You can see clearly from the graph that there are three “eras” of MoC:

• Patch 1.0: Initial release of MoC with 10 Floors max only
• Patch 1.3: Introduction of Floors 11 and 12 as the hardest floors wth increasing focus on Boss and Elite enemies
• Patch 1.6 Part 2: Futher reduction in enemy volume and almot every fight is focused on Boss and Elite enemies

As soon as the AoE game mode launched, Mihoyo got rid of most of the trash mobs in the hardest MoC floors. Instead, they dramatically raised the difficulty with harder enemies and a greater focus on single target damage.

Afterall, players shouldn’t be rewarded twice for owning AoE characters… right?

Because HSR combat was designed to have multiple levers Mihoyo can manipulate, they can also create Blessings that are hyper-targeted. This means they can heavily influence how they want players to problem solve and avoid brute forcing content.

For example, compare the June 2024 HSR MOC versus the June 2024 Genshin Abyss:

• HSR Blessing: Provides up to 75% ATK and 150% Break Effect in addition to a 100% Action Advance. Even ignoring the action advance, the raw value of stats is worth 40 substats for every single team member.
• Genshin Blessing: Provides up to 60% Elemental or Physical Damage %. This is worth approximately 10 substats.

To come up with a Blessing as extreme as the one in HSR, Genshin would need to offer something equivalent to: When you trigger Dendro reactions 5 times, your entire team gains 944 EM

Likewise, Pure Fiction has also been a game mode that has rotated between a fixed set of 3 buffs rewarding:

• Ultimates (Patches 1.6 and 2.1);
• DoT damage (Patches 1.6 and 2.2);
• and FUA damage (Patches 2.0, 2.1, and 2.3).

It is very clear at this point that Mihoyo explicitly expects players to build teams around these themes and pull for the required supporting characters in the gacha.

3cii. Genshin has fewer levers for restrictive gameplay so its design looks different

HSR was built from the ground up to have multiple combat systems that could explicitly reward or punish players. Genshin was not.

Geshin also has a larger focus on other components in the power equation which contributes to variance between players (e.g. player skill, Artifact quality). This in turn lets players brute force content.

For example, do you know someone who basically plays the exact same teams every single Abyss (and completely ignores the Spiral Abyss blessing)?

Since Genshin cannot rely on the same explicit levers as HSR, it requires a different approach to game design to pressure spending.

3cii-1. Combat: Shield Breaking

This is one of the classic approaches to Abyss combat design. Elemental shields (generally) cannot be brute forced. This means that players must make sacrifices in team building to handle them.

A classic example is the 3.7 Spiral Abyss which had a combination of Hydro and Cryo Heralds. This is an encounter design that is explicitly hostile to Hydro characters and more specifically Nilou Bloom (which was a very strong and popular team).

As I wrote in my 3.7 Spiral Abyss Guide, Elemental Shield challenges such as these are designed as a “sink” for key characters. In this case, the 3.7 Spiral Abyss Left Half was designed as a Bennett and (to a lesser extent) Nahida “sink”.

Structuring Abyss layouts to create team building challenges therefore punishes players who lack a deep enough character roster.

3cii-2. Combat: Enemy wave structures

Teams in Genshin have specific rotation structures and damage profiles. Encounters can be designed to punish or reward these team structures.

For example, Ayaka Freeze is a team which has:

• Initial set-up period to cast buffs and pile them onto Ayaka;
• Frontloaded spike in damage concentrated in her Burst; and
• Period of downtime before the second rotation can begin.

This team therefore is good at greeting a pile of AoE mobs and then asking the question: “Will it Blend?”

But it can also be easily punished. During Patch 3.x, Mihoyo wanted to promote its latest new teams and that meant punishing older popular teams from the 2.x era.

Patch 3.4 Abyss Floor is an excellent example of punishing setup teams from the 2.x era. Abyss Floor 1-1 has 4 waves of single target enemies:

• If all 4 Ruin Machines spawned at the same time, it’d be a pretty easy clear for Ayaya Freeze;
• But when they spawn separately, the threshold to brute force this is so much higher. A 20 sec rotation across 4 waves = 80 sec which is nearly half of your 180 sec limit for a 3★ clear;
• If any single enemy requires two rotations, your time allowance basically dies; and
• As a front loaded Burst team, if you overkill the first wave then your CDs are down for the next Wave forcing you to run down the clock.

You can see similar patterns in other Abyss encounter designs:

• Most enemies are no longer Venti-able precisely so you cannot solve all your problems with one character;
• The developers decide whether enemies are allowed to spawn close enough to be grouped by Kazuha or too far away to be easily grouped
  • e.g. The Eremite spawn pattern in Patch 3.4 Abyss Floor 12-3-2 was nerfed specifically to prevent easy grouping
• The Wenut is a boss that has explicit on / off dps phases and extremely predictable attacks to punish setup based teams and reward teams with flexible rotation structures
  • e.g. C0 Ganyu can solo the Wenut because a constant stream of CAs line up very well against a boss that has low HP and is extremely predictable

Adjusting combat encounter design is another method similar to shield breaking that can indirectly pressure player rosters.

3cii-3. Combat: Imaginarium Theater

Genshin has also evolved to the point where the variance in even accounts without vertical investment is huge due to factors such as Artifact quality, player skill and game knowledge (do you know how to use i-frames?), etc.

Genshin also can’t create highly restrictive rules such as “the AoE mode” and “the non-AoE mode” in a game where players can just group enemies or manipulate the AI.

Genshin also has a problem where eventually it just cannot convince players to roll for characters with overlapping roles.

For example, HSR can convince you Black Swan vs Blade are Wind DPS characters that are both worth owning because they have different Paths and uses (Nihility DoT vs Destruction Crit Scaling).

But why should someone in Genshin own Hutao vs Yoimiya vs Arlecchino vs Lyney when their team structures are so similar? Do you really need a 4th Pyro on-field DPS character when you can’t own more than one Kazuha / Chevreuse / etc.?

At this point, there are only heavy handed options available to create restrictive gameplay. And so we arrive at the magic world of the Imaginarium Theater, which:

• Forcibly locks accounts to specific elements; and
• Restricting the number of times a character can be used per run

This form of ham-fisted restrictions is the natural conclusion if you create a game where:

• The game systems were not built from the ground up to allow for multiple ways to differentiate between characters that perform the same role;
• The power equation is sufficiently skewed to the point where players can brute force combat with highly invested characters; and
• The game developers do not want to aggressively power creep characters and instead want characters to retain value over time.

It is telling that one of the few things Wuthering Waves did not copy 1-for-1 from Genshin was the Spiral Abyss. Instead, their Tower of Adversity game mode has the same Vigor system that Imaginarium Theater and Triumphant Frenzy Event use.

3cii-4. Character Kits: The “Bait Constellations”

Mihoyo needs to create additional avenues of impulse spending to drain free income from players and encourage impulse spending.

This is especially true for long-term highly invested players who have developed accounts and large character rosters.

• These players don’t experience the same pressures to pull for new characters that a new player with an underdeveloped account does, so may pull on the gacha less; and
• These players can stockpile their free income. So when they do finally pull, they can fully subsidize their gacha with free income only.

The approach Genshin has taken with modern character design is to push for early “bait Constellations”. For developed accounts looking for a taste of vertical investment, bait Constellations helps drain savings and trigger impulse spending.

How successful has this been?

Consider Neuvillette. His C1 Constellation is generally highly regarded within the community. So how did the community respond?

• Neuvillette overall ownership rate: 65.5%
• Neuvillette C1 rate: 43.3%
• Neuvillette overall C1 ownership: 28.4%

So about 1 in every 3.5 players in the entire game owns C1 Neuvillette specifically. This ignores all the players who own C2 and up.

To put this into context, there are 8 characters in the game who have an overall ownership rate less than this.

There are 36 Limited characters in the game as of Patch 4.6. So, in a way, Neuvillette’s C1 Constellation by itself is more popular than 22% of the entire Genshin character roster.

That’s a lot of money at stake here. So it’s not surprising that Mihoyo has applied these lessons to HSR and aggressively adopted bait E1 / E2 Constellations designs.

3d. Horniness is also a form of monetization

The exception is if the motivating factor for pulling characters is horniness. Horniness is evergreen.

If the motivation for spending isn’t gameplay but horniness, then you can get away with a lot. (e.g. NIKKE, Azur Lane, etc.) However, this also requires you to have a clear design vision about building a game focused on eroticism.

As such, this can only be adopted by game studios whose vision is to build a niche game and not a mass-market mainstream game.

3di. What if I do want to make a mainstream game? How can I use this knowledge?

The idea behind horniness as a driver for spending is that it is ultimately about appealing to niche individual tastes. So we can apply the same ideas here for Genshin.

One of the problems Mihoyo needs to solve is that it is running a portfolio business now. Its products Genshin, HSR, and ZZZ are all competing with each other and your monthly entertainment budget.

This means Mihoyo needs to deconflict the marquee character releases across its games.

• For example, you know that Acheron is releasing in March 2024 and will be your blockbuster release that absorbs all the marketing hype;
• You need Genshin to not detract from HSR’s success and overshadow Acheron’s release;
• But you also don’t want to sacrifice Genshin’s revenue for free.

Your goal here is to try and segment your customers as much as possible:

• Allow your blockbuster release in one game to capture the majority of spending from the broad audience;
• Extract marginal revenue with niche designs in your second game that won’t compete for broad attention but drive impulse spending;

What does this look like in practice? Well, consider Chiori. Chiori released in the same month as Acheron, a highly anticipated HSR character.

Character	Player Ownership Rate	% Owners with C6	% Players owning C6
Top 10 C6’ed Chars
Yelan	81.3%	12.1%	9.8%
Furina	83.7%	10.9%	9.2%
Chiori	18.5%	8.8%	1.6%
Neuvillette	65.5%	8.5%	5.6%
Wanderer	43.9%	8.4%	3.7%
Arlecchino	50.4%	7.9%	4.0%
Yae Miko	55.5%	7.9%	4.4%
Ayaka	69.4%	7.3%	5.1%
Eula	34.0%	7.2%	2.5%
Itto	21.9%	6.9%	1.5%
Other chars (for reference)
Navia	36.5%	4.2%	1.5%
Ayato	32.4%	4.7%	1.5%
Alhaitham	32.2%	3.0%	1.0%

• Chiori is a character that is in the bottom 5 for overall ownership. However, Chiori’s fanbase is incredibly intense and is top 3 for C6 Rate¹At 8.8% C6 Rate, this is 2x the median C6 Rate for 5-Star characters;
• Chiori has a comparable number of people who went all-out to C6 her compared to other generically popular character such as Navia, Ayato, and Alhaitham;

Expect this trend to continue with future character releases and designs as Mihoyo experiments with ways to deconflict its character release schedules across multiple games (e.g. the split player reactions with Emilie).

4. Enshittification: When monetization goes wrong

Enshittification may be a new word for you. So let’s first define what it is. Because I am lazy, I am going to steal borrow the Wikipedia definition:

Enshittification is the pattern of decreasing quality observed in online services and products such as Amazon, Facebook, Google Search, Twitter, Bandcamp, Reddit, Uber, and Unity.

How does this occur? The creator of the word enshittification, Cory Doctorow, offered an explanation:

Here is how platforms die: first, they are good to their users; then they abuse their users to make things better for their business customers; finally, they abuse those business customers to claw back all the value for themselves. Then, they die.

I call this enshittification, and it is a seemingly inevitable consequence arising from the combination of the ease of changing how a platform allocates value, combined with the nature of a “two sided market”, where a platform sits between buyers and sellers, hold each hostage to the other, taking off an ever-larger share of the value that passes between them.

This is a pretty good observation by a non-business person about how basic Marketing 101 principles work.

To explain how enshittification (decreasing quality) affects live service games, I think it is helpful to: 

• First cover the formal Marketing theory about how enshittification occurs;
• Secondly, I will propose an alternative reason about why products and services get worse over time;
• Then I will explain how Mihoyo avoids enshittification; and
• Why enshittification can help explain why Mihoyo seems so resistant to releasing skins in Genshin and HSR

4a. Marketing 101: How does pricing work?

The fundamental principle in Pricing 101 is: You can only charge as much as the value you create.

This makes intuitive sense. If you create a product or service worth $10, why would anyone pay more than $10 for it?

However, companies do not typically charge the full amount of the value they create. This is because:

Willingness to Pay:

• People value things differently (e.g. due to brand loyalty, social impact, weather, etc.);
• Competition from other companies might result in price competition. This might lead to consumers forming expectations about what they are willing to pay;
• A great example is the $60 default price for AAA games that people expect. There’s no fundamental law of physics that says games should cost $60. But consumer expectations set a limit to what companies can charge.

Consumer Surplus:

• Companies typically charge lower than the maximum possible Willingness to Pay that Consumers have;
• Sharing some of the total value created with consumers encourages them to buy more because the customer is getting more value than what they paid for;
• While consumer surplus is measured in monetary terms, it can capture intangible aspects such as the customer experience. It’s just a question of whether you paid the amount of money that the experience was worth.

4ai. Okay so why is this related to enshittification?

Enshittification is essentially the Darth Vader approach to pricing: “I Am Altering the Deal, Pray I Don’t Alter It Any Further.”

If you are unhappy with how much profit you are making as a company, you have three choices:

• Expand: Sell to more people at the same profit margin²Expand your sales volume;
• Improve: Make a better and more valuable product that consumer have a higher willingness to pay³And therefore charge more for it;
• Reallocate: Make a power grab and go take away consumer surplus⁴Change how value is shared.

Option 3 is what the core idea behind what Cory Doctorow calls “ease of changing how a platform allocates value”.

An example of this is the attempt by Unity to change its pricing model to charge for installations. This is basically an unashamed power grab to steal consumer surplus leading to enshittification.

4b. Enshittification sometimes isn’t deliberate

Okay so if we don’t want products and services to be enshittified, then companies should focus on Options 1 and 2.

In fact, Option 2 about improving the product sounds pretty good! Afterall, the company invests in the product and makes it better, increasing total value, and gets paid for doing it. Isn’t that a win-win approach?

No.

4bi. What gets measured gets managed

Let’s say you’re a Product Manager at a tech company. You’re tasked with coming up with a new feature for a product and making people pay for it. Maybe you’re coming up with changes to the YouTube algorithm. How do you do this?

Well, first you need data. And then you measure changes in the data over time. And that creates performance metrics. And then you manage the performance metrics.

If the new feature you make means the metrics go up, then you’ve done a great job improving the product. And when it also helps the company make money, then you’ve achieved the win-win!

… right?

But here’s the problem… What metrics are you actually measuring and managing?

In the tech world for example, companies track metric such as Daily Active Users (DAU), Monthly Average Users (MAU), number of times a link is clicked (click-through rate), usage time, etc.

But do you notice something?

• All of these metrics are used because they’re easy and convenient to measure;
• None of these metrics directly actually measure if the product is delivering a good experience for the customer and therefore consumer surplus
• These metrics can go up for reasons unrelated to delivering consumer surplus.

So even if you are truly well-meaning, it’s possible to keep pushing for changes that look like product improvements but instead undermine your product and make it worse over time.

And when revenue is involved, there is the pressure to release changes not because it’s good for the customer. But it makes the revenue line go up and doesn’t appear to be harming consumer surplus…

So ship the change! And ship the next change! And ship the next change! And ship the…

And then one day you wake up and for some reason your product just seems… really terrible. And yet you can’t actually explain when you actually broke your product. The metrics didn’t show anything wrong!

But clearly this is a problem that big and intelligent companies should be able to solve. There’s no way a company like Google could ever break their core product with incredibly stupid decisions and measuring the wrong metrics.

4bii. Revenue is addicting like a drug

Watching your revenue line go up is like a drug. It feels good, you constantly want another hit, and it becomes harder over time to achieve the same high you felt the first time.

This is a problem because once you turn on the revenue tap, it is extremely hard as a business to justify turning it back off. Scope creep and financial pressure will also keep pushing you to just do that little bit more. And you will naturally enshittify your product over time because erosion of consumer surplus in the name of revenue is a one-way road.

For gaming, a simple example to point to would be the trend of microtransactions within MMOs despite also having paid member subscriptions. Afterall, the players themselves claim they’re willing to pay for these features.

But once you can justify to yourself that players should pay for incredibly basic game functionality like name changing, well… the sky’s the limit isn’t it?

This is how you end up in a hellhole of game design such as MMOs with pay to skip or constantly pushing the cash shop in the face of players through the UI. Afterall, DAU didn’t go down… right?

An alternative non-gaming example would be Amazon. Once Amazon discovered the advertising drug, it was all downhill for the search function and the spam of sketchy third-party sellers.

If customers are buying more and clicking more, then surely the user experience must be getting better! It’s best not to ask too many deep questions about what’s really lurking behind those sweet sweet advertising dollars…

4c. How does Mihoyo avoid enshittification?

Avoiding enshittification requires having a very clear design vision and strong company leadership that lets you say “No” to things.

Because commonly used metrics cannot properly measure and monitor consumer surplus, you need to:

• Create principles about what your product will and will not do;
• And then avoid temptation to deviate from those principles;
• Even if they would make you lots of money or some customers say they want it.

You can see this reflected in Mihoyo’s behaviour as a company. For example:

• They are cautious to adopt radical changes to the product just because their customers ask for it and say No to a lot of things;
• They try to minimize potential for player regret when making system-level changes;
• They adamantly refuse to add complexity to the transaction and monetization systems within the game;
• They try to understand whether players are satisfied by just directly asking the players through frequent in-game surveys rather than trying to guess based on wishy-washy alternative metrics.

4d. What does this have to do with skins?

Mihoyo seems incredibly resistant to using skins as a source of monetization in their most recent games Genshin and HSR. What might drive this?

Until they release an official statement, we can at least think about the design factors that would influence this decision.

Design Factor	Impact
Consumer spending behaviours	Does player spending on skins actually result in net new revenue Or do players have a fixed entertainment budget a month and spending on skins substitutes spending on new gacha banners? If players want to show how much they love a character, do they buy the skin or just C6 them?
Resource allocation	Skins require labour hours to produce. Mihoyo is already a world leader for speed of the content releases and their design ambition. How much more can they take on without compromising the core gameplay experience? And even if they had spare labour capacity, would they rather make a few more skins or just make Natlan more epic? What’s actually more important to them?
Character access: Skin target market	Genshin’s primary monetization is through restricting access to characters. This isn’t compatible with a skins based approach. Restricting character access deliberately shrinks your skin audience. How many people are really going to buy a Ganyu skin if they don’t own Ganyu?
Character access: Free Income	Games with a heavier focus on skin monetization either have complete access to all characters (e.g. DotA), make it possible to grind out enough currency to unlock characters (e.g. LoL, Valorant), or have extremely generous free income (e.g. Azur Lane, GBF) precisely to solve the target market problem.
Social play	Skins are more common in games with cooperative / social play because the skins provide social utility e.g. players in Fortnite who don’t use cosmetics get called “Default” as an insult, etc. However, Genshin’s primary focus is a single player experience. Skins therefore do not have the same social value to players.
Client modification	You can mod your game files locally to just reskin entire characters or replace them with new models such as Chiori Ori (KR Duck pun). In a single player game with no social element, why pay for what you can just mod? See also: Bethesda Horse Armour

These factors imply that Mihoyo has a very clear design vision about what they want their product to be:

• The core product is the open world and combat, and the vast majority of development resources go towards this;
• Mihoyo has a single primary monetization vehicle (Characters and Weapons / Light Cones);
• This single vehicle is sufficient for extracting money without requiring multiple channels to upsell players;
• Mihoyo is willing to say no to making more money if it means maintaining quality of everything else it produces (e.g. not splitting development resources)

So this is how we end up where we are here today in Genshin. A low volume pipeline of skins that are only ever released when paired with events, and with nearly half of them given away for free anyway.

And Mihoyo is absolutely okay if you don’t agree with this approach.

This is a consequence of having a very clear design vision and strong company leadership that says “No” to things.

4e. The skins monetization trap

Skins and cosmetics also contain an insidious trap when it comes to monetization.

The traditional thinking behind skins and cosmetics is that they are an easy to develop form of monetization that can exist outside of the core gameplay loop. This is only true up to a limit.

Remember from Section 3 that game developers need to create reasons for people to pull for characters through game design. And in Section 3di I mentioned how players will eventually reach character saturation and no longer need to pull for as many characters on their account.

In many ways, the same is true for cosmetics. You might buy a skin for your favourite character or weapon. Maybe a second skin. But the fifth? Tenth? Twentieth?

Remember the original revenue equation:

Revenue = Player Desire to Consume – Free Income

Characters are at least tied to gameplay. Therefore gameplay content can influence character sales. Pure cosmetics on the other hand cannot use this lever without becoming “pay to win”. The levers for manipulating the player’s desire to consume are more limited.

Skins also need to be distinct to draw spending and create the desire to consume. This in turn places pressure on your design vision. You start with benign changes, maybe breaking the colour palette for a character. But eventually you need to explore more options and start breaking things such as the character silhouette and readability. You introduce fancy effects like new animations or particles.

These new features also set sticky consumer expectations. Players will expect your new features such as particle effects, higher quality meshes and textures, etc. as the new standard of quality. This means that your cosmetics over time can only ever be monotonically increasing in quality. This in turn also drives up the cost of cosmetic development and erodes profits.

Eventually, as a developer you run out of options to get people to buy cosmetics. At this point, the customer base starts to segment:

• Collectors and whales: Much higher satiety limits (e.g. the player that buys every Lux skin no matter what) and willing to pay higher price points as well;
• Lower spenders: Players who are more sensitive to “value” and become satiated over time.

A company therefore needs to both cultivate a population of collectors as well as offer them products to collect. And this is how you end up with League of Legends announcing a 430 USD commemorative in-game skin.

This also means that your product is now pivoting toward catering to an explicitly smaller and narrower audience. And this has consequences for your priorities when it comes to what you choose to prioritize in product and feature development.

This is the trap when it comes to cosmetic monetization: Player satiation shrinks your customer base the same way that character releases can as well. And without the core gameplay loop offering levers to drive demand, satiety is much harder to break.

5. Conclusion

So what are the key lessons we have learned during this journey together?

Section 2. How does revenue even work for gacha companies?

1. Revenue for gacha games is determined by
   • Revenue = Player Desire to Consume (e.g. gacha / Resin refresh / BP / etc.) – Free Income
2. Free income acts as a subsidy for players and should be calibrated based on expected player elasticity of demand;

Section 3. Game Design meets Monetization

3. The key goal in monetization is for your game’s systems to create continuous and regular impulses to spend;
4. A healthy long-term monetization system should therefore have repeatable design levers that can be used to reliably generate demand without compromising the core gameplay experience;
5. RPG gacha games cannot rely on traditional design tools because the variation in power between players in a gacha game is too wide; 
6. Game design must rely on imposing restrictions and these restrictions should synergize with the monetization approach of the game;
7. For character driven games, the rate of acceptable character releases is governed by how well your game supports excuses to pull for characters;
8. Horniness is a unique factor to encourage player spending but can only be utilised by niche games at scale;

Section 4. Enshittification: When monetization goes wrong

9. Enshittification occurs when companies try to claim too much value and don’t leave enough value for players;
10. Enshittification can occur when companies track the wrong metrics and erode consumer surplus by not properly understanding what they are doing;
11. Even well meaning monetization systems that players themselves ask for can lead to enshittification due to erosion of value;

I hope you enjoyed reading this essay as much as I enjoyed writing it.