The cost of being misunderstood

[00:00:00] Tamara Tirják: You’ve probably heard the common phase of good localization is expensive. A bad one is even more expensive. Yeah, I really like to throw it around a lot. In games, this can play out in two different ways. So games are different from most other industries in the intensity of the competition. So for example, if something goes wrong in a Microsoft release, then yes, it can definitely be a big bad PR problem.

[00:00:22] But will I uninstall my Windows 11? Probably not. Games are different. So if you look at only PC gaming, then there are about 130,000 games available on Steam Store right now. So if the game that I bought is unplayable or even just unenjoyable or unengaging and it’s frustrating me because of the language, I can get a full refund within two hours.

[00:00:46] of playtime with just a click of a button and then go to the next game from my ever-expanding wishlist. And the mobile market is even more intense with the free-to-play and freemium monetization model, but it’s even easier for players to just move on to the next very similar title in the genre.

[00:01:10] Jason Hemingway: Welcome to In Other Words, the podcast from Phrase, where we speak with business leaders shaping how organizations grow, adapt and connect around the world. I’m Jason Hemingway, the CMO at Phrase, and my guest today is Tamara Tirjak. She spent a decade building, leading localization at Frontier Developments, one of the UK’s leading game studios. And during that time, she invented an entirely new fictional language for one of Frontier’s titles, and we’ll come back to that later. She’s since moved on to independent consulting and research, working with organizations on a question that doesn’t probably get nearly enough attention in the boardroom. And that question is, what does getting language wrong really cost a business? So, Tamara, it’s great to have you on the show.

[00:01:51] Tamara Tirják: Hello, Jason. Thank you for inviting me. It’s a pleasure to be here today.

[00:01:54] Well,

[00:01:55] Jason Hemingway: it’s nice to see you too. So, before we get into the kind of big themes and that language you invented, which is intriguing, I’d love to hear the story of how you got here. You know, 10 years at Frontier, building something for nothing, and then a move into this consulting and research side. What was your, take it from the top and then tell us what was your thinking as you were making that call and that change?

[00:02:21] Tamara Tirják: Throughout my career, there has always been a theme of expanding my horizons. So, back in the days, about almost 20 years ago, I entered the industry as an English to Hungarian technical translator. And as I took on various roles at the language service providers, I kept understanding more and more elements of the localization process. Then arriving at Frontier about 10 years ago to build up the localization program from scratch. I discovered how localization fits into the wider business context. And after 10 years there, even though I love the company, I love the games, I had the chance to work on, and I deeply cared for my team, for the team that I built. It was time to expand my scope further to the wider industry. And honestly, I couldn’t have chosen a better time to step up and support the conversations around the role and value of localization at various organizations, because, as you know, with the fundamental changes triggered by new technology, inspiring or forcing a lot of localization teams both on the vendor and the client side to rethink their identity and value. These are both very challenging and very exciting times indeed.

[00:03:46] Jason Hemingway: Yeah, I think you’re absolutely right on that. A curious phrase that you used of both inspiring and forcing is quite, I think you feel it everywhere. It’s not just in teams that are responsible for localization, but global content teams, marketing teams, sales teams, service teams, all over businesses. I think you’ve got those two factors at play. So, let’s go, let’s talk a little bit more before we move on to sort of today’s kind of conundrum for everybody. One of the things that strikes me about your time at Frontier is that you weren’t just running kind of a localization department, you were building an entire ecosystem from the very start, and I’m really curious to know how you approached, you know, particularly those technology decisions and whether the principles that guided you then, let’s talk about them and whether they still hold.

[00:04:40] Tamara Tirják: That’s correct. So, games are highly technical products and localization, just like any other discipline, is expected to meet the challenges and support the innovations on the road towards the game’s creative vision. So, at Frontier, we used, and the team there still uses, a combination of commercial and proprietary tools with MemoQ being the cornerstone TMS connected via APIs to the localization asset management tool sitting within Frontier’s proprietary Cobra game engine. The thing with game localization is that while a lot of the challenges that we face are the same as other niches, such as frequent content updates or the quick turnaround times, and they are handled effectively by pretty much any commercial TMS these days, there are also unique needs as well, which often call for custom plugins or even full tool suites that are built in-house. So, such needs can be handling multilingual voiceover, for example, or relying on video or visual context for translation. You were asking about the guiding principles in building this ecosystem. They were on one hand centralization, and the other one is, proximity to the game teams for alignment on the vision and the delivery. Whether they still hold true and whether I’d approach it the same way now is a great question. Aligning localization efforts with the game team’s creative vision and endeavor is still the main success criteria for game localization programs. So, creating tool ecosystems that connect localization to game writing, UI design, or game production still remain a priority these days. As for centralization, while I believe that still makes perfect sense for most programs, GenAI gave rise to the so-called shadow localization, which is creating a new kind of decentralized model, isn’t it?

[00:06:44] Jason Hemingway: Yeah. Yeah.

[00:06:44] Tamara Tirják: Which can be very risky from a governance point of view, on one hand, and on the other hand, it risks diminishing the specialized value proposition for localization teams, and it does create a new minefield for localization leaders to navigate. At the same time, it makes people outside localization think about localization problems in brand new and sometimes very interesting ways. So, if localization specialists can keep control of the narrative, they can eventually benefit from the wider interest in language operations, and it gives them an opportunity, maybe to step up and lead these discussions themselves.

[00:07:31] Jason Hemingway: Yeah, I think you’re right. There is definitely a huge opportunity for people to raise the profile of, you know, language in general and the mechanisms that you can now use to do that, you know, at greater speed, greater efficiency, but also high quality. But you’re right, it’s that this idea of shadow localization, which I think to the uninitiated is quite, oh, well, isn’t it just, can’t I just do it myself? And it’s like, yeah, you can, but there’s loads of pitfalls that, you know, people that have specialised in this and the professionals in this can kind of point to. That’s not to say that people shouldn’t try and aspire to move quick, I think. So, furthering that kind of idea, many organizations buy tools and assume that that ecosystem those tools fit in sort of follow. At what point, from your view, do you stop buying kind of point solutions and invest in a platform that you can build on as you scale?

[00:08:27] Tamara Tirják: In my experience, building anything in-house is not a decision to make lightly. It’s easy to say, especially for a tech-savvy company, to say, “Hey, look, we have all these smart, talented people. Surely they can build something amazing”, which is true in a way. But what is often ignored is when you look at the total cost of ownership is not just that the work often turns out to be more complex than initially expected, taking longer to complete the tool, but that the development does not stop at the event the proprietary tool is shipped, because the tool will sometimes break and needs fixing. The rest of the ecosystem will keep changing.

[00:09:22] Jason Hemingway: Yeah. Yeah.

[00:09:23] Tamara Tirják: The tools need to adapt to the changing requirements. So, new versions need to be developed, deployed, documented, and bloating into a much larger commitment than anyone anticipated initially. So, unless the control provided by a proprietary tool and the custom system is mission critical, can create market advantage, or it’s simply not available commercially elsewhere, then buying a tool or partnering with an external technology company is almost always the better choice.

[00:10:06] Jason Hemingway: I think that’s a learning that you’ve seen firsthand. And now, in your kind of consulting side of things, you’ll be witnessing much more of that. Before we kind of get more into that detail, I think I’d agree with literally everything you said. It appears simple, but really, you’re taking on a huge… you’re basically running software and that’s not, maybe not your core operation. If you are even a tech company, you’ve probably got your own software to worry about rather than the stuff you’re building internally. But you’ve built, let’s talk about sort of that automation, particularly when you’re running localization, because that’s what it’s all about. How do you automate, get things quicker, how do you orchestrate things to be done? And that pipeline that you were kind of content that you were, or in-game experience, let’s say, that you were kind of turning around. You know, you had like 48-hour turnarounds for games like Elite Dangerous that any gamer will absolutely know and see that. I happen to know it as well. What point does that sort of automation tip over and become the infrastructure that everything else depends on? You know, we talked about the ecosystem, but if you’re thinking about that automation piece, at what point does that become infrastructure?

[00:11:25] Tamara Tirják: Oh yeah, the 48-hour pipeline for Elite Dangerous, yeah, fly safe, commander. So, that pipeline included not just human TEP, but also a human-reviewed text-to-speech conversion as well for the in-game news service that we had in the game. When building automations, my ultimate goal was always to shift attention from manual tasks to actual value creation, such as stakeholder management and aligning localization efforts to the game’s vision. I know that’s a recurring theme in my thinking. But I’d actually like to challenge your question. While I don’t think that any team can afford the luxury of not automating, I do see the risks of everything and anyone depending on an automated pipeline. So, you see, if an automation fails, which every now and then they do, will people know how to work around it manually to complete the task, or will they be just there, sitting completely stuck until someone comes along and fixes the process for them? And is the automated pipeline, and they tend to become eventually a complete black box? So, you wouldn’t even know where you would start troubleshooting it. The kind of automation that I trust most is the kind that is invisible when it’s working, but easy to navigate when it’s not. The system should make it obvious where things break or broke and why. One of the best examples I know is BeLazy, that they get

[00:13:12] Jason Hemingway: this balance right, so it connects; the localization tools remove the repetitive project management tasks, and then if something goes wrong, I can actually go in and see what happened and fix it myself, even if I’m not an engineer. I should say that I’ve worked with BeLazy, so I’m biased, but that’s also why I know the platform so well.

[00:13:13] Yeah, no, but I think, I mean, there’s a principle there, isn’t there, that you can draw out. And when you almost elevate this conversation to any AI that’s operating in any business, there has to be some kind of traceability. Accountability is really important, I think it’s one thing, and we’ve talked about that on this podcast before. And the ultimate thing you’re after is trusting that the box is working. And actually, if the box doesn’t work, there’s alerts, there is traceability of everything that you can go in, and I think you solve that in different ways with different tools, but I think that’s the ultimate objective. So, I think you’re right. My question was probably pointing towards that automation. And you probably answered the question nicely, inasmuch as the automation is working when you don’t need to worry about a problem. And that’s the tipping point, it’s that the workflow is working, but I think you make a fair comment, is once it breaks, that’s the interesting part, and I think you also, would you agree that you need domain expertise to be able to go and then go, “Well, what is the problem here?” Perhaps it’s that, you know, you can get the alerts, but how do you build that workaround? That’s where maybe the localization professionals of the future or the content, you know, multilingual content specialists of the future will be looking at these kinds of pipelines a lot more, perhaps than doing the actual work of translating or whatever it might be, you know?

[00:14:30] Tamara Tirják: Absolutely. The other ones understanding their own pipelines and what maps to what, and how they want to operate it efficiently.

[00:14:38] Jason Hemingway: That’s interesting. And the systems can help as they grow, I think, will help give more context, give more traceability. So, anyway, it’s exciting on that front. But talk about your recent research that you’ve been doing, I think if we switch gears a little to some of the research you’ve done with Nimdzi. And the recent research you contributed to around data challenges that locked teams face makes a really sort of provocative argument about why localization keeps losing the business case conversation. And I think I was thinking, if you had a CFO in the room right now, what would you want them to hear about it? Because I think that’s interesting,

[00:15:18] Tamara Tirják: right?

[00:15:18] Yes, that was a very exciting topic, and I love contributing to that article and collaborating with the other Nimdzi experts. So, that paper was led by Giulia Greco. She’s the main author of that. And one of the main traps that we see is that localization often gets pushed into a cost conversation. And they are trying to prove their ROI. And we also explore why this is often impossible for them to do. So, what we found was that in many cases, localization teams simply just either don’t have access to the data that they could use to justify their contribution to the revenue and the growth of the company, and in other cases, well, localization is positioned as a shared service, a support function, rather than a profit and loss owner. The other teams, such as marketing, customer support, regional managers, would often claim credit for the international success that localization enabled. Not cost, enabled. So, the article works as a survival guide for localization teams operating in a suboptimal data situation, listing a couple of tactics that they can use to still tell a story about their value as a revenue enabler rather than a cost center. So, you were asking about CFOs.

[00:16:55] Jason Hemingway: Yeah, if you had one in the room.

[00:16:56] Tamara Tirják: So, there is something that I think CFOs almost never hear, which is when localization goes over the budget, then the root cause usually is not localization, but initial underscoping of the work or other upstream problems, such as late pivots on content, inefficient communication between teams. So, localization is simply the place where the cost becomes visible because we are at the very end of the chain. So, when a CFO is looking at the cost line for localization and asks why it went over, they are often actually looking at a warning signal for the whole content pipeline. So, this realization can shift the efforts to address the actual root cause of the overspend as an issue, rather than just patching up the symptoms, if that makes sense.

[00:17:58] Jason Hemingway: Yeah, no, I love it. I think that’s a really good way of thinking about it. It’s funny because, you know, the credit is given to the shared services, but when it goes wrong, the blame is squarely with one kind of team. But I think that’s an interesting paradox, I guess. But I think, yeah, CFOs do well to look at the entire process rather than just focusing on the overspend and seeing that single point. And as most teams running multilingual content, global language, operations, whatever you want to call it, as most teams that are running that, you know, they are embedded, like you say, in lots of different teams’ processes. And therein lies both the challenge and, again, I think it’s an opportunity to sort of make the case that you can add value to lots of different places. Let’s move on slightly to sort of about, you know, if I had a CFO in the room as well, I’d say, well, there’s a cost of doing this and a cost of getting it wrong, you know. From your experiences, whether that’s in a game, when language goes wrong. Let’s take a game, for example, what’s the impact on the player, the team, the release schedule, the lessons that you’ve learned there for other businesses and other industries, maybe?

[00:19:20] Tamara Tirják: You’ve probably heard a common phrase, ‘good localization is expensive, a bad one is even more expensive.’ Yeah,

[00:19:30] Jason Hemingway: we

[00:19:30] Tamara Tirják: like to throw it around a lot. In games, this can play out in two different ways. So, games are different from most other industries in the intensity of the competition. So, for example, if something goes wrong in a Microsoft release, then yes, it can definitely be a big, bad PR problem. But will I uninstall my Windows 11? Probably not. Games are different. So, if you look at only PC gaming, then there are about 130,000 games available on Steam Store right now. So, if the game that I bought is unplayable or even just unenjoyable or unengaging, and it’s frustrating me because of the language. I can get a full refund within two hours of playtime with just a click of a button and then go to the next game from my ever-expanding wishlist. And the mobile market is even more intense with the free-to-play and freemium monetization model. But it’s even easier for players to just move on to the next very similar title in the genre. Second point. Language problems can also be a risk for the release schedule. Console games need to be submitted to Sony, Microsoft, or Nintendo to be approved for release on their platform. All three platform holders have very strict guidelines about not just the technical requirements about crashes and such, but also the correct use of branding and terminology, legal text for PlayStation, Xbox and Switch, respectively. For example, in Spanish, both European and Latin American, the left and right stick on a controller is called stick on Xbox, joystick on PlayStation, and palanca on Switch. So, if you use the wrong one, that means the game can fail submission, and the team has to resubmit it with the fix in place. And that means additional work, costs and potential delays for the game team. And localization does not want to be the one called out on that big meeting while we failed submission. So, while this platform compliance requirement is specific to the games, breaking the immersion and failing to communicate with the audience in a way that truly resonates is not a unique challenge.

[00:22:06] Jason Hemingway: Yeah. And I think also, you know, if you think of the Sonys of the world, they’re almost like, you’re almost having to comply with regulation in a sense, aren’t you? So, you can put it in, and then regulated industries must have the same thing when you have to submit things to be approved. So, you mentioned a word there about cost, and I’d like to sort of explore your point of view on cost a bit more. So, lots of businesses treat localization, or multilingual content, or languages as a cost to be managed rather than a capability for the business to build. How do you, and have you seen it done, change that conversation leadership level away from cost to more of the value, which you sort of talked about earlier?

[00:22:52] Tamara Tirják: For framing the contribution of localization, then good, solid, hard data has always been my best friend. So, luckily, I’ve always been in a good situation with data people being my best friends, some of my best friends at the company. Because I knew that it’s hard to argue with, look, we supported XY language in this game, and it made N times as much revenue as this other game, where we didn’t. But at the same time, it’s important to be honest and include it in the report if one language or the other did not perform as well as we thought they would. So, every learning is a new data point for next time and helps flexing our strategy muscles moving forward, but focusing on the value of localization does not mean that costs don’t have to be managed. What is important to remember, though, is that, as we’ve already discussed, our hands are tied in fully controlling the localization costs. So, we are not the ones producing the content to be localized, except when we are. For example, at Frontier, I manage the team of game writers as well, which made it a lot easier for us to have these conversations about working together to efficiently control the costs and the budget through cutting down churn and rework, for example. And on the other side, with marketing teams, we had a lot of conversations about language tiers for the different content types, which is an absolute must to optimize localization spent where it makes the biggest impact. So, working together with game writers on one hand to control the scope and rework, and collaborating closely with marketing on the other, on focusing the budget on what really matters. Those are very effective tools for cost control, while at the same time, you know, building the capability and the strategic value of the function.

[00:25:09] Jason Hemingway: Yeah, I think that’s really important. The strategic value often gets lost in some of those conversations. So, in that kind of high-velocity environment like gaming, as you scale, how do you hold on to that sort of authenticity and trust? Because that’s not easy to do at a very fast pace and at big scale, right?

[00:25:35] Tamara Tirják: Exactly! But let me make a distinction here. So, velocity and scale are often not part of the same equation. So, let me explain what I mean. Short release cycles and quick turnaround times are quite typical challenges for the mobile game industry, where they have 2-week cycles often for new update drops, and they have content pipelines that are geared towards addressing these very challenges, often with machine translation or AI capabilities for a first-pass translation or applying AI

[00:26:13] Jason Hemingway: QE.

[00:26:14] Tamara Tirják: Whereas scale is the challenge for AAA premium games, where the volume of word count is often in the millions. So, in the mobile game scenario, player trust is maintained by staying consistent with the tone, verifying translations in the context, and making decisions based on data to capture the engagement of players with a piece of content. They are very surgical about controlling everything in mobile. I find it fascinating. Whereas in a AAA scenario, robust content pipelines need to be in place to have a solid foundation on which you can build creativity. It’s absolutely impossible to focus on enabling linguists to do the best work in their lives if there are fundamental issues with sharing information, such as reference materials, glossaries, style guides, or file versioning. It doesn’t exist. You have no idea what’s the latest version that you’re working on, or the content planning is completely broken.

[00:27:33] Jason Hemingway: Yeah, I like it. So, it’s two different models, and you hold on to the authenticity and trust in different ways, given the needs. So, look, we’re at the mid-show sort of moment, probably a little bit after that right now. So, we asked you a question that’s sort of non-related to everything, although this one’s quite related, and it’s more fun. So, what’s the best language fail you’ve ever come across?

[00:27:59] Tamara Tirják: The best language fail is the one that doesn’t happen.

[00:28:03] Jason Hemingway: Okay. Yeah.

[00:28:05] Tamara Tirják: The second best is the one that you can learn from and use to improve your processes in the future.

[00:28:12] Jason Hemingway: Yeah, like you said before, it becomes a data point, right, if something’s not working, and as long as you know that.

[00:28:18] Tamara Tirják: And ideally committed by somebody else rather than yourself.

[00:28:23] Jason Hemingway: Yeah, there’s only so many times you can say it was just a data point. Yeah, so let’s think about, you know, from where you sit now, advising across the industry. Do you think business leaders, we talked about shadow localization kind of opening up the eyes of people, you know, that think about, you know, how they should be thinking about different languages and how they communicate with their customers and, you know, partners, etc. Are business leaders starting to understand the commercial value of getting language right? And, you know, where does that responsibility sit from your point of view?

[00:29:05] Tamara Tirják: I don’t believe in silos. Localization has to be put somewhere in the org chart, but in my opinion, that’s an administrative thing and not an operational thing. So, I believe in people from all the different fields coming together and make language-related decisions together. One of my favorite things have always been these discussions around language strategy. What languages should we support in our upcoming game? It looks like a simple chatbox exercise, but in fact, it’s anything but. So, we would come together, and we would look at this question from various different angles, with each of us bringing our unique expertise to the discussion. So, we’d look at the size of the gaming market in different countries as a starting point, the total addressable market. We would look at sales data from similar games, the language choice of players from our telemetry data, just as a baseline, and then they would start scrutinizing the actual content, figuring out which player groups it would resonate with, or where it might be problematic. And if it’s tied to a franchise, how popular is that franchise globally? At Frontier, I had the chance to work with the Jurassic World franchise, Formula One, Age of Sigmar, couple of different ones. And once these opportunities are identified, then a reality check also should be done. So, how many languages can we realistically budget for, considering the scale of the game? Obviously, smaller games can come with a longer list, while for bigger ones, you really need the level of confidence in each of them before we commit. Also, are there any technical barriers to overcome? The Middle East, for example, is becoming a very interesting gaming market worth targeting, but if the UI middleware is unable to support right-to-left languages and UI flipping.

[00:31:18] Interesting.

[00:31:18] Jason Hemingway: Yeah.

[00:31:19] Tamara Tirják: Exactly. Then you can have the most beautiful business case in the world. But if this would require months of work to do so, it’s not ready.

[00:31:32] It’s not gonna happen.

[00:31:33] No. No. No. Because there is the question of how this would impact progress on the game and what is the opportunity cost for allocating a UI specialist to work on supporting this rather than working on other parts of the games that might have a higher impact on the overall player community. So, that’s a very difficult call. And the well-informed and well-connected localization team should be aware of these technical constraints and bring these to the table because ultimately it will be their job to execute the language strategy.

[00:32:08] Yeah, no,

[00:32:08] Jason Hemingway: I love it. I love it.

[00:32:10] Tamara Tirják: So, their insights are absolutely invaluable to draw in realistic language strategy.

[00:32:17] Jason Hemingway: Yeah, I love that. I love

[00:32:18] Tamara Tirják: that.

[00:32:19] And I discussed this, my game localization strategy framework for global growth, in more detail in another recent article for

[00:32:28] Jason Hemingway: Nimdzi.

[00:32:29] Yeah. Well, we will make sure it’s downloadable and linked to in the write-up. And that strategy planning side of things, when you’ve done it, just to delve into it just very quickly, did you have all parts of the business coming? As you said, you know, you don’t believe in silos, but how easy was it to bring all those parts of the business together? I mean, in gaming, I guess it’s just a thing that people want to do. But am I right or am I wrong in saying that?

[00:32:55] Tamara Tirják: I don’t think it’s special to gaming. I think people in general love talking to people with different backgrounds. It’s just they bring in so many interesting points of view because commercial teams would look at the market data from one point of view, and then marketing would add their insights, and then localization comes with their cultural point of view. So, it’s just all this looking at the same problem from the different angles.

[00:33:21] Jason Hemingway: Good. Yeah, I think it’s good if you can get all those people in a room. That’s excellent. So, we’re going to talk now about a topic which is, you know, pervasive through every business, every vertical, every… and I feel like we have to talk about it. It’s not going away. It’s increasing and getting quicker, and it’s hard to keep up with most of the time. But we’re in this interesting moment, I think, where, you know, AI is pervasive everywhere, and it’s making it faster, particularly in content terms, faster to make more content at a huge scale. And teams are just doing it everywhere. But speed and volume aren’t the same as impact. What do you think, you know, what’s still missing from your point of view?

[00:34:08] Tamara Tirják: So, to clarify, the AI adoption in games, especially AA and AAA, is still fairly low compared to other industries, because for a lot of studios, game content is not just content, it’s art. So, what’s missing from that segment is the creative human vision, which is still highly valued in games, and AI-generated text, voice and visual content is undoubtedly useful and very functional in many ways and contexts. But games are not meant to be useful products per se. Right? They are creative, they are immersive and fun experiences crafted by humans for humans. At the moment, players are very harsh critics and very skeptical, if not outright rejecting any AI content. And games on the Steam store, PC games on Steam, for example, have to disclose whether any AI was used in the making of this game. And this can immediately trigger negative reviews, which we will then risk their future success. And to give you an example, last year’s absolute landslide success was a game called Clair Obscure Expedition 33. And it was the debut title of Sandfall Interactive. And it’s an absolutely stunning game with a rich narrative and a masterful localization by, need to give a shout out, the amazing team of Riotloc. They’ve done a beautiful job. So, as expected, the game won many awards, including the game of the year. However, they ended up being disqualified from that award because, in the early stages of the development, they used some AI placeholder assets for prototyping, which were then replaced by proper human content. But the award committee for Game of the Year has a zero-tolerance policy against AI, and it’s very strict. So, I don’t think we are ready to accept AI-generated books, music, film, and imagery as valid forms of art and entertainment just yet. And the same applies to games to a certain extent, especially AA and AAA. As I mentioned, mobile games do use AI. So, I definitely see use cases for AI to support production processes, for example, such as a more intelligent bug triage or streamlining operations, like making sense of live service data, but maybe not actually producing game assets apart from a few notable exceptions, such as using text-to-speech for computer voices, especially if it’s procedurally generated content in runtime, kind of like a sat nav, but there’s just no way to use a voice talent.

[00:37:33] Jason Hemingway: Yeah, I find it fascinating that there’s a sort of zero tolerance. I saw the gaming awards stuff. I’d read about that. They’d won an award. I didn’t realize it had been taken off them for some use of AI in early stages. That’s really interesting. And I think that that will normalize at some point you know, maybe not used in the creative way but in all those ways that you need it, you know, in the sort of just starting off with their automation as one kind of thing it’s sort of, like, well you’re kind of making it really difficult and incredibly more costly probably, but it’s a really interesting point that they’re not doing it at all. Let’s move on to a sort of another more lighter note. I find it sort of incredible that you invented a completely fictional language. Talk about making your life harder, but for the game Planet Zoo, you made a completely fictional language. And I was just interested in that and what it taught you about what makes communication authentic as more of it gets automated, like we’ve said. Where does the human sort of fit in that? But just tell us about that, you know, making up a language. That’s so interesting.

[00:38:44] Tamara Tirják: Yeah. So, first, let me start with a clarification. You is a plural you. So, it was, I was part of the team.

[00:38:51] Jason Hemingway: Yeah. Yeah. Yeah. Yeah.

[00:38:53] Tamara Tirják: Of course.

[00:38:53] The actual creator of the language is James Stant, the current dialogue manager at Frontier Developments. I did work a lot in organizing and managing this language and doing the translation.

[00:39:03] Yeah. Yeah. But

[00:39:04] Jason Hemingway: you

[00:39:05] Tamara Tirják: didn’t invent the language.

[00:39:06] No. I contributed to it. So, crafting a fictional language is an amazing example of human creativity. And if you look at some of the popular constructed languages, such as the Quenya spoken by Tolkien’s Elves or Star Trek’s Klingon.

[00:39:26] Yeah. Yeah

[00:39:27] It’s not just about creating a random language, but they are very intentional about the kind of society with their values and their ways of thinking that would shape this language for Planco, the language that we use in Planet Coaster and Planet Zoo. One of the underlying principles was that positive emotions should be described using happy-sounding words. The word for happy is ‘wippi’, which is a really happy word to say, while frustrations, so this helps the players get an idea of the sentiment of the park guests, visitors, even if they don’t understand the actual words that they say. In the creation of the language, a lot of personal stories and inspirations were used by the creator, James Stant, which makes this language very unique and authentic. And yes, very personal and full of Easter eggs and really fun stories. And of course, I know that it’s not impossible to prompt an LLM to create a language, but it would completely shift the focus to the outcome and ignore the process. If you talk to any conlanger, the creator of a constructed language, then they will explain to you how excited they are and how engaged they are with the actual creative process of crafting this language, of pondering about sociolinguistic questions, figuring out how this society would use grammar, how they would think about different concepts, how many words they would have for certain abstract ideas, so they create the language because they enjoy the journey, not because at the end of the process there is an outcome, which is what happens when the thinking process gets outsourced to AI.

[00:41:35] Jason Hemingway: I love it. That’s very interesting. So, the human is there to do the thinking. I still think ruling any AI out of helping in that process is probably a bad idea. But actually, the creativity is enhanced by it, perhaps, in the process. But we’re getting towards the end now. It’s been a fascinating conversation. Love the gaming angle. But let’s talk about your advice now to sort of business leaders who want to scale language globally, without losing that, what makes them distinctive? What would your advice be to those business leaders?

[00:42:10] Tamara Tirják: While looking for ways to deploy technology to increase efficiency, one might find that technology is more efficient than humans, individuals for certain tasks, but not more efficient than a group of humans with different backgrounds and approaches, feeding off each other’s ideas and creativity and knowledge. So, I actively look for ways to enable groups of humans to become multipliers of business value and a place for innovation. So, this tacit knowledge of the brand’s voice, and the company’s audience, and the market that they operate in is very hard to transfer to a tool or a vendor. And if ultimately every team uses the same tools, what will make them distinctive? So, figuring out how to carry this shared knowledge and collaboration framework forward as the business grows is the key challenge to solve.

[00:43:20] Jason Hemingway: Yeah, I think there’s something to do with that expertise that somebody might bring to that, or a group of humans, as you said. The context, how do you feed that, kind of, in? How do you get that going together? So, it’s interesting. And just moving slightly as we get towards the end. When you think about growth, what indicators would tell you that scaling efforts are delivering not just short-term reach or revenue, but that kind of lasting business impact? What are the indicators that you would look at?

[00:43:51] Tamara Tirják: The indicator I find most telling is the shape of the revenue curve across markets over time. What’s really interesting about the gaming markets, and I’m pretty certain that other B2C businesses share similar experiences in how the revenue curve looks across the different markets. But, for example, players from the U.S or Northern and Western Europe are more likely to pre-order games, pay the full price right after release, and Eastern Europe and Latin America come in bigger numbers later, once the game goes into sales. And in case of PC games, once the players have a configuration, a PC configuration that is powerful enough to play the game smoothly, because sometimes that also requires some time to catch up to the latest technology. And emerging markets such as Brazil, Argentina or Turkey will often also bring in the biggest percentage increase, growth in player numbers when a game is added to a subscription portfolio, for example, Xbox Game Pass. So, in my opinion, a solid long-term strategy understands that certain markets will take some time to activate and to show up in the revenue. But they are very important to ensure the long tail of a game sustained profitability, because if you look at AA, AAA games, but even mobile, I mean, how long has Angry Birds or Candy Crush Saga been out there?

[00:45:30] Jason Hemingway: Yeah.

[00:45:30] Tamara Tirják: Yeah. Decades now almost. So, they need a very long, stable tail.

[00:45:35] Jason Hemingway: Okay, so sort of final bit now. What’s the biggest opportunity that you think most global businesses still underestimate?

[00:45:43] Tamara Tirják: The biggest opportunity is one that’s sitting right there in most studios. Often completely untapped. It’s localization teams who are not just translators. We’re a network of people who understand the nuances of language, of culture and player behavior across multiple regions simultaneously. And in gaming, especially, we’re often gamers ourselves. So, we’re part of the same communities that the game is trying to reach, but doesn’t always know how to read and how to talk, too. So, that kind of embedded cultural intelligence is genuinely valuable and difficult to dig up in other ways. And most businesses never think to ask for it.

[00:46:33] Jason Hemingway: Interesting. Love it. Right. So, just a very quick-fire couple of last questions. One word to describe language intelligence?

[00:46:42] Tamara Tirják: Sorry, two words. Deep connection.

[00:46:46] Jason Hemingway: Deep connection. Yes. Okay, that’s good. Deep connection, better than just connection. I love it. So good. And then a book, a podcast or a newsletter that you’re loving right now.

[00:46:55] Tamara Tirják: I love Natalie Kelly’s Making Global Work newsletter on LinkedIn. She has such an amazing way of bringing clarity and new angle, such as when she was reframing work-life balance as sustained high performance, or more recently. She had a post about how AI actually highlights the value of language workers rather than diminish. And I think her views are such a valuable thing for anyone working in global business.

[00:47:31] Jason Hemingway: Yes, she is brilliant. And she has been a guest on the pod as well. So, yeah, she’s great. So, you’re in good company. And I love that. Tamara, that has been a great conversation. I love it. Fascinating. Games are very interesting. And I think, you know, you put across really well the business side of what you’re talking about, talking about revenue curves and strategic language and use of language. It’s really interesting. So, thanks very much for joining us today.

[00:47:59] Tamara Tirják: Thank you very much for having me.

[00:48:00] Jason Hemingway: Tamara, what a genuinely fascinating conversation. You’ve given business leaders a lot to think about, from what it costs a business when they’re misunderstood to what it takes to scale without losing trust, relevance, or identity. And thank you for bringing both the practitioner view and the gaming side of things view, and then the researcher’s eye to the conversation. And with that, that’s it for another episode of In Other Words, a podcast from Phrase. I’ve been your host, Jason Hemingway, and a big thank you to Tamara Tirjak for joining us today. If this episode made you think about what it really takes to connect with customers across markets, we’d love to continue that conversation. Head to phrase.com to find out more. For more conversations on leadership, growth and global strategy, subscribe to In Other Words on Spotify, Apple Podcasts or indeed your favorite podcast platform. Thanks again for listening, and I’ll see you again next time.