| Renato |
I’m Renato Beninatto. |
| Michael |
And I’m Michael Stevens. And today on Globally Speaking, we’re talking about games. |
| Renato |
Yes. And big people games. |
| Michael |
Big people games and a little bit about kid games. Do you play any games, Renato? |
| Renato |
You know, I’m not. I’m the father of a gamer, but I’m a typical old guy. I play Words with Friends. I play Solitaire. I don’t have much patience for…I skipped that generation. I thing I stopped at Frogger back in Atari. |
| Michael |
The Atari. Well, today it’s a really insightful conversation about games, about localization. And I think you’ll really enjoy it. |
| Renato |
So let our guest introduce himself. |
| Andy |
Hi, I’m Andy Johnson, I’ve been working in video games and web since 1991. I have more grey hairs than I’d like to admit. I’m an avid gamer and I love playing games; and I like to take that passion and try to convert that into localized content for non-English players. So, what I’ve done in the past with localization and on the data side, it’s very interesting to look at the number of players that come in from any given country and also the language that they are using. Before, when you try and justify a localization budget it was pretty much “what languages are we going to do?” and everyone shot from the hip and just said EFIGS because that was the norm, everyone did EFIGS. But, now, we can with the data that comes in you can analyze that data, look at it, and really see where players are coming in from both country and language and you can see their behavior. The great thing about that is on a game called Battle Nations I looked at the data that was coming in and saw that players playing in Russian really liked the game. It was a larger than normal number, and they were all playing in English. So, what I could do from that was look at, and start burrowing down into the cost of how much the word count, the asset count, for the actual game, how much it would cost to load, how much it would cost to loc-QA and then really build a business case around that; look at how long, if the player base increased by X number, X percent, then how much would that increase the revenue. I did that for three or four different options, possibilities, something like 25%, 50%, 100% and then crazy if it doubled or quadrupled. Then, picked one of those and just said we think that this is the most likely given experience, most likely growth pattern for players if we localize to Russian, presented a business case and said we think it’s going to be profitable within X number of months. |
| Michael |
Was it? |
| Andy |
According to the data scientist who ran the numbers it was spot-on. It was profitable within three months and after that point it was paying for itself with updates that went in. So there are two fascinating things about this for me. Battle Nations is a free-to-play game. So, by the time we localized it and this was through a company called Z2, which was later acquired by King, which was then acquired by Activision, crazy. So, the fascinating thing is that when we localized, at the point when we localized Battle Nations, the game content, it was very verbose, it was lots of text, it was very comedy driven, very humor, very tricky to localize. By the time we localized it, the word count was higher than the book, The Hobbit. This was not a small feat. It wasn’t a Words with Friends or a Candy Crush, this was kind of an epic console-sized game that already had had a two-year lifespan, I think at that time, and all the content updates and unit updates had a massive community following, and I managed to build a business case around that and still showed it would be profitable by looking at the data. That was purely looking at data. |
| Michael |
Is it normal for a game that is two years away from launch to be able to recoup revenue like that. |
| Andy |
This was two years after launch. We are in a new place with games. We used to be in a place on console before free-to-play, before always-on devices where games were fire and forget, where you would have this massive spike at the end of the development cycle where everything came in and was crammed in together and localization somehow was done at the same time when all the content was changing and being polished in the last, probably really, three months is when you do the real localization on like a two-year console game but, realistically, all the real work is done in the last month and a half. Then, you would submit it and the game would go in and then it would go to VABC or wherever they do the disc manufacturing, it would be released, you would see it in the stores. And that would be it. And then you’re on to the next one. But, now, with mobile and always-on, always-connected gaming devices, the ecosystem has kind of changed where we’re not looking, companies are not looking at releasing a game and then following up with the game and trying to keep – I don’t like to use the word users, I use the word players – players in their games. So, a player would go from – take EA as an example – from one EA game in the old style to the next EA game, to the next EA game. So, they would chew through content, through the game, it could be an eight- or ten-hour game, and then by the time they’d finished that there would be another release that, hopefully, would capture them when they have that money in their pocket to purchase the next game. So, now, we’re in a very different place where we are in a much more software as a service with you have your initial release and then you have incremental updates. Those incremental updates go on as long as the product, the game, is sticking with players and they like it. |
| Michael |
Yes, and then they can continue to even develop in that same game. |
| Andy |
Yes. |
| Michael |
Something like Destiny that rolls out different adventures on a regular basis and so people never need to leave, ever. Why play anything else? |
| Andy |
Even small mobile games are a lot more like the MMO model where World of Warcraft has been out over 10 years. So, that model is now in things as small as the mobile game that you get where you are playing and then there will be a new content update or a holiday pack or there will be some new skins that come through. Localization has to… it has much smaller, tighter cycles or loops to constantly release. So, it’s gone from this project management style of we do this as much prep work as we can and we have this crazy busy period and then done; to now we do as much prep work as we can, we have this crazy busy period, try and release the best quality that we can and then incrementally update and then as content updates come out, localize and release those. |
| Renato |
I want to go back. I’m fascinated by this concept of using data to drive decisions. You were still building business cases to define what languages you’re going to do and what you are going to do. What other parameters did you take into consideration? Just usage? I remember having a conversation with a gaming company when they said that they realized that any game they released they would have 20 million users in Indonesia overnight. But, they would get 10,000 in Norway and they would get a lot more revenue from Norway than they would get from Indonesia, so they localized into Norwegian and not into Indonesian. |
| Andy |
Yeah. So, they will look at the lifetime value potential of the player, and they’ll say “this cohort, this group of players, are very valuable”. Like, it’s a Facebook group that on a certain day they’re in this certain Facebook grouping; they are worth $10. The cost to acquire each user is $2, therefore the simple math is 10-2; they’re going to get eight bucks back. So, they’re always looking for that positive ROI, return on investment, on players. Everything to do with free-to-play is ROI-based, or should be. But some companies take that too far and are very business and not so much game, and that’s where I like to look at localization and preach that, evangelize that there are players at the end here and it has to be a good experience. That’s one side of it, the quality of the player. And that’s evident in places like Japan versus China. So, if you look at China, China has a massive, just simply massive amount of players, but their average spend is significantly less than players in Japan. So, where there are a lot less players that you can acquire, and the cost to acquire in Japan is much greater than in China, the output of that is that Japanese players, individually, are going to be worth more. But, then, in China you’re looking at the Walmart model of you’re going to get a lot more, but your return’s going to be less, but because of the size of that group of players that are going to come in that’s worth the investment. |
| Michael |
So, you’ve talked about two uses of big data that I think are applicable whether you’re in gaming or not. The first was instead of relying on just the standard GDP model for markets to move in, maybe taking advantage of some things that would not be the most obvious like “hey, one of our primary first languages should be Russian”. The second one you talked about is using it to segment users and be able to create an ROI model on there. What about the creative side of games, how is big data being used on that? |
| Andy |
On the creative side, what tends to happen is they will look at the return from how players are working through a game, how they’re playing a game and then they will adjust the game design accordingly. So, in old-style games it would be a designer would come up with an idea, a designer or design team. Then they would say “this is the game we’re going to make.” They were the experts. But, now, the source of truth is how players work through the game, how players play the game and what their interactions are. And you can look at that and see and say, “well, our assumptions were not correct, players are actually branching off and going in this direction in the game” where we didn’t previously think that was something they’d be interested in. So, now we can change gears and look at creating more content, more features around this area of the game. |
| Renato |
Let’s talk a little bit more about the other trend, the virtual reality and how you think that affects localization. Was that something that you had the chance to work on? |
| Andy |
I have not worked on localizing virtual reality content, or augmented content, or augmented reality content, yet. I’m sure that time is coming, but I have worked on, I’ve seen development cycles and console cycles change over the years, and I worked on 2D games back on Sega Genesis, the Megadrive and then when everything when to 3D, I worked on 3D, and then when everything went to online, I worked online and then when it went to free-to-play. So, I’ve worked through these changes in the industry. I don’t feel that other than the display and interaction because there’s going to be, obviously, differences in the way that you interact with the environment and with text and voice and subtitling and how it’s going to display, I don’t think the actual content is going to change. So, localizing from a content standpoint hasn’t really changed over the years. The quality and the volume, the drops that come in, the volume of drops has changed. So, I think that if we get to a point where we’re all in a Matrix or something similar, Lawnmower Man or something like that, and games are… or Ready Player One where players are in a game for, like, entire days at a time, then the amount of content that’s going to need to be in there is going to be of the size of the World of Warcraft, or even larger. |
| Renato |
One of the things I find interesting is that Japan is very advanced in gaming and game development, and from things that I track there is a lot of interest of having Japanese games localized into English and Western languages, and it seems there is more interest in getting that direction done than from English into Japanese because I think that comes standard. It’s just a speculation, things that I notice in forums and things like that, is there a reason for that? |
| Michael |
If you look at Twitter, that’s about half the comments around translation and localization, it’s about Japanese gaming. |
| Andy |
So, I don’t know if Japanese gaming is the best because I’ve seen some Japanese games that would not work over here and probably don’t do too well in Japan, but they have a very niche player-base. I do know that there’s also China and Korea, and they’re all very unique markets. Whereas, the West is more or less it’s one market so we can create one game for all of the West, meaning US, Europe, South America. I know that there are some companies that are taking games from China and adapting them and bringing them across, but that’s more of acculturalization than just a straight localization. You can’t just take… in many games you can’t just take the content and then localize the content and then bring it across; you have to look at the game holistically and then you start getting into mini-development cycles where you have to update and change assets; something that may be acceptable in Japan, some of their… what’s acceptable over there may not be acceptable in the US, for instance. Some of that content may have to be adapted and changed. So, then, you’re looking at not just having your localizers loc-QA and then implementing the content and then releasing. You’re looking at having developers and artists, and I know that some companies are doing that already, but the tricky thing is finding the right game. There are a lot of people on Twitter and Reddit who love games and want this particular game brought over from Japan to the US, and they really want to play it in English, but going back to Battle Nations, one of the very interesting things that I’ve found after localizing that was that we had this really passionate Russian community, and we released the Russian version, and these guys loved the game and they knew the game, and they’d been screaming for Russian for ages, and the unexpected output of that was the day after we released I came in and I had a Russian player that I’d been talking to and he said “people hate it”. Why! My entire soul dropped. I’d worked on this project, and I’d done this thing that the players had said they wanted and the data said it was correct and my math said it was profitable and the first thing this guy said was people hate it. And the problem was they had fallen in love with the characterization of the English and got used to that and even though I think we did a really good job on the localization, we really took time to hone the comedy and make sure it fit, when it went across to Russian that change was so abrasive that the players who had been playing for that period of time didn’t like that switch. It was such a ripping off the Band-Aid, ripping off the plaster moment. |
| Renato |
That brings me into this topic of who does what. What is the difference between a good supplier of gaming localization and one that is not that good. How do you select a good partner to localize games? |
| Andy |
That’s a really good question. Do you have an hour?! |
| Renato |
We do! |
| Andy |
So, for me, I think it’s very important to say that I don’t come from a traditional localization background. I come from a development background. So, I’m very process, quality, creativity. I always say my second language is development, so I talk to developers on their level; I don’t necessarily talk in French to a French localizer. I think that’s very important to distinguish. So, my selection process may be very different from someone who comes from a translation background. And if I put this out there then vendors will come to me after listening to this and saying “we can do this; we can do that exactly how you said” which will be quite amusing, I’m sure! I can see that happening now. My selection process normally goes honestly on something you touched on in a previous podcast which was content. If a vendor is willing to give something away for free, some piece of information, a walk-through guide, they have a blog, they have something that I can look at that’s substantial that I can quantify and say “these guys definitely know what they’re talking about”. Or, at least, “this piece of content looks like they know what they’re talking about” then I’m much more willing to go talk to them. I tend to stick with vendors that I have worked with previously; trust is a very big thing with me because when push comes to shove you can do all the tests you want, you can have the perfect localization vendor but when you’re up against it and you need 10,000 localized in two days and it needs to be creatively localized and you need it back spot on, perfect, then you need to make sure that vendor can do that work. |
