Gameforge was founded in 2003 and released their first game — the space-themed simulator OGame. The company now distributes dozens of games translated into more than 50 languages. The Inlingo team spoke to Head of Localization Sarah Müller, who’s been part of the company for 17 years now. We discovered how AI is affecting the translation process, what’s in store for indie developers over the next few years, and how the localization department developed at Gameforge.
“There’s a massive amount of user content in the industry”
— You’ve been working at Gameforge for almost 17 years. How did it all start?
— Before working at Gameforge, I played OGame, and was a member of the project’s voluntary gaming team. I later went to university in Karlsruhe, where our company actually happens to be headquartered, and began looking for a part-time job. I applied for the position of community-manager as a student job, and they hired me.
— How has the company changed over the past 17 years?
— The biggest change has been the company’s growth. There were 20 employees on the payroll 17 years ago, but now there are around 300 people working at our company. Apart from that, I think another important stage was launching the Metin2, NosTale, AION, Elsword and Ikariam projects. It’s been over 10 years since they were launched, but they’re still popular with gamers to this day.
If we’re talking about new achievements, then it’s got to be localizing Swords of Legends Online. It was a mammoth task: the translation of around 3 million words into 3 languages within tight deadlines. The large volume of information and tough deadlines weren’t the only challenges, we also had to coordinate the work of payroll employees and freelancers. We did it, and gave gamers the gift of quality content that immersed them in the wonderful world of SOLO.
SOLO — action MMORPG from Gameforge. Source: Steam
— In an industry where young people are inclined to change jobs every couple of years, staying in one company for such a long time seems fairly unusual. What keeps you working at Gameforge?
— One of the main reasons is people. There’s a very close-knit team at our company. We love working together, overcoming each new obstacle that comes our way. Each of us is genuinely inspired by what we do. And not only is this passion a source of motivation, it also tends to be contagious.
Challenging tasks are another source of motivation for me. We have a lot of work and face challenges every day, but we can even get great satisfaction out of them. Developers and games are constantly changing, so even after 17 years, I’m still facing new problems and learning new things.
— Gameforge started with client-and browser-based games, then transitioned to mobile games, and has recently begun working with indie game developers. Which projects do you think you’ll be working with 5 years from now?
— Creating the IndieForge label showed us that the indie sector is growing and evolving (ed. IndieForge is a division of Gameforge that supports small independent game studios from across the globe). We were caught up in a MMORPG mania of sorts for a while, but we’ve been exploring the indie sphere for the last few years. Our team discovered a number of fascinating indie games, and we wanted to help bring them to the market. We’re likely to continue working with this particular sector over the coming years. There’s a massive amount of user content in the industry — this trend is only set to increase.
We continue to support some of our classic MMO games and their global communities, but Gameforge is growing and changing in parallel. We now have a whole New Games team dedicated to new projects. Some of them are still being kept secret, but we’ll be making a couple of interesting announcements in the coming months.
“If a developer wants their game to be successful, they need to think about the global market”
— You started out as a community manager, but then turned your attention to localization. Why?
— When I arrived at Gameforge, there was no special person or department responsible for localization. The community managers themselves had to manage the translation of games into their native languages — typically English or German. When we had a volunteer who was prepared to complete the translation of one game or another, a relevant manager would be assigned to oversee their work.
The company grew over time, as did the volume of translation required. The global community broadened, and so did the number of languages that projects needed to be translated into. My boss at the time suggested broadening the structure of our studio and creating a localization department. I took on all the organizational matters: tools, processes, staff, and vendors. That’s how an entire department was built from scratch at Gameforge.
— What did the localization industry look like in the 2000s?
— The main problem was no one had any real idea of how the localization process should be configured — we didn’t know the optimal methods for solving the problems we encountered. At the beginning of my journey, I was sort of isolated from the industry: I just did what I thought was necessary. And, it turned out not to be a completely wrong approach.
There are still a lot of newcomers in this field. Even now, I still have to explain how the localization process works sometimes — some developers don’t even take it into consideration when they’re creating their games.
— A couple of years ago, you told Slator that one of the biggest problems in your job is getting developers to understand what exactly localization involves. We often hear industry experts say that developers underestimate the importance of localization. Is it possible to say the situation has improved somehow over the years?
— Some teams now understand what localization is and which points are important for the process to run smoothly. These are generally developers who didn’t think about adapting a project for different regions at first, but learned the hard way by then having to correct their own mistakes.
This task is now made easier by engines like Unity and Unreal with built-in translation dashboards — they help a little. However, some things still need to be repeated time and time again — that we need more context, and that the automatic conversion of lowercase letters into uppercase may not be appropriate for certain languages.
It gets even more interesting when changes are made to the game, e.g., gender markers. But we can solve these kinds of problems fairly quickly and effectively thanks to our vast experience.
— Why should developers in particular think about how their game will be localized?
— It all depends on business objectives. If a developer wants their game to be successful, they need to think about the global market, not about the requirements of one specific country or even one region. It’s very important to deliver content in the languages of potential users, which will significantly increase the project’s popularity. English has become a lingua franca of sorts, but researchers still believe people are more than likely to spend time and money on a project which has been localized for their countries.
— At this stage, are the developers working with you at Gameforge beginning to plan localization? Could you tell us what developers do to make your job easier?
— For new products developed at Gameforge, we provide a list of requirements — it should be taken into account from the very first stages of work. It becomes a well-honed process over time: when new functions appear or changes are planned, developers already know what they should do that’s important.
When it comes to translating products offered by third-party developers, it all depends on how soon in the development stage we receive the product, and how much experience the developer has of localization. We take meeting the client’s requirements and deadlines very seriously. It’s absolutely crucial to adapt to the other company’s workflow — it’s very often different from our own.
— How do you tell a good translation of a game from a bad one? Can an excellent translator from a different field quickly adapt to game localization?
— A good translator of games should have a deep linguistic knowledge of both the source- and target-language they work with. They should take an interest in the products they translate, and should also be familiar with gaming terminology.
Apart from that, they should have an idea of the technical side of the text reflected in the game — tags, formatting and variables. They need to be careful with these elements so that they appear correctly in the translation. Amongst other things, a general understanding of gaming mechanics will help them find the appropriate terms for skills, characteristics and other details.
A translator from a different industry can pivot to game localization if they possess all the above, and can quickly pick up gaming terminology. Sometimes a creative approach can be even more important than just a good translation.
— With the arrival of machine translation post-editing (MTPE), do you think the role of editors is becoming more important?
— We still need people to achieve high quality, but their role is gradually changing. Translators need to pivot their skills towards editing, familiarize themselves with new technologies, and learn how to use them.
— How can you make localization high-quality and fast? Do you have some sort of secret?
— I think everyone has their own secrets. One of them is a team of translators with a proven track record who can ensure high-quality even when deadlines are tight. Another thing that helps achieve a result is having experienced localization managers monitor the workflow and answer all the questions.
One factor of no small importance is having the right tools and good skills to work with them. I recommend getting familiar with continuous localization, and running quality assessment throughout the entire course of the project, not just at the end.
“AI is just a means of creating a lot of content in a very short space of time”
— Does Gameforge use any machine translation in the localization process?
— In certain cases we do that, but only for documentation which isn’t intended for the user. The technologies available at present still aren’t good enough to grasp the context and style of the source text. When using machine translation, problems also arise with variables — ones like {1}, {2}, {3}. They’re often used in sentences to replace relevant numbers, names, and named items. But variables can even occasionally cause problems for translators.
— Some experts say AI will change localization in the next 5 years, while others predict faster changes. In your view, what should we expect?
— I’m confident AI will begin to be used for translation sooner or later. When exactly? I don’t know the answer to that. But we should always be prepared for changes. Experts should already be studying post-editing methods and ways to interact with systems such as ChatGPT, in order to get the best results in the shortest possible time frame.
— Do you believe ChatGPT will change the way game studios or the creators of games work with localization forever?
— I believe a lot of things will soon be reworked and adapted for new technologies. The use of AI is just a means of creating a lot of content in a very short space of time. It could lead to an increase in the number of indie games with plots that no longer need to be created by a huge team.
We tested AI for our objectives and found it still can’t be used as the starting point to create content for games and marketing purposes — only as a supplement.
— Developers have begun integrating AI dialog generators into their projects to make dialogs with non-game characters more captivating. If this trend catches on, how will it affect the localization process?
— If that happens, the part of localization that still uses humans will be no more. If AI is already there, why try to extract it from the game in order to translate it and then reintroduce it back into the game? Developers want to use it to translate content directly. But they’re unaware of the problems AI can lead to.
We’re now encountering cases time and time again, where machine translation can’t understand the precise meaning or intention in the source language. Even the most advanced technology needs to be checked by a human, as robots aren’t constantly immersed in the culture, traditions or the history of another country. It’s important to keep all of this in mind during localization.
“The geography of game launches is only set to expand”
— Have you made any mistakes entering new markets? Tell us what can go wrong.
— It’s usually small mistakes. For instance, cultural issues which hadn’t crossed our minds before can always crop up. Thanks to reviews from the community team, we can spot them and make changes which take the nuances of a language into account or use the appropriate terms.
In rare cases there are also problems with the text — sometimes it gets cut if it’s too long. On even rarer occasions, the meaning of a word in the source language can be drastically different from the one that ends up in the game — that happens if the meaning of the word can change depending on the context.
Certain games still in the trial stages have featured funny translations by AI, e.g., the word “wrap” in the source language was changed to “burrito” in the target language. That’s exactly why in our case, the text undergoes multiple levels of translation, editing, import, export and correction before ending up in the game.
— The game industry is on the rise in Middle Eastern and North African countries. In your experience, what are the biggest problems when translating games into Arabic?
— Arabic is written and read from right to left. This is one of the biggest problems in and of itself — not only do you need to have the appropriate translations, you also need to adapt the game by changing the whole user interface and icons.
Apart from that, problems can also arise with cultural adaptation. The content shouldn’t mention short skirts or alcohol. These points are hard to change. We sometimes edited visuals where the developers thought it was possible. In other cases, we were only able to change the wording to avoid delaying the game’s launch.
— And what’s the situation with the game industry in India? How can game developers and publishers reach audiences there, where people speak so many different languages?
— As the head of a Western publisher, we’re primarily oriented towards audiences in Europe, North America and Latin America. This mainly has to do with licensing, as our partner developers often have publishers in other regions, including MENA (ed. Middle East and North Africa). If a developer comes along in the future who’d like us to launch their game in India, there’s nothing to stop us from doing that.
We’re very aware of India’s size and potential as a market, and also have great respect for Indian people and culture. We’re happy when we see the Indian media discussing some of our games and updates. I’m sure it’s only the tip of the iceberg — we see what’s written and published in European languages.
Gamers in India are known to be highly proficient in English. We hope they’ll be able to enjoy a taster of some of our games. Many of them are free — you can play for a couple of hours to see whether it’s worth continuing.
— Are there any other developing markets where more localization efforts will be needed in the coming years?
— As Western publishers, we sometimes transcend our regional boundaries out of interest, and publish games in Brazil, China and Japan. We have various ideas when it comes to MENA, but I can’t share them publicly at this stage.
As far as the overall game industry is concerned… My sincere hope is that the geography of game launches is only set to expand. We live in times of globalization with an increasingly competent workforce, and the quality of tools is increasing, so the future looks very promising. I hope we can continue bringing people together with the help of video games and entertainment.
