Anastasiya Nikolayeva, director of localization for G5 Entertainment group, a developer and publisher of casual and free-to-play games for smartphones and tablets, gave an in-depth interview to Pavel Tokarev, CEO of INLINGO game localization studio.
They talked about the structure of working processes within the localization department, and also attempted to answer the question of what model works best for localization: partnering with localization studios or working directly with freelancers. They also discussed the localization industry’s development prospects.
Pavel Tokarev: Tell me about your experience in the industry and the stages of your career.
Anastasiya Nikolayeva: Let’s start at the beginning. In 2011, I began my career as a QA for G5 Entertainment. I really liked testing and I planned to build my career in that area, but in 2013 I was assigned to streamline the system that we used to work with our outsourced team of localization testers.
After that, all work related with localization gradually started to get transferred to me. Now my job title is Localization Director, and I do everything in the company that’s connected with localization, from management of internal and outsourced employees to development strategy across all areas of localization.
So that means that in eight years you managed to go from being a regular employee to managing the department. Can you tell us what difficulties you encountered along the way?
All my work involves changing established processes at some point. The most difficult thing about my job is persuading everybody to adopt new methods. People get used to working in a certain way, and even if their way isn’t the easiest or most efficient way, they are rarely eager to change things.
An obvious example is introducing a CAT program for localization managers. At some point everyone got used to sending Excel files back and forth, so it was difficult to convince people to learn a new program and prove to them that their old method wasn’t efficient.
In my experience, I’ve found that if I approach an established team of ten people to introduce changes, I can manage to get five of them onto my side, and the rest either have to be fired or they leave of their own accord. How do you deal with this?
I try to be careful in my approach to situations like that: I find the people who will support me, and I try to rely on them as I introduce changes.
We get the department together for a general meeting, and those guys talk about what they’ve done and the results they achieved — it’s like internal PR for whatever solutions may be unpopular at that particular time.
Then I record those solutions in case studies and make them available to everyone.
So you always use “white magic”?
Yes, I try to. But you do sometimes get situations where employees really can’t deal with the changes, and that’s when we part ways.
Tell me about the company today and about the projects you manage.
I work for G5 Entertainment group. It’s a Swedish publisher and developer of casual free-to-play games for smartphones and tablets. We develop games for Apple, Google, Amazon, and Windows platforms.
Our games are intended for the whole family. They have simple rules that are easy for novices and experienced players alike to understand. Our portfolio includes: Hidden City®, Mahjong Journey®, Survivors: The Quest®, The Secret Society®, Pirates & Pearls®, and Twin Moons®.
You said that localization for some projects is done in-house, and for others it is outsourced to agencies. How did you decide on this approach?
That’s not quite right. Previously, we outsourced the localization of all our projects to agencies. Then, at some point we started getting negative feedback on several localizations from users and journalists. At first, we passed this feedback on to the translators and corrected specific problems, but our efforts didn’t improve the overall quality of the translations. So we decided that the problem must be the agency, and we found another one. But the same thing happened.
As a result, we came to the conclusion that for some languages we needed to hire freelance translators independently.
Now we are continuing to expand our base of freelancers, primarily in languages where we want to improve localization quality.
What languages had these agency problems? And what exactly were the problems?
It was always Asian languages. They were full of every kind of mistake imaginable, and it turned out the translations weren’t being done by native speakers.
What is the structure of your localization department?
I lead a department of 15 localization managers, and they each work on a specific project. For some projects there’s a considerable volume of work, and in those cases several managers are involved simultaneously.
The localization manager plays an active role in a project’s development and can influence the development process. For example, they can say that one mock-up or other needs to be changed to avoid localization problems. The manager also directs a team of translators and localization testers.
What translation management and quality control programs do you use?
We use one of the CAT programs, and we do our QA checks there as well, with custom settings. Basically, we do all our different types of work in one program.
What services do you use for task management on projects?
We have a task tracker where we list tasks, and a Gantt chart where we plan all the work. Everything is shared between projects, and then the data is collected in one overall chart.
Was it difficult to segway from planning by hand to a Gantt chart?
Transferring to a new system is always a cumbersome process. The chart helps us make the right decisions when choosing contractors for each project, and everyone understands that. If anyone sees mistakes made by someone else, then they just say something.
In your opinion, is localization more of an expense or an investment?
We treat localization as an investment. We have a standard cost list for localization. If there are any doubts about the profitability of a language, then we investigate the viability of supporting it and make a decision on whether to continue working with it or not. When we’re trying out new localizations, we calculate their profitability at the beginning.
What period do you normally set when calculating profitability?
A minimum of six months. A year is ideal, because several events with different profitability can happen in the course of a year.
What pluses and minuses do you see in working on localization in-house versus working with freelancers?
Unfortunately, I don’t have any experience of working in-house. I’ve only worked with freelancers. The pluses are as follows:
First is the quality. The translation of a game is a creative process, and you can’t do it if you don’t like the game. We had a case where a translator messed up badly in every way, from deadlines to quality, even though we were very happy with his work on another project. We asked him directly what the problem was, and it turned out that he didn’t like the most recent game, but played the previous game frequently with his wife, and even organized mini-competitions. That was when we understood that it’s very important to take the translator’s tastes into account, as it has a significant effect on the quality of their work.
The second plus is that working with freelancers is cheaper than working with agencies, because you cut out all the incidental expenses of running the agency.
However, we have one translator who has extremely high rates even when compared to agencies and, amazingly, her schedule is always full. She has stable statistics of zero errors per 1000 words. The testers never even have suggestions for improvements to her texts. Furthermore, she offers advice on how best to adapt games to the market, she plays all our games and all our competitors’ games, and she knows what games in the same genre are popular now in her region. She also regularly suggests corrections based on her user experience. For that level of work, we’re happy to pay more.
So, those are the pluses. Are there any minuses?
There are certain risks when you work with freelancers, of course. But any potential risks can be dealt with. Working with freelancers means we need a larger staff of in-house localization managers. And as well as our numbers, we also had to raise the requirements for our staff.
You also need to have a backup list of freelancers, as your main ones might go on vacation or get sick. We have a calendar where we keep track of vacations and public holidays — international and local — in every country where we employ freelancers.
What do you consider the minimum percentage of backup providers you need for each language pair to maintain stable operations?
You need a minimum of half the total number as backup, in other words, five backups for every ten freelancers.
Do you ever run into the problem where a freelancer “deteriorates” if you don’t work with them for some time? If so, what do you do about it?
In addition to an assigned localization manager, we also have an assigned team of translators and testers. If we notice that one regular provider has started to make mistakes, then we immediately look in our database to see who we can transfer the project to.
What criteria do you use to select localization providers?
There’s a range of formal criteria: provision of all required services, ability to work with our software, the right price range, experience in our game genre, and the right quality of work.
If a provider meets all the preliminary criteria, then we test them “in battle”, if you will. We check how well they adhere to deadlines, their attention to detail, their capacity for working overtime and on holidays, and basically their ability to fit into our systems well.
Could you say a few words about each of the criteria?
Briefly, this is the checklist:
- Provision of all required services. I don’t see any point in working with an LQA provider who can only provide services in some languages.
- Work experience in our genre. If a translator hasn’t worked with hidden object games before, then we have to spend a lot of time explaining nuances, which is why we work with people who have already translated games in the genre.
- Attention to detail. As part of the test assignment, we provide a pool of documents that we ask the provider to study. Going by the results of the assignment, we can see whether or not the candidate has familiarized themselves with the documents. If they haven’t looked at them at all, then we won’t work them, because we don’t write those rules to make the provider’s life more difficult — we do it to make our staff’s lives easier.
- Consistency. We place a great deal of weight on the plot of our games, so it’s important for us that the texts for new updates are translated in accordance with the existing context. We prepare accompanying documentation — a glossary, style guide, and game design document. If the mistakes we find show that a translator hasn’t used the reference materials, then we probably won’t work with them any further.
- Adherence to deadlines. Our games have events scheduled for specific holidays, and we can’t move them. That’s why it’s very important for us that providers complete assignments by the stipulated deadlines. We had one translator who finished a project just a few hours before giving birth. The nurses literally had to wrest the computer out of her hands. That’s an extreme situation, of course, but the level of commitment it shows is just incredible.
Was the translator from Japan, by any chance? The level of commitment that people have to their work there often leads to career burnout.
Yes, she was Japanese. They even have a word for the situation where people just disappear for a while. Not because they’re irresponsible, but because they’re burnt out from working.
Karoshi (japanese: 過労死) — a Japanese term meaning death from overwork. Japan is one of a handful of countries that collect statistics specifically about Karoshi. The main medical causes of Karoshi are stress-related strokes and heart attacks.
What do you advise localization providers do to immerse themselves in the world of a project?
The most important thing is to play the game you’re working on. You can’t do a good translation if you don’t see the product.
Apart from that, it’s important to play competitors’ games. That’s why our providers have most likely played other, similar games and are used to the specific terminology, names, etc. A player can get used to the fact that a booster is an object that helps you complete a level, but they may not actually understand the meaning of the word “booster”. For us it’s very important to lower the barriers to entry into the game and reduce the number of new words at the start, thus helping the player get drawn into the game.
We also welcome deeper immersion, by which I mean getting even more familiar with the game world and the setting of the game. We recently released a game about Ancient Rome, and all of us had to refresh our knowledge of that period in history.
Is there a minimum barrier to entry? How many hours does a translator need to play?
I can’t give you a rough estimate, but I can say it has to be enough for them to get an understanding of the main content. Of course, some games don’t have a plot, but they still normally have some kind of tutorial that you need to complete.
What’s the future of the localization industry? What services will be in demand in the near future?
Localization really isn’t just translation, it’s working with the developers so that the game is as financially successful as possible in the target market.
That’s the responsibility and the motivation of the people working within the company, who are really immersed in the project. I’ve yet to see any manager from an agency who’s thinking along the lines of: “What can we do to help our client earn more money in this or that market?”
[content-wide]I think that the localization industry is moving towards in-house developers.[/content-wide]
Over time, more and more game producers will get personally involved in the search for translators, or use the same freelancers on a regular basis. There will be an additional role for MLVs [multi-language vendors] — localization consulting. In other words, developing processes and finding providers for game producers.
In any case, there will still be companies that outsource their localization services, but achieving a high-quality product will be very difficult with this method.
That’s an interesting idea about consulting — I hadn’t thought of that. You know, I want to tell you one thing I’m afraid of. Sometimes I wake up thinking: “What if Google learns to localize?” What do you think, will machine translation be able to replace human translators in the next 5-10 years?
There are different types of text. For UI texts, help pages, and FAQs, automated translation techniques are already widely used.
However, I sincerely doubt that will ever work for creative texts. It is an art, after all. You need to convey character through dialogue, and engage the user in the story. I don’t yet see a way to do that with machine translation.
Thanks for the reassurance. Anastasiya, I understand that you have a massive workload — 15 people in your department and around 100 freelance translators. Could you tell us your approach to staff management and to your personal time management?
I can’t say that I’m good at managing my own time — here we are talking on a Saturday evening, for example. But I think that if you recognize a problem, it’s much easier to solve it.
Last year, I recognized that because I was not getting enough sleep, my brain was just spending half the day wondering where the next cup of coffee was coming from. I tried to make myself go to bed earlier, but there was always something super-important to do. Then I thought that I should plan some things for the early morning, but getting up at 6am just to read a book wasn’t easy.
Then at around the same time I decided I wanted to learn Japanese, and the only teacher I could find was in a different time zone. We scheduled lessons for 6am, and that helped me get used to getting up early. It was extremely difficult, but I’m proud that I managed to do it.
Anastasiya, you once sent me some recommendations on correspondence. Could you list three of the biggest mistakes that you come across in letters from freelancers?
Being over-formal. It’s not like we’re working for Gazprom, and emails of the “I would hereby like to inform you” or “My Dear Ms. Nikolayeva” variety sound very strange in the casual world of game dev. To be honest, I don’t remember exactly what I sent you. It was probably Ilyakhov’s book Pishi, Sokrashchai [“Write and Shorten”], which I recommend to everyone.
Do you really still get messages like that? Ones addressed to “My Dear Ms. Nikolayeva”?
It does happen.
I think that people in general are reading less and less. So, even in games, we need to work out how to get everything we want to say across in as few words as possible.
[content-wide]I don’t think that users play games to read text.[/content-wide]
Looking to the future, can you tell us some of the positions that you’re trying to fill now or that regularly open up? Basically, who are you hiring?
We have a project-based management system, and when a new project begins we start looking for a manager for that project. Right now we don’t have any vacancies, but they do appear periodically, and I’m always looking for team members who are engaged in their work and who are always focused on advancing game localization.
You can check for vacancies at G5 here.
Last question. Could you tell us what three qualities you value most in your employees?
I like working with people who are engaged in their profession, and who want to develop within their industry. I value honesty and directness in employees, as well as an unorthodox approach to routine assignments.