I went to the Game Quality Forum and all I got was this lousy t-shirt*

Ewa Dacko

The Game Quality Forum 2022 took place at the end of June in Amsterdam. The event had a three-day program supported by three function-specific streams: QA, localization, and player community. It aimed to give gamedev professionals “everything they need to deliver a next-generation gaming experience.”

Below I give a handful of takeaways and insights from the conference. It is, by all means, a subjective summary, written from my personal perspective (one of a localization lead in a five-hundred-people studio creating mobile games), and you know, your mileage can vary. But as far as I could tell, most gaming companies fight similar battles. I hope you enjoy my battlefield report!

Localization in game development

All l10n professionals agree that localization should be a significant part of game development. It must be taken into account from the very beginning, not an afterthought. Localization teams (whether in-house or not) should be an integral part of the product team, not external stakeholders treated as outsourcers or contractors. 

Changing the general mindset is challenging, and

we localizers must fight for our place in the design and development process.

The transformation—changing the way people think, raising awareness, creating new localization processes or streamlining existing ones, and implementing and integrating new tools—is a long and painful process. For example, it took Capcom (the studio behind Monster Hunter) ten years to go from managing translations in Excel to a fully-fledged process where localization is an integral part of the game development.

It is further complicated by the fact that no specific and simple metrics can measure the impact and value of localization and even the localization quality itself. The most commonly mentioned localization KPIs were the number of players (DAU, MAU, DARPU), retention and drop-out rate, NPS, and LQS scores. However, all these must be analyzed in conjunction with other factors, metrics, and data. For example, quantitative and qualitative market data, user research, general game performance, global and local economic situation, etc.

Many localizers have highlighted the need to “do our homework”—conduct the necessary research to gather and analyze reliable data to base decisions on. It includes working with local partners and using their expertise (the good old think global, act local). It is a standard best practice, and all studios do it or at least plan to do it.

There was a lot of discussion around the apparently rhetorical question:

should localization be something that directly pays off? Or should it be viewed as a part of a bigger picture: as a tool to increase accessibility, enhance UX, and remove barriers for players?

(A valid point by Miguel Sepulveda from King: Nobody actually asks graphic designers how much revenue they bring into the company.) I think we all know the correct answer, but, well, we’re not there yet 😉 

Many studios (just like us) have struggled with internationalization issues, for example, localizing their games into “exotic” languages, just because they didn’t consider integrating them early enough. They’ve had problems with rendering right-to-left (RTL) languages and creating RTL interfaces, using rare fonts, implementing different word- or line-breaking rules, etc. The costs of fixing such issues are often tremendous, both in terms of money and time. And they could have been avoided if someone had thought of localization earlier in the game development process.

There was also a general agreement that localization and QA were closely related and should be integrated, but, to my disappointment, nobody had (or shared) any particular examples of how exactly they achieved this. 

Machine translation and game localization

Pretty much everyone I talked to had some concerns and/or hopes about machine translation. Still, they were somewhat skeptical. It seems that nobody believes they can put MT to good use in localizing games (even as a part of the process, like MTPE) because games need a human touch, and MT can’t guarantee a sufficient level of quality. Nobody admitted they actually used MT, but to be honest, I didn’t quite believe it. I think MT is something everybody sometimes does, but they are sort of ashamed of it. Generally, game localizers perceive MT as a substandard, cutting-corners solution.

Knowing a bit about MT/ML development in recent years, I suppose we’ll be surprised by the progress in this field sooner rather than later. There’s a general tendency among linguists to underestimate MT and its impact on the industry—after all, nobody wants to be replaced by a machine, don’t they?

Tools and technologies

This area was probably the most disappointing for me: I expected to learn new tools and solutions we could use to streamline, integrate, and optimize our processes, but nothing surprised or inspired me.

The most mentioned tools were the same that almost every 2022 company uses: Slack, Jira, Confluence, Excel, Zoom/Meet/Teams, CATs, and (sometimes) TMS tools. Many studios use their proprietary tools (also for QA or gathering feedback from players) that fit or have fitted their particular needs best. It has some obvious drawbacks. Such tools must be maintained and scaled up as the company grows, which is very costly and not always feasible.

So it’s advisable to look for third-party “straight-out-of-the-box” solutions—if an external tool satisfies your needs 80-90%, it is good enough in most cases, and you should consider adapting your processes for the remaining 10-20%.

Globalization and localization

Globalization changes the way we think of languages and accessibility (thanks, Captain Obvious). Some (more-or-less fun) facts: 

  • Some words and terms become understood globally (for example, English phrases in Japanese games), and players sometimes even expect us to “mix” languages.
  • In Spanish, the gap between regional variants (ES-ES, ES-MX) is shrinking because players consume Spanish content from all over the world.
  • On the other hand, in Portuguese, the distinction between the variants is crucial for the target audience, and they are very sensitive to differences and “errors” (like using European words or syntax in Brazilian text).
  • The importance of so-called “underserved” languages (Arabic, Turkish) is growing. It can be a sensitive topic. There are many language variants, and we must choose and use them carefully. For example, Modern Standard Arabic is the more formal variant used in official situations and not particularly suitable for games
  • For Finnish players, even minor language errors are off-putting (and they switch to English or abandon the game).
  • Regarding subtitles, the Nordic audience hates the “hearing impaired” version. They want the subtitles to be a short summary of the dialogs between characters and consider all other info redundant and sometimes even offensive.

My final thoughts

Regarding current localization trends and challenges (both language- and technology-related), the experiences, practices, and problems are similar, whether you are a tiny start-up working on an indie game or a AAA studio publishing several million-dollar titles each year.

It all boils down to thinking beforehand, constantly advocating for localization, and gaining allies at all organizational levels.

As they say: good localization is neither a marathon nor a sprint but a journey with no finish line. It was inspiring and reassuring to see that I’m not walking this path alone, just my team and me, but there are many other localization professionals sharing the struggle. 

* Actually, it was a pen and a notebook.