An Introduction That Is Really a Job Posting

Whether you are a fresh graduate just out of school, a seasoned veteran who has worked in many places, or a creator used to working alone, no matter how different our backgrounds may be, this article will tell you what we share in creating value for our users: a common set of values and principles of conduct that apply to everyone at PhantomWorks.

On paper, PhantomWorks is a company. But the word "company" carries an image of employer-employee relationships, and that is not the direction we aspire to. We would much rather be a team that supports one another, one that values "people" and the collaboration between them. People are not machines, and there is no need to work just for the sake of working.

Rather than a recipe that teaches you how to cook a dish, this article is more like a guide telling you which cookware is suited to which dish. And when Cooking Master Boy reaches for the wrong pan, we have to remind him. Because what we want to cook together is a dish that brings value to many people, and only the right cookware leads the way to perfection.

So, to make sure our creative passion doesn't go to waste, these principles are something we must never forget at PhantomWorks.

1. What PhantomWorks Means: Experiences That Linger

The culture and principles of PhantomWorks shape the way we do things. We respect people whose ideas differ from the team's core values, but we also prefer to work with partners we click with. Put another way, we are a group of people chasing the same goal, and we tend to use similar means to get there. That means there are behaviors that are readily celebrated here, and things that simply aren't allowed. So if you find that you can no longer follow our core values, this may not be the best place for you.

What Is PhantomWorks' Core Output?

We have made independent works such as audio dramas and music, and we also want to make games. So what kind of team are we, essentially?

If we are talking about a business model that can keep the team going over the long term, free works like audio dramas and music might earn some income through things like advertising, but games are certainly what bring in the most sales. This has indirectly led users to form a first impression of our brand as "games."

In reality, though, neither where we spend our money nor where our revenue comes from defines our roots. We need to dig into what motivates users to pay, which means we must know exactly what makes our products appealing to them. Sometimes there is an unexpected gap between the target audience (TA) we envisioned and the actual response after release (reducing this kind of misjudgment is, of course, one of our marketing team's jobs). But the qualities our core TA loves across all our products are what define the kind of team we are, and those qualities are rooted in writing and music, brought together with characters, art, and design to give birth to IP that people identify with.

By this point, we actually have a very concrete picture of our TA in our heads. If we had to say what kind of team we want to become, we hope to be a "digital content team" behind world-class IP.

Using the many technology-driven vehicles for sound and writing, we want to create a landmark IP that, like the Xeno series, gives users a complete fantasy world through story and music. That, rather than any single product, is the key to going the distance.

So we are grateful to you for reading this right now, because only if we get the chance to set out on this journey together with partners will we ever see it through to the end.

Why Build Our IP on Sound and Writing?

This comes down to both personal taste and the market. Let's talk about the market first.

"People are always searching for stories." This idea has become practically received wisdom across the entire digital content industry, and countless films, TV series, games, manga, novels, and even music have proven it. Beyond an instinctive thirst for knowledge, the deeper motivation is that people are quietly seeking a kind of spiritual elevation. To put it plainly, people strive to climb higher, and the benefits of broadening their horizons far outweigh the risks of picking up bad influences along the way.

But what if you can't understand stories in a foreign language, and even learning the language doesn't necessarily help you understand the culture? This is where "music," the universal language of the world, and the emotion-filled human "voice" come into play. The simplest example is to compare a "silent film" with a "radio drama."

Chaplin's silent films are such classics. Beyond the brisk pacing of the shots, his vividly expressive physical acting is truly astonishing. But as everyone probably knows, his silent films are interspersed with many title cards of white text on black, used to explain the more dramatic turns of the situation. Combined with the emotional information carried by the music, these are what create the complete silent-film experience. There are also many "true silent films" with no text or sound at all, and because their high artistic barrier makes them too highbrow to reach a wide audience, they represent exactly what we least want to run into when trying to sustain our business.

Radio dramas, meanwhile, are a format that has flourished more in modern times but has yet to become mainstream. In terms of how easily media spreads, the most effective is always the static image, followed by moving video, and last of all audio, which takes a long time to absorb. This comes from how easily we use our senses: the eyes will always win over the ears. So although radio dramas offer better comprehension of the story and stronger emotional understanding, they cost more to consume, and without visual memory it is much harder for them to leave a lasting impression.

So if we want to solve all of these problems, why not just combine the two? Yes, but such formats are expensive, whether we're talking about animation, film, or games, the three major popular formats. And if we're going to do it anyway, why not choose the one with the most room for fantasy?

Finally, a word about personal taste... though there really isn't much to it. It's simply that we are people who are easily moved by sound, so we want to try to move others with the sound we ourselves can create. And the key to elevating music and voice acting is the story behind them. If you share this kind of stubbornness, or if you love Maaya Sakamoto's The Garden of Everything, then we are a few steps closer to our goal.

2. PhantomWorks Works by Principles, Not Complicated Rules

We love working with creators who can turn complicated rules into principles. If rules are the most rigid thing there is, principles are what you get when you abstract them, and abstract them further still and you arrive at the culture mentioned above.

Building a culture is crucial while a team is still small and in its early days. It anchors your ship so you don't drift aimlessly. We also hope that, as we grow, this culture will show its value more and more rather than becoming a slogan no one can live up to. To build this culture, we take a very careful approach to hiring: we would rather leave a seat empty than fill it with the wrong person.

Principles aren't always easy to understand, so when someone has well-rounded abilities and a good grasp of the principles, they can comfortably figure out how to reach any goal. Large teams use more detailed rules to cope with the diversity of their people, but the process costs those rules create don't ultimately translate into value for users. The more detailed the rules, the harder they are to optimize for the team's situation, and the restrictions they impose further dampen people's willingness to produce.

So does this mean the advantages of a small team that relies on principles rather than rules, such as flat management and holistic decision-making, should be defended at all costs, no matter the price? In a creative industry that changes as quickly as ours, yes. This habit is what allows us to adapt to every change in the industry.

Flat Management: Nimbly Spotting Small, Concrete Problems to Defuse Latent Abstract Ones

Because startup teams are small, they only need to focus on the product and don't have to deal with too many marketing and operations issues. Wearing many hats is perfectly normal and very much needed, and everyone works by a single principle: for the product.

This also means that everyone deals directly with all of the product's users on the front line, and knows the problems that come up when the various aspects of the product are brought together, so they can quickly understand needs and find answers. Of course, so far we are only talking about the product side. So when we grow, how do we handle tasks that are more tightly coupled, like marketing and business?

If you believe in "user stories," then marketing, business, and so on can all be treated as products. As long as you can figure out who you are serving and what they truly need deep down, you can focus on each product and deliver real value toward the goal.

All of this relies on "minimal goal-based squads" under flat management. Regardless of whether people sit at the same vertical or horizontal level, we flexibly form teams of 2 to 4 people based purely on the goal, and choose the person best able to judge whether the people being served are satisfied (acceptance criteria). Anyone could end up in that role. The people you work with won't always be the ones who draw or write code; they might take part in business goals or even promote the product. It all depends on each person's abilities and willingness to define their own role.

If you look at how the team collaborates from the perspective of each individual and draw a relationship map, you'll find a complex, decentralized web, with some individuals quietly becoming the hubs of small areas, and several hubs in turn possibly working together on more abstract tasks. This is completely different from the tree-shaped org chart of a typical company, with none of the information gaps or time costs that come from passing things up and down a hierarchy. Meteors out of nowhere always fall from the heights where the information gap is greatest, and lowering the height from which those meteors fall is our goal.

Only with equal autonomy can we evolve quickly. You need well-rounded abilities to recognize problems, put a lot of effort into proactively syncing information with other teams, and make decisions logically. Only then can you ultimately create a high-value product.

Holistic Decision-Making: Understand Information From All Sides, Weigh the Trade-offs, and Make the Right Call

"Context, not control" is a principle Netflix has long followed. What does it mean?

Context refers to the information that influences a decision, while control refers to the means used in making it. But if execution matters so much, why not champion control?

Because what PhantomWorks cares about is a process that solves the root cause, not one that robs Peter to pay Paul. Here is a common design problem as an example.

You are a game designer, and the game you're responsible for has just launched. It's a linear level-based game, and you need to look at its retention rate to optimize the level design. Unfortunately, by the time players go from launching the game to reaching the first level, retention has already dropped to 90%. What do you do? Two possible scenarios unfold:

On a team that uses Control, the producer or your direct manager plays the game for a bit, decides the problem is slow loading, and asks a programmer from another department to optimize the loading time for entering levels. This task will take at least three days, and those three days might cost the product a lot of users, but for the sake of the long term, everyone grudgingly accepts it and gets it fixed. The programmer thinks to themselves, isn't the loading time only about five seconds anyway? Unexpectedly, because of the time pressure, the fix is a hack, and it causes some additional backend data to stop being collected. It doesn't affect players, but the operations team can no longer get any of its data. Worse still, retention for the first level only goes up by 1%.

On a team that uses Context, the person working with you on improving retention notices the situation and cross-references tracking data previously set up in other main menus. They find that as many as 15% of players adjusted the graphics quality in the options screen, and all of them ended up sending error reports when entering the game. They immediately report this to the producer, and after confirming, ask a programmer on the same squad to look for problems caused by changing the graphics quality. It turns out to be a visual glitch. They have QA quickly test the changes, and within a single day they successfully win back 8 to 9% of retention, with no other problems arising.

Between these two scenarios, besides the obvious difference in outcome, the biggest difference by far is the shared understanding among everyone involved in carrying out the task. When decision-makers aren't given enough information, they make decisions based on their own sense of reason and experience, which usually can't be applied to the actual problem. On the other hand, when the people doing the work don't know or don't agree with the value of what they're doing, after a few rounds of this their willingness and efficiency wear down, and in the end you may lose a talented person.

Fast execution, complete information, lower costs for spreading information across levels, and motivated people are the greatest values of Context. But in a team's early days, or when a task is urgent, Control does its job of stopping the bleeding.

3. How We Put PhantomWorks' Principles Into Practice

We never cut corners when hiring. Cultural fit and sustained growth are the first standard, while whether a particular skill is above the team average, whether we're short-handed, or whether someone has enough potential to grow quickly make up the second. These criteria ensure that everyone has endless things to learn from each new partner.

And most important of all: we're not afraid of making mistakes. What we fear is not knowing we've made one, and failing to learn something from it. Trying more and not fearing mistakes is what makes people the team's power source and lets us go further.

In an environment with partners like these, everyone is eager to learn new things, raise the level of their output, and communicate smoothly and happily with one another. This is not only a perpetual-motion machine for running a team; combined with financial rewards, it becomes the best motivation for everyone to create. Although we were able to get started thanks to the shared ideals of the first few people, going the distance depends on everyone's understanding of, and trust in, these ideals.

Holding Firm to What's Right to Raise Quality

At PhantomWorks, a UI artist might even write code. There's only one reason for it: I can do better than anyone else in this area, and that quality is what satisfies user value.

So how do you know what user value is? Basically, the proof is in the pudding: you only know once it's been tested. And before a product faces real users, internal testing takes its place. Setting aside any extra weight from personal taste or creative experience, what users need becomes our definition of quality.

In other words, anything unrelated to user value isn't worth doing. We don't accept our products falling outside the top ten; our only goal is to be number one in the genre. That's how we set our products apart from the competition and achieve a true "winner takes all."

Because users will only choose the best product, and this is especially true in the digital content industry. We believe anything can be done; it's only a question of how much time it takes.

Share Objective Information Generously, and Focus on the Issue, Not the Person

Everyone at PhantomWorks should have the courage to speak their mind, because none of us is perfect and we need to complement one another's strengths and weaknesses. Even if it ends in an argument, all opinions should be laid on the table rather than kept inside oneself or among a small few, where they breed resentment.

Someone who can code may not be able to draw, but they might be extremely sensitive to musical rhythm and notice a rushed beat the composer missed. That's why voicing opinions matters: anyone could spot an error that players would see as a problem. And even if it was the composer's carelessness, blaming them for it won't help anything. Understanding the whole story and then offering them another way of thinking about it, so they can find a way to reconcile it with their own ideas, is what respecting your partners truly looks like.

If we can't create an atmosphere where people dare to speak up, product quality will sink into ivory-tower fantasy, and everyone's engagement will gradually decline. Every sincere opinion you share can help the team keep evolving, not only sparking more creativity but also killing off, early on, the cancer cells that would cause a project to deteriorate.

This is also the direction that leads, as organizers, must set: let the team voice all kinds of opinions, let everyone take responsibility for their own work, and avoid making the final call unless it's necessary. Leads listen to and weigh every opinion, but they also filter out unsuitable noise, because their knowledge is still the most complete, so they still hold final decision-making authority.

Respect and Trust Others, and Be Trusted by Them

Working at PhantomWorks, trusting one another is both an obligation and a blessing. It means that as you move toward your goals, you never feel like you're fighting alone; you can trust them to have your back and set out for even bigger goals.

At the same time, to repay that help, you must also be someone others can rely on. How do you do that?

Think about the ways you communicate with your partners, whether it's a status in the project management tool, a message on a chat app, an email thread, or bumping into each other in the hallway. Whenever you talk about a task you're working on together, all the information must be "accurate."

What does accurate mean? The request, the acceptance criteria, the deadline, and the outcome of a task: all four deeply affect how the whole project moves forward. Only accurate information can ensure a successful execution. The bigger the goal, the greater the complexity, and as soon as the information is off, the risk rises exponentially.

This also means that not everything can be estimated accurately, so "taking the time to confirm before giving an answer" is also a great way to communicate. There's no need to reply hastily because of time pressure. For anything you aren't clear on, always make sure you ask and understand what it means before carrying out the task. This doesn't just make you someone who can be trusted; it also makes your partners more reliable in turn.

Reaching Big Goals by Stacking Small Ones

At PhantomWorks, any task estimated to take more than 12 hours should be broken down into smaller tasks. Why?

Besides helping everyone stay focused on the value they're delivering as they carry out their tasks, breaking a task down into its details allows for more precise time estimates. This not only increases product value but also gives us better control over project costs.

Here's an example. I'm a senior 2D artist, and I receive a task from the producer: "To help the team understand the script more deeply, please design a character prototype for the heroine." Hmm, very abstract. The task owner wants me to estimate how long it will take, so I picture in my head what I'll need to do. First, communicate clearly with the designer about what "understand more deeply" means. Second, plan how many directions I should come up with. Third, keep checking with the designer whether these directions are what they want. Fourth, actually start designing. Fifth, have the art director check the style. Sixth, have the producer and designer pick the final direction... That's a lot. Let's say 35, I should be able to finish within a week.

But then! Just confirming the direction alone took nearly a full day with the designer, so 35 is definitely not enough. But they said earlier they wanted to show the team something next week, so I guess I'll have to apologize to the PM... yet it still feels like I can't estimate it accurately. What now?

This is a case of underestimation that happens every single day. This senior 2D artist actually already had a very good sense of the stages involved. What was missing in the end was simply listing out the people involved in each of those stages and the more granular items that actually need to be done. Once you list those out, you can get a much more accurate estimate.

Once these items are broken down, you'll also find it easier to feel the satisfaction of reaching a small goal as you complete each one, rather than waiting until the big goal is done to get any sense of reward.

4. PhantomWorks Wants to Inspire Those Who Don't Have a Dream Yet

We believe that you, reading this article, have surely been moved by certain works, and that this led you to make up your mind and become who you are today. Athletes, producers, singers, directors, and others have all, to some degree, sparked a small creative flame in each of us.

Interestingly, most of these predecessors have said, "My responsibility is to inspire the next generation." It's a kind of inheritance, and also a kind of revolution. Because what they've made will inevitably be refined, keeping the most valuable parts and improving on what originally succeeded, so that the business models the next generation develops can inspire even more people.

Most of the time, the innovative products we feel so confident about are proposals from those before us that users didn't embrace. It's a bit like the survival-of-the-fittest theory of natural selection: it's not that no one has done it before, just that it's been proven to be a dead end.

But is it possible they focused on the wrong thing? Of course it is.

So we keep looking for new formats. As long as we can identify users' core preferences and test what users love in the most minimal viable way, everything else is the charm of innovative creation. Sometimes our fire isn't burning hot enough, sometimes the wind is blowing the wrong way. Beyond giving it our all, the rest is luck.

If, within a wave, we can grasp its core and innovate on it, the business value it brings will carry us into the next wave of innovation, creating a never-ending business cycle.

Depending on the size of the market, such an innovation cycle can range from 3 to 12 years. We don't have that many 12-year spans, so we aspire to create a culture of constant renewal that unleashes everyone's creativity.

Closing Thoughts

You've made it through this long-winded article. Congratulations! But, as is customary, I also have to tell you that many things involve some sad realities. Whether you finish this feeling fired up or with plenty of counterarguments, you're welcome to chat with all of us about it at the right time.

I originally thought there was no need to write this article, but after reading about miHoYo's philosophy, I gradually came to be convinced that an article like this is necessary. First, because the hiring process doesn't involve much deep exchange of ideas. Second, because many concepts are gradually forgotten amid the hard-to-overcome difficulties of reality, and over time people no longer know how to put them into practice.

I don't think PhantomWorks will grow so big that it has lots of people (at least that's what I think right after writing this), but what I am sure of is that if these concepts can be well understood and extended to every partner, we will create highly competitive products. After all, a team's treasure map is its people's ideals, and only principles that gather many ideals together and are followed by all can lead us to that legendary treasure.

I hope everyone can understand these ideas, come to see PhantomWorks as a place to call home, and embrace the path we walk together.

Best Regards,
PhantomWorks