyer; however, my definitely-ask-a-lawyer-about-this-before-doing-anything recollection of the FCC’s interpretation of COPPA is that a platform can determine information about players covered by COPPA, as long as it is neither PII nor being used to correlate them with a unique user ID on another platform.
  36. But these sorts of things can be fixed.
  37. The obligatory comment here being “And not just in an economic sense; have you seen JoJo’s Bizarre Adventure/Evangelion/Akira/etc.?”
  38. And it gets even worse. On UGC games platforms, tolerance of IP violations blocks (the equivalent of) foreign investment. Let’s imagine the following dialogue: ENTREPRENEUR: Hello, Professional Capital Allocator. I would like $10 million to start a studio to target UGC games platforms. PROFESSIONAL CAPITAL ALLOCATOR: Maybe. How are you going to compete with other games on the platform? E: We are going to outperform current studios at product and marketing. PCA: Wait, doesn’t most of the revenue on UGC games platforms currently go to unlicensed anime games? And aren’t you and I, as publicly-known professionals with LinkedIn profiles and registered corporations and bank accounts, unable to compete with them because we can’t infringe on IP as well as an anonymous developer who can disappear into the depths of Discord at a moment’s notice? E: … PCA: The answer is no. (Beat) Have you considered Apple Vision Pro?
  39. At least not unless they want to nail themselves to the bottom rung of the supply-side ladder.
  40. Which is incidentally why it’s acceptable to laugh at people who say “It’s an investment!” about things like fur coats and bottles of wine and speedboats. It’s because they’re not investments.
  41. And that continues to apply even if you capture rungs above the lowest one.
  42. Note that intellectual capital is not the same as intelligence. No one is saying that a given group of developers with less intellectual capital is any less intelligent! But there’s a reason for the saying “Age and guile beat youth and enthusiasm.”
  43. See Stay Awhile and Listen.
  44. A broad statement, but again, one confirmed in conversations with major developers and publishers on the platform.
  45. Let’s hope Taiwan doesn’t get invaded, but what if it does and phones suddenly drop in performance/price ratio for a decade? Stranger things have happened.
  46. Don’t laugh. It happened with ARM and RISC and power consumption.
  47. A pattern of IP violation! Sound familiar?

Discover more from Seeing the Chessboard

Subscribe now to keep reading and get access to the full archive.

Continue reading