1. Introduction

Every day, it seems major game companies and publishers talk about how much they want to make Live Services. Or if they don’t have one yet, how much they wish they had one.

95% of studios are working on or aim to release a live service game

GamesIndustry.biz: Report: 95% of studios are working on or aim to release a live service game (Jeffrey Rousseau, 2024)

And when it comes to games, companies just seem to follow the latest trend and copy whatever was recently successful. Even if, quite frankly, they really should not.

I was in the board room where they presented HYENAS game concept internally in 2017.
"PUBG, Overwatch, Destiny & Tarkov. Will take 5 years to make."

It was soul-crushing to see that's how games were made at my favourite studio.

Such a waste, hope the impact is minimal. https://t.co/gnn7qvHDHj

— Darren (@whatdarrensays) September 28, 2023

And they’re willing to spend big money chasing trends and making Live Services:

According to developers who worked on the ill-fated project, Hyenas was the single most expensive video game Sega’s ever made.

Kotaku: Report: Sega Just Canceled Its Most Expensive Game Ever (Zack Zwiezen, 2023)

On May 9, during an earnings call, Warner Bros. revealed that it was taking a $200 million loss on Suicide Squad — making it one of the gaming world’s worst blunders.

Behind ‘Suicide Squad,’ the Year’s Biggest Video-Game Flop (Jason Schreier, 2024)

And players don’t seem to be happy with this either.

Live service games were a mistake
byu/stanthetulip ingaming

And the failures keep rolling on:

This is a brutal sequence of events. Sony bought the makers of Concord, Firewalk Studios, in 2023. Concord had been in development for eight years, and it was an expensive game, with bespoke cinematics and a long-term plan that would have cost $100m or more to develop. In its two weeks on the market, it sold fewer than 25,000 copies, according to estimates. This is a shocker, even compared with the year’s other bad news for developers and studios.

The Guardian: Sony’s big-budget hero shooter Concord failed spectacularly – here’s where it went wrong (Keza MacDonald, 2024)

Everything seems to be broken. How did we even get here? And what’s driving this?

“Greed” is not a sufficient answer. If you want to solve a problem you need to properly understand the dynamics and factors that influence how people in charge actually make decisions. Otherwise you will fail to ever change anything and your complaints will mean nothing.

So let’s do that.

• To understand how we got here, first we need to understand how companies actually evaluate projects and decide what to do;
• Then we need to understand why companies all like to follow trends; and
• Finally, we need to understand the specifics about the Live Services model which make it so attractive.

2. How do companies decide what projects to do?

Companies are organizations that undertake risky projects in order to make money. Companies therefore need a framework to decide what projects to undertake. The process generally looks something like this:

• What is the risk / reward payoff?
• What is the risk that your understanding of the payoff itself is wrong?
• What is the risk you can even execute?

Or, to put it crudely:

• Is this worth doing?
• Did I make sure not to f@#! up step 1?
• Should I even do this?

2a. Step 1: Is this worth doing?

Projects come in many forms and varieties with different features. This makes direct comparison across options difficult. For example, should you spend on a new marketing campaign or should you spend on expanding customer service?

To solve this problem, we can try to create some framework to facilitate direct comparison. The most common approach that nearly all companies use is to use money as the single unifying factor. This is because money is:

• Quantifiable which allows you to easily compare by seeing which number is bigger;
• Easy to understand compared to many other complex measures; and
• Allows you to measure the impact of a decision because money Is used as both an input to fund a project as well as an output from the project.

This means that almost all business decisions are converted into some form of financial calculation in order to answer the question: Is this worth doing?

Let’s do a very short Corporate Finance 101 lesson. In general, the calculations to prove if something is worth doing will be some variation of:¹Which of these methods is chosen will depend on the internal sophistication within the company, its priorities, and its culture. The NPV or IRR approach is generally recommended in literature but their use within large companies is… varied, to say the least.

• Direct calculation: Cashflow / P&L projection; or
• Adjusted calculation: Discounted Cashflow (DCF) / Net Present Value (NPV) / Internal Rate of Return (IRR)²There may be additional variations or approaches that are considered depending on the industry characteristics. For example, capital intensive industries such as Oil & Gas may consider metrics such as Return on Assets or Return on Invested Capital.

In general, adjusted calculations are preferred. This is because money is not free. There is a cost to money known as the cost of capital.

Every decision has an opportunity cost.³Opportunity cost is the cost of choosing an option. By choosing something, it means you are no longer able to pick another option. For example, instead of spending $10M in a company’s new project, you could just invest that money in a financial asset.

For the new project to be worth doing, it needs to be better than the alternative of just investing that money.

How you determine the opportunity cost and then reflect it in your financial calculation depends on the risk of your project.

Very roughly speaking, different activities have different levels of risk. Riskier activities should generate a higher return in order to compensate for the risk.⁴Some people misunderstand this as implying that doing something risky means your rewards might be higher. This is WRONG! You can take on a lot of risk for absolutely nothing in return. And there are lots of low risk activities that are more rewarding than high risk activities. The primary focus of risk assessment is to take the right types of risk that are actually worth it. And not the dumb risks that are just stupid. This needs to be reflected in the financial calculation for project evaluation.⁵You might naturally wonder if different industries should assess the cost of capital differently given that different industries have different levels of risk. And they do! Furthermore, the method for sourcing money also influences the minimum return required. I don’t intend to go in-depth into this here because it’s not important to my essay’s goals. But if you are interested, I recommend reading about the Weighted Average Cost of Capital (WACC).

Furthermore, there is also what is known as the time value of money. It is better to have $100 now rather than receive $100 in the future. That’s because you can actually do something with the money now instead of waiting. Therefore, the time value of money says you prefer to have money now than in the future. So $100 in the future is worth less to you than $100 now.

Calculations such as IRR or NPV explicitly recognize these issues. Money is adjusted downwards based at some value known as the discount rate.

	NPV calculation	IRR calculation
Discount Rate	You specify what the discount rate is. So you can increase or decrease this discount rate to reflect how risky the project is.	You are reverse engineering the implied rate.
A good project…	Will generate positive cash flow after you adjust for the time value of money and risk.	Will have a rate of return higher than your cost of capital.

So if your project fails the NPV or IRR assessment, then it is probably not worth doing.⁶There are a LOT of caveats that will apply here of course, which is why I say “probably”. For example, let’s say that your industry is currently going through an advertising war. Everyone is burning money on advertising even though it won’t increase sales. So you expect the calculation to say this is a bad idea because you’re spending money for zero gain. But it may be the best decision to also set your money on fire as well because the alternative is not advertising and losing market share and sales. Your financial calculation might be negative, but every other alternative action you take might be even more negative.

2b. Step 2: Did I make sure not to f@#! up step 1?

The problem with financial forecasting is that it is a forecast.

One of the biggest mistakes you can make is to assume your model predicts reality. It does not. A financial model is a way to conveniently make approximations of reality and then ask: “What would happen if my assumptions were correct?”

You can probably immediately see the limitations of this approach.

All models are wrong, but some are useful.

Generally attributed to George E. P. Box

Even if you have the best possible data, you may still be in a situation where you are not certain of the outcome such as:

• What if you don’t actually know how risky a project is because it’s the first time anyone has ever tried something like this and there is no data available?
• What if the cost of acquiring the data you need is too prohibitive?
• What if you don’t know whether assumptions you use today will still be true in the future?⁷This is especially problematic if your project is very long! For example, imagine a 20 yr infrastructure investment. How different will the world even be in 20 yrs? Just 20 yrs ago at time of writing, we still had not gone through the 2008 Global Financial crisis, technological change such as the first iPhone and massive shift to mobile devices, Generative AI, or COVID. And that’s before any country / region specific changes such as the Arab Spring, ISIS, Brexit, Ukraine War, election results, etc.

This makes reassessing what you did in Step 1 critical. For example, you might conduct sensitivity analysis to determine how exposed you are to changes in different variables and the impact changes in assumptions could have.

This is especially important because each successive layer of uncertainty increases the complexity of your assessment. And companies may not have the appetite to take on certain levels of risk.

You should always know what you are getting yourself into. Don’t take on risks you aren’t aware of without some compensation in exchange for that additional risk.

2c. Step 3: Can I even do this?

The final step is an overall evaluation of the options available and the willingness to proceed. As I wrote in my previous essay on corporate decision making:

An important lesson in decision making is avoiding one-way doors where possible. This means “avoid making decisions which are super difficult to reverse.”

[…]

A type of two-way door you can use is called “just make up your mind later”.

Even if a project is highly valuable, an organization always has the option to just delay. Gaining additional information or knowledge / expertise can be highly valuable especially when the risks are too significant.

For example, let’s say you work at a small oil & gas company worth $300M and can choose between two projects:⁸Let’s pretend for a moment that options such as risk sharing approaches through JVs / consortiums / etc. aren’t available. I’m just trying to prove a point here about the “Wait and see” approach.

Project A: Develop a new oil field in country that has just finished a war

• Costs $500M and cost of capital requires 10% return
• 90% of the time has a 900% IRR, and 10% of the time it has -10% IRR instead
• You think there’s a 30% chance war might break out again and all your calculations only work for that 70% chance peace continues
• If war breaks out again, you probably lose all your money

Project B: Expand production in an existing safe oil field in the US

• Costs $50M
• 60% of the time has a 15% IRR, and 40% of the time has a 10% IRR
• You have 100% certainty in this calculation.

Project A is clearly a project that could catapult you into the big leagues and make you a major player. But it’s also very clearly a project that could absolutely destroy your company if things don’t work out.

Project B is practically zero risk and will on average be profitable. But it won’t massively improve your company.

Ideally you’d want to do Project A. But the existential risk it poses to your company might be too much for you to accept.

Well… you can just wait and see instead! Oil fields can take 5+ years before becoming online so you don’t have to make a decision today. Instead you can wait and see if new information improves your ability to make a decision. For example:

• Does the country’s political situation seem like it is improving or getting worse over time?
• Can you infer additional private information by observing how other companies are reacting? (e.g. Are they eager to be the first to open a new field or are they also playing a waiting game?)
• Can your company use the time to gain additional resources / expertise that will help reduce the potential risks involved? (e.g. hiring private security to protect your oil fields so even if war breaks out you can ignore it or reduce the impact to you)

While waiting might lower the potential return / reward (e.g. loss of exclusivity), reducing the existential risk to your company may be a very worthwhile tradeoff.

Whew. Okay. That was a lot to quickly cover as a crash course to basic corporate decision making and project evaluation.

With this basic understanding covered, we can now begin to unpack why the AAA games industry is as broken as it is.

3. Why herd into trends?

Doing something bold, innovative, or different as a company is hard. It’s hard for many reasons, but one of the most critical is knowing if there’s even a market for what you want to do.

This problem, and how companies deal with it, is what encourages trend following.

3a. Trend following is a risk reduction strategy

This is part of why games companies tend to follow trends and copy successes. Success acts as proof that something is possible and is worth doing.

Remember that project risk evaluation requires performing a financial calculation and understanding the potential outcomes. In Step 2 you question your assumptions and try to validate them. In Step 3, you decide if you want to proceed based on your risk appetite.

The best way to solve many of these risks and uncertainties is to just have real proof. A real example is better than anything else. Remember the example of the oil field in a country that just finished a war? The ultimate proof that it is safe to operate an oil field would just be another oil field that is already operating in that area safely for a long period of time.

The same idea applies in other industries. Imagine you’re an investor in AAA-tier games and it’s 2001. A studio comes to you and they say there’s a new genre called a Massively Multiplayer Online Role-Playing Game (MMORPG). It’s like an RPG game but it takes advantage of the internet to offer more to players and it has the potential to be a big hit.

Do you believe them? Are you willing to take this risk? At the time, the main MMORPG was Everquest. A decently popular game, but its peak player base would not come until 3 years later around 2004 when it hit 450k players. So it was niche, but not mainstream. This type of niche genre might be worth niche funding. But a niche game is not worth AAA mass market funding.

There’s too much risk and uncertainty in your risk assessment. You can’t afford to commit to this idea.

On the other hand, there’s this new game called Halo that was just released this year. And it used an innovative dual analog stick control system while also offering online multiplayer.⁹In fairness, Halo was not the first game to use a dual analog stick control system. GoldenEye had this as a feature and Alien Resurrection would use it as the default scheme, leading to this famous review. However, it was Halo that truly standardized and popularized this system leading it to be the default approach that is still used today.

There’s even advanced AI where the enemy will respond dynamically to player actions. Halo is an absolute industry revolutionizing game that defines the XBox console and players want more. And First-Person Shooters (FPS) aren’t a new idea.

Halo has proven there is a viable and massive market for this so you have immense confidence that investing in this has a higher likelihood of working out. And there’s this new studio that says they want to make a World War II inspired shooter. They’ve even previously made a World War II inspired shooter that sold well, proving that people want to play World War II themed games. They say their new IP is called… Call of Duty?¹⁰I am deliberately mangling the timeline here for a more dramatic narrative. Medal of Honor: Allied Assault would not be released until January 2002, and Infinity Ward would not be founded until May 2002.

World of Warcraft (WoW) would be announced later in 2001 and then released in 2004. It was this point at which MMOs finally became a mainstream gaming genre rather than a nice genre. WoW demonstrating that MMOs would be a commercial success is what then lowered the perceived risk in launching an MMO.

This is how we then get the MMO boom with a big rush of “WoW-killers” all trying to copy its success in the late 2000s and 2010s.¹¹It took WoW less than a year after its release in November 2004 to achieve millions of subscribers, far more than anything Everquest would achieve. This success would provide the “proof” for funding future MMOs. For example, Age of Conan, Rift, and Aion started development in 2006 when WoW breached the 6M player mark. TERA would start in 2007, after WoW surpassed 8M players. Wildstar technically started in 2005 not long after WoW’s launch but it was acquired by NCSoft in 2007 which would have been the biggest source of funding. Part of this is people chasing a trend. But part of this is people who had the same idea now getting permission to try and do it themselves.

3b. Trend following is also a strategy for innovation

Businesses love to talk about innovation and trying to be innovative. It should come as no surprise then that there are formal frameworks to think about the process for innovation that have been deeply studied.

The primary framework that is most commonly known is called First Mover vs Fast Follower.

	First Mover	Fast Follower
What is it?	Be the first to come up with a new idea	Copy someone else’s new idea very quickly
Why do you do it?	You get to be the very first company to launch the new product or service.	Focus on proven ideas that have been successful only. This reduces waste and risk to the business.
Benefits	You have a head start in developing the experience and knowledge to understand what makes a good product or service. Being first means you have a head start in building customer loyalty.	Being second means you get to learn from someone else’s mistakes. Ideally, you get to make the better product. Especially important if things like technical debt are a big issue.
Good approach when…	There are high barriers to entry. You have unique knowledge or capabilities. e.g. A steep learning curve to learn what “good” looks like, patents or other legal protections, market is only big enough for one company, etc.	Your company has high operational efficiency and focus to quickly focus on a new product / service. This allows you to follow as fast as possible. Customers prefer the best, and not just what came first.

Copying other companies and following trends is therefore a time tested and very popular strategy in many industries.¹²Consider the smartphone boom after the iPhone, explosion of energy drinks after Red Bull, everyone wanting to be the “Uber for [X]”, etc.

A clear example of this in modern times is the Battle Royale genre. A modder called PlayerUnknown (Brendan Greene) initially made a mod in Arma 2 in 2012 (and later Arma 3 in 2014) which emphasized survival and last man standing mechanics.¹³These ideas themselves were not new. Elimination and Deathmatch type modes have been around ever since the 1980s and multiplayer shooters existed. What makes the Battle Royale genre unique is the combination of these ideas with a shrinking map to force player interaction. In 2016 he joins a new South Korean company Bluehole Studio. They launch Player Unknown’s Battlegrounds (PUBG) in early access March 2017, and then a full release in December 2017.

This game takes the world by storm. Just a few months after its early access release on Steam, it’s already the most played game ever and breaking new concurrent player count records month after month.

.@PUBATTLEGROUNDS has finally reached 3 million concurrent players! https://t.co/HIkZMh7mCB pic.twitter.com/eJwXXSfHJD

— SteamDB (@SteamDB) December 29, 2017

In fact, PUBG is so popular, it single handedly drives Steam adoption in China. By the end of 2017, half of Steam users use Chinese.

@Steam_Spy #Steam #PUBG #Chinese pic.twitter.com/4AdEq1Jzru

— Andrew (@Befzz) November 17, 2017

Now, enter Fortnite.

Fornite is a survival game that has been in development at Epic Games since 2011.¹⁴Remember when survival games was the big trend? How quickly time flies. Epic Games has been going through a tumultuous time, so development has been stalled for quite a while.¹⁵In particular, Epic Games has struggled with the Live Services mode. This is what prompts them to sell 40% of their company to Tencent in exchange for support setting up Live Services as well as accessing the lucrative Chinese market. Because development has taken so long, this game isn’t expected to come out until 2018.¹⁶I’m going to ignore Epic’s failed MOBA game Paragon for the purpose of this essay. But it’s another example of trying to follow trends in the gaming industry. That being said, contrast how fast Epic Games took to launch their MOBA product versus their Battle Royale product. Time to market is critical if you’re going to fast follow.

So while Fortnite is still being developed, PUBG releases. And Epic Games looks at this smash hit and realizes there’s potential. Not just monetary potential of course. But the Battle Royale style gameplay is… just plain ol’ fun as well!

So they decided to fast follow. Within 4 months of PUBG’s release in July 2017, the Fortnite team was already working on their own Battle Royale gameplay. Epic Games was no stranger to shooters, having produced high quality titles such as Unreal and Gears of War before. And to try and follow as fast as possible, Epic Games brings over the entire Unreal Tournament team to work on this project.

Fortnite Battle Royale releases in September 2017. It took Epic just 2 months to fast follow.

And the rest, as they say, is history.

Fortnite proceeds to crush records of its own. In less than a year, it had 125M players. And the numbers just seem to keep growing.

Fortnite now has over 350 million registered players! In April, players spent over 3.2 billion hours in game. 🙌🥳

Let’s keep the party going with our Party Royale Premiere LIVE on May 8 at 9PM ET featuring @DillonFrancis @steveaoki @deadmau5: https://t.co/H18c3UgBL1 pic.twitter.com/Cgt3r7LXQO

— Fortnite (@FortniteGame) May 6, 2020

At the same time, Respawn Entertainment¹⁷The former Call of Duty folks. Remember them from my story earlier in Section 3a? is just wrapping up their latest game Titanfall 2. And they’re wondering what they want to work on next. And they watch this early access game called PUBG come out. And succeed.

So the studio kicks into overdrive. Unlike Fortnite, they have no pre-existing assets to work with, so these all need to be developed from scratch. It takes a while. But 1.5 years later, Apex Legends is released in February 2019 with its own take on the genre.

It hits 50M players within a month.

This demonstrates the high upside potential of being a fast follower. Being able to iterate on a proven idea and expand upon it can lead to success that can match or exceed the original innovator.

4. Traditional AAA games model is deeply broken

At this point we can understand how companies make decisions. And we understand why companies naturally will follow trends.

So let’s talk about why the economics of AAA games is broken and how that influences decision makers and investors at gaming companies and the broader industry.

4a. The economics of AAA games doesn’t work

Since we know how the basics of a finance calculation works, let’s actually do one.

A major AAA game might cost around $100-200M to develop these days due to ballooning costs and player demands for high quality graphics. Checking a quick sample to validate:¹⁸As an additional data point, the CMA report during the Blizzard / Activision merger review also indicated that AAA games often have $200M+ development budgets.

• Cyberpunk 2077: $174M
• Horizon Forbidden West: $212M
• The Last Of Us Part 2: $220M
• Battlefield 4: $100M+
• Callisto Protocol: $150M+
• God of War: Ragnarok: $200M+

In general, the rule of thumb is that marketing costs are an additional 1-2x of the development budget.

So let’s say you’re at a large publisher such as MacroHard Game Studios. You have a studio asking for $200M for the development budget for its new game. AAA games take longer to develop so instead of 1-2 years let’s say it will take 4 years to develop. Let’s also assume:

• the relevant cost of capital is 10%;
• the development cost is evenly distributed across the 4 years;
• the marketing budget is incurred all in year 4 and is only 1x the development budget; and
• you want to break even in the first year of sales.

This means that the NPV of the cash outlay is:

The sales required in Year 4 to make this cash outlay net neutral on an NPV basis is then:

So just surpass the minimum viable financial requirements, you don’t need to earn back the $400M spent on development and marketing. Your minimum target is actually $432M in sales.

The standard AAA game is sold at $60. Let’s assume with Day 1 DLC, premium collector editions, etc. that the average price paid is actually $70. But that’s not what the developer actually receives. We need to adjust for the 30% platform fees¹⁹e.g. From Steam, iOS or Google Play store, etc. We’ll assume no additional fees such as IP licensing²⁰e.g. if you’re making a Star Wars game, publisher costs²¹Such as recouping exclusivity funding or distribution and licensing costs for physical game discs, sales and discounts, etc. In general developers can see 50% or less in many cases from sales. This means we’re being generous here.

So with $49 actually received per game sold, we need about 8.8M copies to be sold just to hit our minimum viable financial requirements.

Remember, technically with 8.8M in sales you haven’t really made an economic profit yet! This is a minimum required to not be an economic loss.²²More technically, a zero NPV project functionally adds no net value to the company. What you ideally want is a highly positive NPV for the project to be value accretive.

Let’s see some sales performances for major AAA games shall we?

• Horizon Forbidden West: 8.4M sold in the first year
• Final Fantasy VII Remake: 7M sold in three years
• The Last Of Us Part 2: 10M in two years
• God of War: Ragnarok: 15M in one year 

So even out of several highly acclaimed major AAA titles that used their own IP with brand recognition, not all of them would pass our test of 8.8M sales.²³To be clear, not all of them would have actually required 9M+ sales to succeed. For example, the Callisto Protocol’s sales target was initially 5M. And if you factor in any additional fees that need to be paid or sales / discounts that reduce revenue from sales, the required sales get even more grim.

This leads to a bifurcation in the AAA market. Only the absolute best sellers even have a chance of making their money back. Being “just okay” is the same as failure.

So that demands studios pour more money and resources into the games they develop to increase the likelihood of success. But that too increases the cost of development and increases the sales target needed. That increases the risk, demanding a higher return to compensate. So studios pour more money into…

It’s a vicious feedback loop that makes the economics in AAA increasingly untenable.

4b. Live Services crowd out the competition

There are two fundamental limitations in the modern entertainment market:

• Can you get the attention of people when entertainment is functionally unlimited?
• Can you convince people to spend money on you when they have functionally unlimited alternative options?

When you launch a new AAA game, you need to attract people to your new release and break their existing behaviour patterns. In the previous age of gaming, this meant primarily competing with other AAA games releasing approximately at the same time as you.

This is problematic in an environment where you now need to compete with Live Services. The rules are different. And Live Services crowd out the competition.

4bi. Live Services functionally act as a subscription or tax on players.

Even if you are fully free-to-play, the Live Service still eats a fixed proportion of your free time and attention.

You maintain contact points with the IP both in-game and out of game²⁴Such as community interaction, social media, discussions with friends, etc. which take time away from other marketing efforts from other games.

This is a serious problem when entertainment and content is functionally unlimited in an always online world.

Remember our calculation of costs in Section 4a? The marketing budget is 1-2x the development budget. Why is so much spent on marketing? Because it doesn’t matter how good your game is unless people know about it and play it.

But Live Services are still a tax on players monetarily.

While the average player may not always have a predefined spending behaviour, overall Live Service microtransactions suck out some % of spend on entertainment every month. This can be from explicit subscriptions such as in WoW, regularly occurring products such as Battle Passes, or from content releases such as cosmetics.

In this sense, Live Services eat up what is known as Share of Wallet. This is the proportion of total spend from customers within an industry.

The constant microtransactions claiming a % of spend every month therefore acts very similar to a tax on the entertainment industry. It cuts the available budget and wallet spend available for all other products and services.

So now as a AAA developer you have a hard choice to make. Do you fight over that shrinking volume of wallet for non-Live Services spending? Or do you decide that it’s better to be the one shrinking a customer’s wallet rather than getting shrunk yourself?

4bii. The attention economy is just as critical as the actual money economy.

People can only spend money on things they are aware of and have an attachment to. So if you don’t even have someone’s attention to begin with, you disqualify yourself from even asking them for money.

So clearly you need attention as a prerequisite to make money. But what if you can turn attention from a hurdle you need to overcome and convert it into an asset?

This is true when a Live Service game hits a critical mass moment and becomes a cultural phenomenon.

Much has been made about how technology companies can create large user bases that then lock its users in because of network effects. If you want to sign up for a social media site for example, why would you not just pick TikTok or Instagram where everyone else already is?

But this isn’t a new phenomenon. And there is a second order effect when we move away from goods and services that provide utility to goods and services which are a reflection of taste.

Why do so many people watch the latest most popular TV series or movie? Why do people all try to wear the same fashion trend? It is not necessarily because they enjoy the underlying material. Rather, it is because consumption of the good or service acts as a social signalling function.

Watching Game of Thrones or Breaking Bad is required if you want to keep up with the latest social buzz and participate in conversations with friends and acquaintances (even if you don’t care for the story). It shows you’re someone who understands and is in tune with popular culture.

Wearing double denim isn’t necessarily because you like double denim. It is to show that you are the type of person who knows about the latest fashion and can afford to change your wardrobe to match.

Cultural phenomena therefore is about social peer pressure. “I do it because everyone else does.”

It is a network effect that has been a thing forever in the fashion and entertainment industry. But unlike network effects, it pulls in people who are not natural customers for that particular product or service because failure to keep up leads to social ostracization.
This also creates an incentive to make your AAA games as big as possible with as much content as possible.²⁵See the Ubisoft trend of endless chores and radio towers and how many gamers rate a game based on hours of gameplay per dollar. That’s not the same as a good gameplay experience! But people care about these factors so developers need to meet this need. This has several important side effects:

• Firstly, producing more content bloats costs and makes the financial calculation worse and exacerbates the risk in the project; and
• Secondly, because everyone else is doing this too, it means that what little remaining time people have not playing Live Services is no longer able to accommodate as many AAA releases.

This further leads to additional attention deficit problems. And if people like your game but don’t have time to play it, they’re less likely to buy your game. And once again, the attention economy eats into the money economy thanks to the Live Services attention tax.

4biii. Perversely, Live Services double down by turning their natural content updates into their own marketing.

In the war that is the attention economy, it is critical to have as many touch points with your customers as possible and to monopolize as much of their eyeball time as possible.

Because Live Services offer constant content updates, these content updates themselves become news. The broader community and fandom that exists around these games also generates their own content that eats up time to consume and participate in.

While this phenomenon exists for other AAA games, the problem is that eventually the conversations become stale. You can only talk about the secret valkyrie bosses in God of War so many times before you’ve beaten that horse to death. This means that attention and community has some finite lifespan where it is most active.

Think of it like radioactive material.²⁶A decently apt and amusing analogy to make with respect to online discourse given how some of it can be radioactive in its own way. Different materials have their own half-life, and naturally over time they will decay at their own rates until there are only dregs left over. The more starting material (i.e. the bigger spectacle your AAA game is) the longer it will take to decay, but decay it still will.

But with Live Services, you can replenish this attention with every content release and update. This refresh of interest keeps the community going. And with the right dynamics, you might even invert the attention decay into attention growth if you become a cultural phenomenon.

4biv. Live Services also create habit forming behaviours. Behaviours are the foundation for increasing switching costs.

It is difficult to convince someone to give up an existing product or service they use for another product or service. The barrier you need to overcome to convince someone to swap is called the switching cost.

For example, let’s say you want to convince a company to move away from Azure to use Google Cloud Platform (GCP). There are monetary costs from switching, such as the cost of labour to perform the changes and adjustments necessary.

But there are also non-monetary costs as well. More critically, the company will have already built workflows on Azure. Changing to GCP requires changing these workflows and the day-to-day activities of employees who have a routine.

Monetary costs can be solved with money. That’s an “easy” problem to solve because it is simple and straightforward. Non-monetary costs are often specific and require personalized solutions. This makes them very hard to solve.

For instance, each of these behaviours requires a different solution to reduce the cost of switching:

• I play Fortnite because it’s not just a game. It’s a place to hang out with my friends after school!
• I play FGO because I don’t want to lose my log-in streak. I don’t really like the game but since I have the game open, I’ll do my dailies anyway.
• I play League of Legends because I am addicted. This game makes me miserable and I hate it but I can’t stop.

Breaking human established habits and behaviours is extremely hard. Just think about either you or a friend who says they will exercise more but then… doesn’t. It doesn’t matter how much they really want to. Actually changing your behaviour is hard.

Games are also unique because people form deep emotional and personal connections to the products and services.

Just think about how deeply personal and intense online debates and arguments get over things that are, quite frankly, basically not important or even pointless in the grand scheme of life. Like headcanon / shipping wars / fanfic / tier lists / Rule 34 / etc.

I have strong opinions on, say, Microsoft 365 vs Google Workspace vs iWork. But it’s not mainstream to identify yourself based on your productivity application of choice despite each having user bases far larger than any game. Likewise, there’s almost no Rule 34 of this either because it’s just not something people care about.

But entertainment IP and games? This is ride or die my friend.

How high do you think the switching costs of asking someone to betray their #1 crack ship is?

In the old world, AAA games competed against each other. So if you timed your game’s release right, you could avoid the direct attention wars with other games.

But in today’s world, there is no “good time” to launch a game in order to avoid attention wars. Every day is a bad day to launch a game. Some are just less bad than others.

4bv. Live Services can price discriminate better and therefore capture more value and therefore revenue.

The holy grail in business is being able to extract the maximum possible value from all of your customers.

I wrote about Pricing 101 in my previous essay on monetization. As a quick refresher, here is the diagram on the basics of pricing:

The problem with the traditional AAA business model is that offering discrete products at singular price points is the worst way to extract value from your customers:

• The default $60 price tag means that people who value your game at more than $60 get massive consumer surplus;
  • At the same time, you lock yourself out of the market of people who value your game at less than $60.
• Discounting your game can help you capture the market of people who value the game at less than $60
  • However, it also teaches people that your games are not worth $60. If people are busy, then they can also choose to just wait for a sale or discount.
  • This is especially true for the Millennial generation that works full-time time. When time is at a premium, there is less pressure to play games the week of release. You can afford to wait and pick it up on sale and after initial launch bugs are patched.
  • The lower price you pay during the sale is the exchange you make for not being part of the initial launch hype. Although you can still participate in the community for zero cost.
  • Frequent sales and discounts can lead to reduced revenue in the long-term and undermine your brand positioning, further encouraging a “wait and see” approach.²⁷This is part of why Nintendo has been so stubborn about not discounting its games and trying to maintain its MSRP. Discounting undermines its brand. So they give up short-term revenue in exchange for longer-term brand equity. This is also related to the Nintendo Seal of Quality. The 1983 video game crash was due to a massive glut of low quality games that destroyed consumer confidence in the market. Nintendo was one of the survivors of that era. The Seal of Quality and refusal to discount games is a by-product of that era. It represents Nintendo’s confidence to consumers that their games aren’t low quality slop and therefore should be treated with respect.
• Collector Editions and other variations can help create offerings that customers can buy allowing the company to collect some of the consumer surplus for itself;
  • However, this also has the negative impact of making your other customers mad;
  • Producing these special versions also increases the costs of production;
  • If these special editions are bad quality or have production defects, it can lead to negative PR.²⁸For example, the response to the Fallout 76 Collector Edition. This is especially problematic because it is the most enfranchised customers who you are pissing off.
  • Various sales channels such as physical retailers may also demand their own special edition or refuse to carry your product. This increases complexity for the business.
  • Collector Editions are generally only sold at launch in order to capitalize on FOMO. But this means that you cannot extract value from future customers such as late-comers to your IP or people who weren’t sure about your product, bought it, liked it, and now want to give you more money but cannot.
  • Scalpers suck.
• The single purchase model means that you extract value from your customers once. However, many games these days have long lifespans;
  • Every time someone loads up a game and derives enjoyment from it well into the future, they are gaining consumer surplus that you have no method to extract value from
  • I want to be clear so I am writing this as a bullet point and not a footnote that could be hidden. It is not desirable for the gaming industry to move to a state where it tries to extract value from you every time you load a game.
  • However, purely from an analysis from the perspective of price discrimination, it is an example of the mismatch between the consumer value created versus the ability of companies to monetize and extract value.

So the traditional AAA business model is very far off from what would be considered commercially ideal.²⁹Many of these features are due to limitations and constraints before the internet. For example, the fact you had to go to a physical store to buy goods creates frictions and costs just in the act of buying itself, let alone marketing considerations. Similarly, the limitations on payment processing or shipping and delivery prevented many companies from even being able to sell things directly to the customer even if they wanted to.

Live Services with microtransactions however are very close to the holy grail.

• If you offer enough products within your Live Service, your customers can self-select themselves into their desired optimal level of spending commensurate to how much they like your game;
  • For example, giga whales will buy anything and everything you offer. So you can extract far more than a $200 Collector Edition ever could;
• You can create multiple product options inside the game that deliberately cater to different needs and spending behaviours;
  • Price sensitive players for example may focus on buying only Battle Passes;³⁰It is not a coincidence that the Battle Pass in many games is deliberately designed to be highly “cost efficient”. That’s the whole point!
• Even players who spend nothing at all and are fully free-to-pay (F2P) can still be valuable assets you can monetize;
  • In my prior essay about Mihoyo I covered the main benefits of F2P players.
• These are:
  • Word-of-mouth marketing;
  • Providing a large player base to sustain co-operative gameplay;
  • Providing a community that leads to increased retention of people who actually make you money; and
  • Finding ways to convert F2P players to actual paying customers.
• Because customers self-select themselves into their desired level of spending, you don’t need to do the hard work of trying to identify and segment them yourself! Your customers will reveal themselves to you instead.

Isn’t this business model just… so much better than the traditional AAA model?

And this is before we talk about the general arguments that people know about such as the benefits of recurring revenue vs one-time revenue.³¹I’ll cover this in more detail in Section 4bvii. Or the broader benefits of freemium models where the initial F2P experience encourages greater adoption and large player bases.

4bvi. Continuous revenue and continuous development leads to a continually improving product.

Live Services follows a broader trend in technology with the trend towards SaaS products.

Product vs product head-to-head competition is no longer static. If you plan to launch a shooter game in 2026, you will be competing against Fortnite / Apex Legends / Call of Duty Warzone / Counterstrike / Valorant / etc. as they will be in 2026.

These games will continue to get better as more content is added to those games. Each update they make has a compounding effect making their product better and better. So you’re constantly under pressure to make your launch better as well. What might be good enough in 2024 won’t necessarily be good enough in the future.³²Some readers may recognize this as a description of the flywheel effect in action.

This constantly escalating quality demand acts as a barrier to entry, limiting the potential for new competitors. Back in the old days, if you wanted to compete against Halo or Call of Duty when they were King, then the main competition would be whenever they launched a new sequel. So you had time to leapfrog them and build up a following.

This is not to say it is impossible to achieve. But it becomes increasingly harder to do so. And the increased resource demands mean that a smaller number of studios will be able to compete.

4bvii. Platforms love Live Services.

Platforms control distribution in today’s world. Imagine trying to release a game without using Playstation / Xbox / Steam / Epic / iOS / Google Play. It would be extremely difficult without significant backing and marketing.

And Platforms love Live Services because when you use their store they get their juicy 30% cut. They have a vested interest in seeing more Live Services get made. So they’re willing to pony up funding, exclusivity agreements, technical support and more.

If you’re a developer, these are very attractive terms. Think back to Section 2 where we went over project assessment.

• Upfront funding means early money that makes your NPV or IRR calculation look better;
• Technical support means lowering risk involved in the actual production of your game and therefore lower likelihood of a bad outcome;
• Exclusivity agreements might also bring marketing support, which is critical to break through in the attention economy and also improve the calculation.

This creates strong incentives for multiple stakeholders in the creation process of large big budget games to make these games Live Services.

To show the impact of this, let’s look at Sony Game & Network Services revenue over the last 10 yrs:

Data Source: Sony Corporation Investor Relations. (Note: Network Services here refers to the PlayStation Plus subscription and not Live Service games more broadly)

The trend is pretty clear. Hardware and physical game sales have grown modestly over the past ten years. But digital game sales (including microtransactions) have absolutely exploded in value.

This is how you get Sony talking about the PS5 being the most profitable generation ever despite half of its customers still playing on a PS4 instead. It’s all about the incremental spending from additional content add-on and services.

People spend less on full games. But they are more than willing to spend on content add-on.

It is little wonder then that Platforms love Live Services so much.

Even Apple has recognized the importance of services revenue more broadly and is pivoting the company away from just the iPhone to expanding the importance of the iOS store and services.

In 2016, when iPhone sales stopped growing the first thing the CFO did was to try and tell everyone how much Apple was becoming a services company, primarily thanks to its exclusive control of the iOS App Store. And this absolute control over the store is what led to the Epic v Apple lawsuit.

5. Live Services are the new apex predator. What now?

Several thousand words later, we have a better understanding of how projects are decided upon, why companies follow trends, and why the Live Services model offers massive advantages over the traditional AAA model.

It’s likely at this point that the non-Live Services traditional AAA market is going to shrink.

If games have to achieve 9M+ sales just to meet minimum requirements, then the only games that can ever be made are smash hits. When dealing with these types of extremes, even one single failure can outweigh a large number of successes.

Ideally, what you would like to see is increased experimentation from AAA games studios with smaller lower budget games at A or AA quality instead. Small experiments within a carefully managed portfolio that allows for innovation and creation of new IP.

However, with the increasing rise of indie games, the room in the market for that is limited. Indie games such as Terraria, Slay the Spire, Balatro, and Stardew Valley priced at $20 or less with single person or tiny development teams are more optimized and better suited for developing these types of experimental titles.

This also doesn’t deal with the attention economy difficulties that existing live services games pose. Producing more games just runs you into the attention problem faster.

Breaking through the attention wall requires increasingly more effort. As such, it will require increased cost. And this will further encourage consolidation and a shrinking AAA market.

The second conclusion is that the AAA market is likely to focus more on the niche.

If Live Services are providing a specific value proposition, then you want to get out of their way as much as possible to avoid competition.

In biology, crowding out means that species who have been crowded out of their original habitat instead move to a niche area and specialize in that. Expect to see the same here.

For example, a general trend in Live Services is to have highly accessible gameplay where complex mechanics are simplified to appeal to a broader audience. Compare this to Elden Ring, a game that is unashamed about its difficulty level and complex mechanics.

Similarly, Nintendo happily abandons any semblance of competing on cutting-edge graphical performance (but maintains a good aesthetic for the player experience). By significantly reducing their graphical ambitions and operating in this niche, they can run at lower costs. This in turn reduces the required financial performance and increases risk tolerance.

The other potential domain is branded IP products leveraging non-gaming IP such as Marvel’s Spider-Man. The IP licensing fees create a high barrier to entry that means only AAA-studios with large budgets will even be eligible to develop these games. These games can also leverage the core IP’s fanbase to convert into a minimum level of expected sales.

The final conclusion is that consumers are more and not less likely to pile into Live Services

The attention economy is only going to get even worse. The increasing rise of indie games and ever expanding unlimited content and entertainment online means that choice becomes even harder. And with increasing choice comes decision paralysis.

Established titles offer several benefits. Firstly, they are comfortable and familiar to people who already know those titles.

Secondly, they solve the coordination problem between friends for multiplayer purposes. It is easier for the whole friendship group to just get back onto Apex Legends rather than coordinate multiple people to try a new game.

And finally, the winner-takes-all dynamics of Live Services means that Live Services have the resources to maintain eyeballs in the attention economy. Most critically, games will need to expand beyond traditional channels to create additional touchpoints to defend the attention they have.

Establishing new marketing vehicles including live events and merchandising as defensive barriers is capital intensive and best suited for established players.

5a. Some miscellaneous predictions

Taking a slight detour, I’m going to deviate from just explaining business processes and instead make some predictions about the trends in gaming over the next few years:³³I mostly just want to come back in a few years and see what I got right and what I got wrong.

• AAA studios will increasingly focus on singular massive game releases for traditional games;
  • This will likely be underwritten by Platform owners looking to drive platform adoption to fuel Live Service fee revenue;
• AAA studios are more likely to fast-follow but time to market is critical;
  • The risk embedded in new development is too high, especially given the long development times and high costs;
• Studio brand will become an increasing differentiator given the nihilistic state of the attention economy;
  • Games are slowly progressing to a state where the primary way to outcompete for attention is finding new touchpoints to interact with customers;
  • The CPG industry transitioned from basic IP positioning to brands as statements of value. This transition can be adopted by game brands as well;
  • Currently activities such as merchandising and licensing are seen as supplemental revenue streams for companies. However, this approach is too naive;
  • In an attention limited world, merchandising and licensing should instead be used as defensive levers that defend the “attention capital” accrued by your game;
  • The fact that these levers happen to also make money and therefore can be self-funding is just a bonus;
  • The critical limitation on Live Services will always be customer lock-in. So the more touchpoints and connections that allow your product to align with and / or imprint onto your customer’s identity, the greater the defensive moat you can construct around your revenue stream;
• Live Service gameplay will continue. However, it will consolidate around a single-digit number of titles per genre;
  • The winner-takes-all and cultural phenomenon dynamics of Live Services discourage an open competitive market;
  • Consider how almost all attempts at a WoW-killer or League-killer all failed. Part of this is that each of these games do not necessarily grow the market for the genre in question. Instead, they are their own market in question;
  • Likewise, the idea of “just create a looter shooter” or “just create a Battle Royale game” don’t work because there’s not exactly a market for “Battle Royale” the same way that Breaking Bad succeeding does not create a market for “TV shows about teachers making meth”;
  • What Live Services do however is suck the oxygen out of the room and discourage the creation of competition. Consider how Facebook and Instagram made it extremely difficult to acquire venture capital funding for new social media startups after establishing dominance. It wasn’t that a competitor wasn’t possible. It was that investors and decision makers were scared from trying because their personal risk appetite could not allow them to approve a decision;
  • TikTok’s success is a reflection that there is always room for additional competitors who have a unique value proposition. However, TikTok paid dearly in CAC to fill their funnel and survive long enough to prove itself in the Western market. This is a level of risk taking that is difficult to convince others to agree to;
• Generative AI (GenAI) has been discussed at length about how it can be used to create interactive worlds and content. I do not see any major AAA game title or Live Service utilizing GenAI as a primary differentiating feature for at least 3 years;³⁴Not what the marketing team might speculatively call a feature such as funny NPC chat interactions. But an actual core gameplay feature such as persistent dynamic overworld NPC behaviour and content generation that cannot be achieved through existing techniques. So generating a random world would not count because that’s achievable today with existing algorithms such as in Civilization or Minecraft. GenAI will likely be heavily used as a tool during the development process however.
  • The primary issue is that the inferencing costs of GenAI are current too exorbitant;
  • For enterprise workflows, clear cost-benefit calculations can be performed that justify the use of GenAI. For example, if an activity costs $6 to perform and GenAI inferencing costs $2 to perform, then it is possible to build a business case to deploy GenAI;
  • However, for game interactions it is not practical to develop such a cost-benefit calculation. At an abstract level, you can argue that fixed costs such as servers can be spread on a per-user basis. However, the incremental server cost of one additional user is generally minimal;
  • GenAI inferencing however is not minimal. And a game with millions of users hammering your AI with queries will very quickly rack up substantial costs;
  • Part of this is also a psychological issue. Gamers will not treat any GenAI interaction as a workflow designed to optimize some economic output;
  • Games are meant to be fun. So if the GenAI does not produce the desired output, the gamer will repeatedly pull the RNG lever of the GenAI slot machine until it produces an output they are satisfied with;
  • A simple example is how Civilization players are willing to reroll their starting seed until they get a starting position they feel good about. Now imagine every time they reroll it costs Firaxis $1. There is clearly a big problem here!
  • Trying to cover costs through monetization also has problems. Gamers have intense reactions to hostile monetization. Charging users on a per-inferencing basis works for enterprise contracts, but not for gaming experiences;
  • An average subscription cost news to be carefully priced to reflect extreme users, and giving disgruntled players a way to make you lose money every time they press a button also sounds like a disaster!
  • Absent any sensible approach to cover the costs of using GenAI, it therefore seems extremely unlikely that this can be a major feature in AAA games until significant improvements in GenAI cost-efficiency are achieved.

• 1
  Which of these methods is chosen will depend on the internal sophistication within the company, its priorities, and its culture. The NPV or IRR approach is generally recommended in literature but their use within large companies is… varied, to say the least.
• 2
  There may be additional variations or approaches that are considered depending on the industry characteristics. For example, capital intensive industries such as Oil & Gas may consider metrics such as Return on Assets or Return on Invested Capital.
• 3
  Opportunity cost is the cost of choosing an option. By choosing something, it means you are no longer able to pick another option.
• 4
  Some people misunderstand this as implying that doing something risky means your rewards might be higher. This is WRONG! You can take on a lot of risk for absolutely nothing in return. And there are lots of low risk activities that are more rewarding than high risk activities. The primary focus of risk assessment is to take the right types of risk that are actually worth it. And not the dumb risks that are just stupid.
• 5
  You might naturally wonder if different industries should assess the cost of capital differently given that different industries have different levels of risk. And they do! Furthermore, the method for sourcing money also influences the minimum return required. I don’t intend to go in-depth into this here because it’s not important to my essay’s goals. But if you are interested, I recommend reading about the Weighted Average Cost of Capital (WACC).
• 6
  There are a LOT of caveats that will apply here of course, which is why I say “probably”. For example, let’s say that your industry is currently going through an advertising war. Everyone is burning money on advertising even though it won’t increase sales. So you expect the calculation to say this is a bad idea because you’re spending money for zero gain. But it may be the best decision to also set your money on fire as well because the alternative is not advertising and losing market share and sales. Your financial calculation might be negative, but every other alternative action you take might be even more negative.
• 7
  This is especially problematic if your project is very long! For example, imagine a 20 yr infrastructure investment. How different will the world even be in 20 yrs? Just 20 yrs ago at time of writing, we still had not gone through the 2008 Global Financial crisis, technological change such as the first iPhone and massive shift to mobile devices, Generative AI, or COVID. And that’s before any country / region specific changes such as the Arab Spring, ISIS, Brexit, Ukraine War, election results, etc.
• 8
  Let’s pretend for a moment that options such as risk sharing approaches through JVs / consortiums / etc. aren’t available. I’m just trying to prove a point here about the “Wait and see” approach.
• 9
  In fairness, Halo was not the first game to use a dual analog stick control system. GoldenEye had this as a feature and Alien Resurrection would use it as the default scheme, leading to this famous review. However, it was Halo that truly standardized and popularized this system leading it to be the default approach that is still used today.
• 10
  I am deliberately mangling the timeline here for a more dramatic narrative. Medal of Honor: Allied Assault would not be released until January 2002, and Infinity Ward would not be founded until May 2002.
• 11
  It took WoW less than a year after its release in November 2004 to achieve millions of subscribers, far more than anything Everquest would achieve. This success would provide the “proof” for funding future MMOs. For example, Age of Conan, Rift, and Aion started development in 2006 when WoW breached the 6M player mark. TERA would start in 2007, after WoW surpassed 8M players. Wildstar technically started in 2005 not long after WoW’s launch but it was acquired by NCSoft in 2007 which would have been the biggest source of funding.
• 12
  Consider the smartphone boom after the iPhone, explosion of energy drinks after Red Bull, everyone wanting to be the “Uber for [X]”, etc.
• 13
  These ideas themselves were not new. Elimination and Deathmatch type modes have been around ever since the 1980s and multiplayer shooters existed. What makes the Battle Royale genre unique is the combination of these ideas with a shrinking map to force player interaction.
• 14
  Remember when survival games was the big trend? How quickly time flies.
• 15
  In particular, Epic Games has struggled with the Live Services mode. This is what prompts them to sell 40% of their company to Tencent in exchange for support setting up Live Services as well as accessing the lucrative Chinese market.
• 16
  I’m going to ignore Epic’s failed MOBA game Paragon for the purpose of this essay. But it’s another example of trying to follow trends in the gaming industry. That being said, contrast how fast Epic Games took to launch their MOBA product versus their Battle Royale product. Time to market is critical if you’re going to fast follow.
• 17
  The former Call of Duty folks. Remember them from my story earlier in Section 3a?
• 18
  As an additional data point, the CMA report during the Blizzard / Activision merger review also indicated that AAA games often have $200M+ development budgets.
• 19
  e.g. From Steam, iOS or Google Play store, etc.
• 20
  e.g. if you’re making a Star Wars game
• 21
  Such as recouping exclusivity funding or distribution and licensing costs for physical game discs
• 22
  More technically, a zero NPV project functionally adds no net value to the company. What you ideally want is a highly positive NPV for the project to be value accretive.
• 23
  To be clear, not all of them would have actually required 9M+ sales to succeed.
• 24
  Such as community interaction, social media, discussions with friends, etc.
• 25
  See the Ubisoft trend of endless chores and radio towers and how many gamers rate a game based on hours of gameplay per dollar. That’s not the same as a good gameplay experience! But people care about these factors so developers need to meet this need.
• 26
  A decently apt and amusing analogy to make with respect to online discourse given how some of it can be radioactive in its own way.
• 27
  This is part of why Nintendo has been so stubborn about not discounting its games and trying to maintain its MSRP. Discounting undermines its brand. So they give up short-term revenue in exchange for longer-term brand equity. This is also related to the Nintendo Seal of Quality. The 1983 video game crash was due to a massive glut of low quality games that destroyed consumer confidence in the market. Nintendo was one of the survivors of that era. The Seal of Quality and refusal to discount games is a by-product of that era. It represents Nintendo’s confidence to consumers that their games aren’t low quality slop and therefore should be treated with respect.
• 28
  For example, the response to the Fallout 76 Collector Edition. This is especially problematic because it is the most enfranchised customers who you are pissing off.
• 29
  Many of these features are due to limitations and constraints before the internet. For example, the fact you had to go to a physical store to buy goods creates frictions and costs just in the act of buying itself, let alone marketing considerations. Similarly, the limitations on payment processing or shipping and delivery prevented many companies from even being able to sell things directly to the customer even if they wanted to.
• 30
  It is not a coincidence that the Battle Pass in many games is deliberately designed to be highly “cost efficient”. That’s the whole point!
• 31
  I’ll cover this in more detail in Section 4bvii.
• 32
  Some readers may recognize this as a description of the flywheel effect in action.
• 33
  I mostly just want to come back in a few years and see what I got right and what I got wrong.
• 34
  Not what the marketing team might speculatively call a feature such as funny NPC chat interactions. But an actual core gameplay feature such as persistent dynamic overworld NPC behaviour and content generation that cannot be achieved through existing techniques. So generating a random world would not count because that’s achievable today with existing algorithms such as in Civilization or Minecraft. GenAI will likely be heavily used as a tool during the development process however.