In Episode 2, Vasco talks with Silvia Oviedo Lopez, Pinterest’s Head of International Product & Operations.

Localization is optional, but internationalization is not. The moment you start writing, you need to be thinking about internationalizing your code.

Episode 2 Highlights

Part 1

  • Silvia’s path to the world of translation and interpretation, and how it prepared her for localization for startups
  • How to think about internationalizing your startup
  • The ROI of localization and Silvia’s “one metric” for localization
  • Why all languages are not created equally
  • What makes Pinterest magical to users all over the world

Part 2

  • The agile localization workflows and processes that Pinterest employs
  • Silvia’s opinions on user feedback
  • A discussion of the value of machine translation
  • How Silvia believes our minds are becoming more like search engines
  • Why Japan and Korea are such tough markets to localize within

If what you want is growth, then what you need to think about is internationalization.

Read the full Episode 2 transcript below.


Episode 2 Transcript

Part 1

Vasco (intro): Hi guys. Welcome back to Reach with Unbabel. I’m Vasco Pedro and this week we’ll be talking with Silvia Oviedo Lopez, the head of internationalization of Pinterest. For this episode we’re gonna split the conversation into two parts so that it’s easier to listen to. The first part is going to cover Silvia’s background in Translation and Interpretation, the translation process that Pinterest employs, the question of when a startup or business should internationalize, and how does she think about it. And finally, Silvia closes this episode with her opinion on what makes Pinterest magical. Let’s get into the conversation. Without further ado, Silvia Oviado Lopez, Pinterest, Head of Internationalization.

Vasco: Hi today we have with us Silvia Oviedo Lopez. She leads internationalization at Pinterest and she focuses on growing Pinterest’s international footprints. And since the beginning of her career, she has managed communities, content localization of international companies with Yahoo, eBay and Pinterest now. And for a while you also ran your own internationalization and blogging and search engine optimization company? And you studied Translation and Interpreting at University of Complutense of Madrid and Strategic Decision and Risk Management at Stanford. You also speak 8 languages and you’re also a poet.

Silvia: Kind of. Yeah I try. I can get by with like 5 and speak 3, 4 somewhat decently.

Vasco: That’s pretty awesome. It seems like language is always a very important part of your life.

Silvia: It’s funny because I wasn’t raised in a bilingual household. Until I was twelve all I can speak was Spanish. But ever since I took my first English lesson that was kinda like the world opened up for me. And if I look back there was always an interest in language itself aside from second languages or third languages. As a kid I was very obsessed with words and language and the use of language. So, I guess, that it kind of was a natural inclination for me to end up in the language world.

Vasco: Yeah. Why Translation? I mean you always knew that … it seems so you write poems also that it’s something that you do for a long time?

Silvia: I studied after college actually. I’ve always written and that was when it kinda got serious. I wanted to be an engineer, initially. That was my plan. I’ve always want to become a computer scientist. I wanted work with computers, to build stuff, and I did a bunch of that while in high school but I was extremely interested in both languages and history. So, one month before applying to … for college I faced this decision. Do I want to become an engineer, do I want to study linguistics or do I want to go for history which was also pretty passionate about. And in the end, the decision came easily. Translation was a fairly new degree in Spain at that time. It didn’t exist before. I think at the time there were only five or six schools that offered it. And for me it was a door to the world.

Vasco: Did you say translation specifically in Spanish to English or?

Silvia: I was doing English, German and Russian, the languages that I was studying. And a lot of the focus was around the theory and practice of translation and interpreting. After my second year my goal was to become a UN interpreter. I don’t know it may happen someday.

Vasco: Is that something you’d like to do?

Silvia: Yeah definitely. it’s something that my dream and would definitely love to visit an interpreter’s booth maybe in the UN.

Vasco: Must be a very stressful job.

Silvia: Definitely. But I was trained to be an interpreter. That was my plan when I finished college. In Spain, at least last two years of your translation studies, you can … translation and interpreting studies you can specialize…

Vasco: Oh, because it’s translation and interpreting studies?

Silvia: Yeah. So you can kinda focus on one or the other. I did both but with a very strong focus on interpreting and I think that it actually shaped my brain for translation or localization at fast speeds. Like localization for startups. You kind of need to change your mindset when … that’s something that I alway say and it may sound unpopular but the translation industry I think needs to evolve a bit faster than it’s currently doing.

Vasco: That’s a great topic because that was one of the questions that I had. Your thought in the field as a very progressive thinking person in localization and internationalization. Recently we were at Localization World, and I mean there was a panel you were in and it was fascinating because compared to the general field of localization, you’re kind of like way out there an kinda you know, pushing the boundaries forward. And what’s your position in terms of where the field is now, what the companies are doing, and where you think it’s headed? You know, because I feel that one of the things we’re seeing is exactly that mindset, as you identified, the interpretation of fast moving kind of adapting to content where it needs to be versus stagnating around the content, and thinking of this as a craft and we need hours and hours to produce the perfect content.

Silvia: I think there’s layers. When you think about localization in the online industry and for better or worse everything is moving to the online industry. Everything’s available on every device at any time and that has shaped the way you develop products. One of my key priorities when thinking about how do I want to run localization was what does our development process look like? Is it agile? Are we shipping things every day? And the answer is yes we’re shipping things every day so I cannot be the gatekeeper and say ok this will be waiting a month from now and when it’s ready at the same time in all the languages that we offer. And that has always been a personal frustration of mine that localization and translation is always perceived as the gatekeeper. It doesn’t necessarily need to be that way. I feel that this fast-paced development has given us a tool or has equipped us with the right mindset to start pushing the limits as opposed to saying we okay, we need to come up with this perfect process where it will provide a final product and we won’t be touched. Luckily right now, fixing a bug is not a 4 week process. A localization bug can be fixed in a couple of hours.

Vasco: Did you find that … so when you got to Pinterest for example, did you find that … that was because the development process was already in place such as you were pushing things everyday or were you thinking about that even before the developmment process was like that? So like, my point being, if you were a smaller company, I see this a lot in startups nowadays is, even though it seems like it should be easier and easier for a company to be able to internationalize earlier, there’s a lot of startups that delay it. Especially that first initial effort to for example localize a website into one language. It’s always the hardest part. Once that happens it’s kinda easier to work to go two languages then it’s a nightmare to manage more than a few languages. But it seems like it’s still happening too late. I wonder if part of that is because the deployment process itself, the pushing you code, new features, it’s a little bit messier than in a company like Pinterest where things are probably a little more organized and kind of more scheduled.

Silvia: I feel that … going back to the … should companies or startups start thinking about going international earlier, the answer is definitely yes. I would say that localization is optional but internationalization is not. So the moment you start writing code you need to start thinking about internationalizing your code. Making sure that you’re following best practices, that things would be readily available in the future and there’s tons of documentation out there to make things right and eventually you can turn on the switch at any point and start localizing and start shipping that first language and that should be a significantly easier experience if you didn’t think about internationalizing your code. Besides then it’s going to mean that you need to rewrite everything and that’s something that no one wants to do.

Vasco: Yeah. You know what … it’s funny because I think every startup you talk to, for example, knows this right? If you ask them like shouldn’t you know get your code ready or something? “Oh sure, of course, it is very important.” But somehow most people when doing this for the first time they’ll realize that you know, even when they thought they did a good job that there’s a lot of strings missing that there’s a lot of things weren’t ready.

Silvia: I think that’s going to happen that’s where pseudo localization or even machine translation can help a ton because you can create dummy versions and that’s how … I start my process. Basically, if we need to translate something new or localize a new site up, anything, is … first thing you need to do is run it through pseudo localization and see what your missing. That’s the first place where you do a round of QA. To say okay where are the gaps. And these things can be fixed, for me, easily. If you got them right early on your localization bugs are not blockers. It’s really strange that you’ll be at a point where internationalization bugs will be stopping your development.

Vasco: That’s a good point. For example, what I’ve noticed and going through it with our own website and we do translation so and it should be easier for us and still hard. And I think part of the thing that I realize it that the required information like for example what you just said you know … you should first run pseudo localization to understand where the gaps are that information is not collated in to one place. There’s no user guide for you know how to internationalize or localize your website there’s simply parts of it. You have this, well if you’re in Django generate the PO files, right? But then there’s no one that you need to figure that oh yeah doing internationalization on your URLs and making sure that they’re ready for different domains that’s a different part. And how do you SEO on that, that’s a different part. And oh by the way, before you even translate do pseudo localization that’s a different part. So I think there’s some resources missing that make it very easy like … Ok, you wanna localize your website? Here’s really the steps you need to do from end to finish. I think I find that, especially because every time the company starts thinking about localization, it’s never their core. you know. It’s never their focus. It’s something that they see as something as this can bring us growth. But it’s not something that the CEO or whoever it is that’s doing is stopping … is voluntarily stopping to do this. Like, well, gotta get this done and what’s the easiest and cheapest way of doing it?

Silvia: Yeah I think that’s … I don’t want to say a wrong mindset, or dangerous mindset, because what you want is growth, then you need to think about international. Because the US population or the English speaking population or the whatever, the German or Russian whatever language your app or website or service is originally in, is finite. Like, you’ll hit the ceiling pretty soon. But if you internationalize it then, adding extra languages it’s somewhat trivial. The hardest part is the beginning. Once you’re you can add more languages you can add more markets. It’s fairly easy to go on with that process but if you think about limiting it to the US or English or any language you’re kinda shooting yourself in the foot when it comes to growth. You really need to … you need to think big. You never know. You probably won’t start to localize into 20 languages in week two, you need to be ready to do it if it happens. You don’t want to wait five years.

Vasco: That’s a great segue. How do you know? Like … You’re starting a month, you look for different markets and you think ok maybe I’ve done my products in English or maybe I should do it in Chinese or Russian or something someplace where I think the market is big. But there’s ton of different languages, you know. How do you go about it? Like how do you pick. Pinterest now is in 31 languages that’s amazing. But how did it happen? Was it completely planned or was it more, well, we are gonna try different markets and then you see some metrics and you decide based on that?

Silvia: So, I think and this is not Pinterest specific but for every, any company is what do you care about? Like what is the core of your business and how does that resonate with different markets? Is it ads that you care about? Like, what’s the most interesting ad markets. I use something as specific service that can resonate specifically with a given demograpic and a given market. Are you in the apps downloads business and you have a social app that would be especially attractive to any market? Something that I find interesting these days is that aside from research that you would be doing is users are already telling you what are the most interesting market. Because you will see your product resonate more with a given demographic and a given market and that’s happening organically right now.

Vasco: Ok. So basically you’re saying you get your product in English for example, and there’s gonna be some traction in some demographics even if it’s in English and then you look at them and see well it seems like there’s already some interesting tractions here before you actually localize it to the local language then we can actually grow. You’re saying that it’s about the product fitting that market and then making it even better by presenting it in that one language.

Silvia: Yeah. Unless you have a specific interest in a given market. There’s a reason why you want to be in Europe. Then of course you want to focus there and if you have no possibility in working in Argentina or Latin America in general I wouldn’t prioritize Spanish. Because it’s, yeah of course you may have traction right now but if it’s not something you’re really interested in it’s not the smartest idea to put your effort there, while you could be focusing on France and Germany maybe because that’s where the core of your business will be 3 years from now. So, it’s kind of like making the decision of what do I care about? Or what does my business care about? And where will I find that product market fit?

Vasco: I think that it’s very interesting because one of the challenges that I feel that most people either in any from the local to international line is finding … Well assuming if you are at this stage and you have product market fit, is then how do you identify eaerly signs of that helping determine first internationalize into and once you’ve done that what was the ROI on that? So what was the … maybe sometimes it’s easy. For Pinterest, I imagine that you guys, once you deploy the language, you’ll see an uptick in growth. You’ll see someway like is there any specific metrics that you look at to say that yes this was worth it, let’s invest more here.

Silvia: So going back to the first part of the question, I think there’s two approaches you can be reactive or proactive. Going back, if you see that you’re getting a lot of the traction in a given country and if that’s a fit you should totally go for it. Or you can say, ok, I’m really interested in the Russian market. I’m going to put my efforts behind it and make sure that I’m getting all the benefits I can. Going back to how do you measure the ROI of localization. That’s the million dollar question. Everyone’s asking it. And it’s the same question for a lot of disciplines. What is the ROI of social media, what’s the … In the case of localization, there’s a lot of ways to look at it. First, you can can just look at what’s the growth in general. It’s going to tell you was this worth it. And I see more growth, which you’ll probably see because offering a product in the language of the end user is going to give you benefits in different ways. So, you will increase the activation of the user. The users will more likely to understand the product but understand what they can do with it, engage more with it. It’s also more discoverable. Say you got the benefit from the local SEO you get the added benefit of more awareness because you are speaking your customers’ language.

Vasco: Yeah. Of course.

Silvia: Yeah. Kind of basic but …

Vasco: Do you feel over time, I mean, you’ve been doing this you know at least at two different companies and more actually but, do you feel that it’s gets easier this conversation? Where nowadays if you go to whoever the decision maker of which language to go to, that it becomes more of a no-brainer? It’s like, yeah of course we need to be in multiple languages. It’s more of a matter of do we have resources deploy it rather than do we want to do it? Or do you feel that it’s still a struggle to explain this and make this point of ROI. Will this question of what’s the ROI become obsolete?

Silvia: Hopefully yes, it will become obsolete. One metric that I found has helped me a lot to say okay this is the ROI, because when no languages have the same ROI. Not all markets have the same ROI. So that’s when you need to back to what metrics do I care about or work metrics does my business care about. Because you can look at it from the growth standpoint. You can look at it from the revenue standpoint. What you want to map is your cost of localization per user to the metric you care about. So it’s like cost of localization per monthly active user in the market, cost of localization per pageview in that language. Whatever it is that you care about, those are two metrics that you should map and interestingly you’ll see that it’s a decreasing curve and in some cases will start tends to zero. Which is good. If it tends to zero, it means that there’s definitely an ROI there. And that’s also a metric that … for people who have an established business in many languages it can also help you make decisions. If I need to cut down a language because we are not getting any ROI on that language, then you can easily see what that language would be. We talked about scaling localization, increasing the international footprint but there’s also a point where you need to scale back. And doing that …

Vasco: Does that happen often? That’s very interesting. Usually we think of internationalization as always growing, as moving forward. Does that happen that you feel like, well, this was great but it’s not working as we thought, let’s rethink our strategy?

Silvia: I don’t think it happens very often but it definitely happens. And I think it’s a very smart move to see it coming and make the decision of, hey, we need to stop investing in this particular market or this particular language. Because at the end of the day it may not be the right moment, it may not be the right strategy at the time. I mean if can create a model that reduces your localization costs, there’s generally going to be an upside in offering a language even though that offering may be extremely limited. For example if you have your app but you don’t have anything else. You don’t have customer support, you don’t have any additional resources. Just a mobile app. Honestly, three thousand words is not a ton of money.

Vasco: Right and I’ve been seeing that in apps, it seems to be more of a no-brainer to do it earlier because there’s less amount of words and there’s more of an impact of having your app translated into a particular language enables you to have a better ranking in the market place for example in the app store of that country. If you don’t, users will complain. So there’s more of a direct impact that you can see immediately and there’s lower effort. So it seems like there’s more people that’s starting to localize their apps earlier and that leads to everything else. Which kinda leads to my next question which is when it’s not just about localizing your website of course there’s all this other collateral that you’re not incurring and it’s a little bit of a feature debt. Because it’s not only up there in the website but then there’s custom service, knowledge basis and you know, marking materials, and sales material so on and so forth. How do you manage all that? Because doing in 31 languages seems like quite a challenge.

Silvia: Yeah so that’s where a tiering system for languages. I always like to say that all languages are created equal but that’s not always true. I would love for that to be a reality. For us, my main priority is to offer our core product in every language. For every platform.

Vasco: When you say core product, do you mean like app or websites?

Silvia: Pinterest. Both. Like. Everything, every language, every platform at the same time. So doing simultaneous shipping is really a priority of mine. I always stick to that principle as much as we can. I mean or course there’s always features that may not be available internationally for various reasons but I would say that 95% of what we ship ships everywhere. And then there’s all the rest. So right now we do have countries where we have people on the ground that have different needs in terms of …

Vasco: Do you produce specific contents for those countries?

Silvia: Yes. We’ll produce specific content or translate specific content for them that we are not translating for other markets. But when it comes to customer support or resources for businesses we try to be as comprehensive as possible in our offering and translate them to every language. So make sure that we are helping international users get support in their language.

Vasco: Right! And in the case of Pinterest, I mean it’s really a love brand. As far as I can tell, your users love Pinterest. There’s this huge rise of goodwill that oh everything related to Pinterest is always amazing. And does that help? Does that mean that it’s even more important to be in their own language or do you feel that it actually enables Pinterest to not to care so much in a sense that your users will make more of an effort to understand Pinterest?

Silvia: Actually for us, putting Pinners first is one of our core values and it’s something we want to do and offering Pinterest in the user’s language is an acute priority for us as a company. Something that I always like to clarify is that we don’t translate pins themselves. We just translate the skeleton of the platform. The content is decided by all the users all over the world. But making sure that everyone has the content they need about or where they need it is acute priority to make the experience great.

Vasco: Would you like to? Let’s say that there was an easy way to translate all the pins but the ability to do that wasn’t an issue. Is that something that you think would be useful?

Silvia: I think that in the case of Pinterest the magic of it is that it’s … so we have over 30 billion pins curated and organized by users. It’s a mind-blowing number. And the magic is that the users all over the world have handpicked that content and pinned it to the system. So I think yeah of course that would be useful for some people but I think where the magic is we can have content from any country, any language, all curated by users.

Vasco (outro): That’s it for the first part of Episode 2. Thanks for listening. I’m your host Vasco Pedro and be sure to listen to part 2 to learn more about how Silvia thinks about feedback, how difficult it can be to localize a consumer product in Japan and her unique perspective on how content is changing people’s brains. See you next time on Reach.

Part 2

Vasco (intro): Hi, guys. Welcome back to Reach with Unbabel. I’m Vasco Pedro and this is the second part of our conversation with Silvia Oviedo Lopez, the Head of Internationalization at Pinterest. In this part of the conversation, we’re going to cover a variety of topics, from feedback to machine translation to the value of good copy. Silvia even discusses how she thinks our brains are being rewired to act more like search engines. Let’s get back into it. Here’s the second half of our conversation with Silvia Oviedo Lopez, Head of Internationalization at Pinterest.

Silvia: I employ all translation methods known to humanity in our process. Like me, I like the fact that we have a very lean process, we have a very lean team. In-house we’re one and a half, then we have three Project Managers, a couple agencies, freelancers, et cetera, that help with the load, but I like to keep things as lean as possible. We are using from raw machine translation to post-editing, translation and just …

Vasco: Everything.

Silvia: Corporate, translation plus review, translation plus review plus editor plus copywriting plus review again plus QA, so every content has its needs.

Vasco: And how do you determine? I mean, so for each of those content, basically it’s a different process. I mentioned that it’s both because of speed and quality, right? And so, how do you determine that quality? How do you know, well, like we need to be able to move this up to a different process given that it’s a big amount of content? I mean, you probably don’t have the ability to review it all, right? I mean, given all the language that you speak, that makes it easier, but still, if you have something in Chinese, then someone needs to figure out if that’s good enough or if we need a different process.

Silvia: Yeah, so we have a well-thought process when it comes to quality and we have all the review and QA mechanisms and feedback loops. I think that feedback is an essential part of translation. And, as a translator myself, I remember when I was in college, the professor would come with feedback like, “This is terrible” or … More nicely, but basically like, “This translation sucks, Silvia.” And my first reaction was to get pretty defensive. Feedback in translation’s like the best gift you can get. And when you create the right feedback loops, as translators, reviewers, QA people, local team members, users, it’s like you get this fantastic place where you’re shaping the language of a product and that really doesn’t compare to anything else. But going back to how do you decide where each content should go, like a lot of people will tell you, “Well, everything should be translated and reviewed then, re-reviewed then.” I think you have to go for your MVP for each type of content. Like, what is the least of amount of review that you can get away with? And that will vary.

Vasco: Right.

Silvia: Like for when we are localizing a video, it goes through analysts around soft review.

Vasco: Of course, yeah, that makes sense.

Silvia: Products, it’s like it’s constant. There’s thousands of pairs of eyes everyday, if not more.

Vasco: I imagine your users also let you know, right?

Silvia: Yeah.

Vasco: If something is not accurate.

Silvia: That’s actually a funny story, it happened to me once. A user wrote very persuaded because of a translation error. It was more of a linguistic or a terminology discussion but the user was really passionate. And that happened pre-Pinterest but it was a very good conversation with the translator and they ended up in a really good place. But I find it fantastic that people speak up and say, like, “Hey, I think your translation is terrible because of … blah blah blah. You should now feel terrible. It’s like …

Vasco: But that’s also an issue I feel with users, right? Is that people get really passionate about things they believe are not ideal. And there’s a degree of subjective criticism in translation. Should you have something a little more literal? Should you … ? Some people don’t like the tone or don’t like the way you say things and they get very vocal and how do you deal with that, right? I mean, having those discussions sometimes with users is, obviously, you want to put your Pinners first and you want to put your users first, but on the other hand, if you’re trying to, “Listen, no, no. I believe that this is the correct translation.” And mediating that sometimes can be a little hard.

Silvia: I think it’s always kind of hard to quantify language, but fortunately I had a few people in my career that taught me how to do that and if it’s not objective feedback, I won’t take it.

Vasco: Okay, so the “Oh, this is horrible, this sucks”?

Silvia: Yeah. I don’t buy that because I don’t speak Polish and you told me that sucks, you need to tell me more. You need to tell me how it makes you feel that the grammar is wrong because it doesn’t follow these rules. But saying like, “It’s bad,” that’s something that happens a lot with machine translation or like lower tiers of quality. Like, of course, if you have ten people looking at a translation trying to find exactly the right message, it’s going to be fantastic, of course. But you don’t need fantastic always, you sometimes can …

Vasco: Fill a knowledge base, for example.

Silvia: Yeah, you can …

Vasco: You might even use more information.

Silvia: Yeah. So something that a lot of people say is, “Machine translation is really bad.” I think machine translation is a fantastic human development.

Vasco: Right, yeah, yeah.

Silvia: Machine translation …

Vasco: Is getting better.

Silvia: It’s bad if you speak both languages. If you don’t speak a language, suddenly it becomes the most useful tool that you can use.

Vasco: That’s interesting.

Silvia: My mom speaks Spanish, that’s it. Like 50% of her Internet consumption comes thanks to machine translation.

Vasco: Wow! I guess it’s very easy, especially in the U.S., to forget that the Internet is majority English, 80% English, and if you don’t speak English, it’s complicated. Even in terms of simple things, like instructions or how to save a document or how to browse, go to a website, how to use Google. I see that also in my in-laws, like older generation Portugal, that’s very common.

Silvia: Yeah, and I think we try to be too good, which … And it would be fantastic if you could have every single piece of content on the Internet …

Vasco: Which was a little bit the promise of machine translation, right? I think sometimes the challenge in machine translation is that there was a very high promise. This idea that by now 2014, we would have like this perfect translation machine that would sound exactly like a human. And I agree with you, I think that the advances have been amazing, but somehow, like, I think people see in sci-fi movies, the Star Trek of like universal communicator and expect this ability, this fluency of just communication in any environment. And of course we’re not there yet. I mean, machine translation’s just not there yet. I think there’s fundamental issues with machine translation that we haven’t figured out that in a way have to do with the human brain itself and how we communicate and how we learn language and it’s going to be a while until we do that, right? But I think that’s part of the misexpectations of people and I think also because we’re really good at language. It’s a little bit like this uncanny valley in language where if it sounds kind of like a human but it’s not exactly, we get disturbed like, “Oh, it’s a machine and it sounds robotic,” which is kind of stupid because, as you said, it’s really useful for a lot of people in the world.

Silvia: Yeah, I totally agree with you but I think we don’t stop and admire the view. It’s like I can understand any text in any language by clicking a button.

Vasco: It’s true.

Silvia: And it happens in milliseconds.

Vasco: That’s true, it’s amazing.

Silvia: It’s seriously amazing and of course it’s not perfect but if I can just get the gist of it, I’m happy. It’s like how much time do you spare on a text on the Internet anyway? It’s kind of like read 10% the words and you’re happy with it. I think our brains are probably constructing a new way to process language that relies less on grammar and traditional structures and kind of like traditional text structure and more on key words and therefore our brains are starting to process information more and more like a search engine.

Vasco: That’s interesting. Can you elaborate a little bit on that? That’s a very interesting idea.

Silvia: Yeah. So, if you think about the amount of content you consume per day, like if you think about it in terms of words, it’s a crazy amount.

Vasco: Yeah.

Silvia: I don’t know how much you read every day, but I do read a lot of email, I read a lot of …

Vasco: News, blogs.

Silvia: News, Facebook updates and things like that. It’s like you just cannot retain information. Like, I don’t read, I scan. And I think that I’ve really found myself doing is, when I read a novel that I like, I read it on paper because that forces me to read in a linear way, whereas if I read on a screen, I’m just jumping like every five words.

Vasco: Interesting. So did you ever try reading a book or novel in iPad, for example?

Silvia: Yeah. I’m kind of picky with what and how I read, so if it’s something that I know I will enjoy, I will read it on paper because I want to force myself to pay attention as opposed to scan through a page, because otherwise I’ll read like three times as fast and not process all the information and not stop and stare at how the writer constructed the sentence.

Vasco: Right.

Silvia: It’s not about the message, it’s also about how the …

Vasco: The craft.

Silvia: Yeah, the craft of the writer. Whereas if I just want to gain information, I’ll just scan.

Vasco: Silvia, do you think that that’s because you also write and so you’re kind of also looking from a craftsman perspective? Like, Interesting, I wonder how, seriously, he constructed this.

Silvia: Probably. And I really like reading. I really like complex spelling and like give a name to the what you said, also how you say it. And when it comes to the Internet, there are two layers. Sometimes you care about what is being said, but also the how, [it’s being said is] really, really important.

Vasco: Silvia, do you usually read in Spanish, English? Doesn’t matter?

Silvia: No, it doesn’t matter.

Vasco: Is there a language you prefer to read it?

Silvia: Not really. It’s like I, now this …

Vasco: What’s your favorite book? Or a few of your favorite books?

Silvia: So, I’ll talk about writers. I have been reading a lot of Roberto Bolaño lately.

Vasco: Yeah.

Silvia: He’s a Latin American writer, fantastic. I also like Thomas Pynchon a lot in English. But these are writers that I can spend days and months with. It’s like I …

Vasco: Why do you like them? Why those writers?

Silvia: It’s really complex structures.

Vasco: So not like Da Vinci Code.

Silvia: No.

Vasco: More Umberto Eco.

Silvia: Yeah.

Vasco: Okay.

Silvia: I think more than that, even. It’s like they make you think a lot and it’s not only about the message, but also how they craft books. And I think fortunately a lot of that’s trickling to online services, like how you see things.

Vasco: Yeah, that’s interesting, right. So do you think that what we’re going to be seeing and so, even though people are supposedly reading less books, the amount of books that are being published has been increasing steady, right. I mean, there are more books being published now every year than there were a few years ago and it’s a tremendous amount of books. Now, do you think that this content creation’s going to start percolating to the kinds of content in Pinterest or in general websites, where you’re going to see the type of language that’s being used going up the value chain, let’s say, and you’re going to start seeing things that are better written, more well-constructed and moving away from the Tweet-type mentality of key words?

Silvia: Yeah. So I think that, for example, this is talking about Pinterest, we have I think it’s five or six people on our writing team. They take care of making sure that every single sentence on the service is a well-crafted message, that it has intention, that it’s the best way to say whatever we’re trying to say. And that’s something that I’ve seen proliferating in different services as to just like having that content person or team, if you’re big enough, because it does actually make a huge difference. Like, the two words you write on a button may completely change your conversion rate. So it’s like that’s something that growth hackers always talk about, like optimize a tagline, optimize the wording here and there and it’s totally true.

Vasco: It really makes a difference.

Silvia: Yeah, it does make a difference. And that’s something that I feel in the past was not very well taken care of, in general, in many companies, but now that writing part is taking more and more prominence, which is really nice and it also has that impact on the localization aspect of it. It’s like, “Okay, we have six people thinking about how to say this in English.”

Vasco: Right, so that you say it in a much better …

Silvia: Going back to the different levels of quality, it’s like that’s where we put a lot of thought, behind how do we localize this. Because we can all just say, “Okay, whatever translation you come up with, you need to put the same amount of thought into that translation.”

Vasco: That is really a challenge, right? I mean, because in order to go into languages, you might not have the staff available to do that, but then you need to remind some sort of resource, but the chances that the resource is going to have the same amount of care that you had in constructing the content are fairly low.

Silvia: Well, how I approach that problem is a) I trust the people that I hire. So it’s like people who become part of our pool of translators are tested thoroughly.

Vasco: So all of the people that write the final copy for different languages are translators that you interact with directly?

Silvia: Well, not directly but they have been tested. And we provide them with the thinking behind the sentence.

Vasco: Okay, the context, the semantic structure.

Silvia: It’s like anyone that works on our content has the rationale that the writers have followed.

Vasco: In this channel, Pinterest.

Silvia: Yeah. So it’s like basically, it’s like there are … there’s people with the same accent or they know how the writers think and they can apply that same logic. At the same time we have, as I have mentioned, several feedback loops and in some languages we’ll have more than in others. So for certain types of content, we’ll have more. But at the end of the day, it boils down to synthesizing your spendback and instead of breaking up a 20-page document, make it something that, as a translator, you can tape to your wall. You follow what I’m saying? I like this, “What would blah do?” Now, it’s kind of having that that … I need to put myself in the shoes of company A or to put myself in the shoes of writer B, how would I speak if I was that person.

Vasco: So it’s like being a director, you have to give direction enough to the translator.

Silvia: So you’re saying just like, “Think that you’re this person and then go for it.”

Vasco: What was the hardest language to be in, right? So there’s a lot of challenges, sometimes specific to specific markets and dealing with all of these issues that we’ve been talking about, in all these different languages, what was the marker in a language that you thought, This was unexpected. And why.

Silvia: I think that Japan and Korea are tough markets to get the language right because there’s a lot of social-cultural implications that are not present in other markets to that extent. It’s like you have all the a) nailed down to tone, formality or how do you want to speak to users and b) you have the best language, shape the world or thus, the world shapes the language. It’s like the chicken and egg problem. I think the deep structure of the language, the way ideas are conveyed is so different from the Western culture and the Western languages that crawling that over to Korea or Japan is extremely hard. Like all the …

Vasco: Maybe because it’s not conveying the same way. Like, it’s not just the structure of the language that is different, it’s the structure of the thought.

Silvia: How do you point to the world as a user. I remember one of the things that someone on the ground was telling me, it’s like in English, services speak to people, like your phone speaks to you. “Hey, Silvia, you’ve got a notification.” Like you cannot do that in those languages.

Vasco: How would you do that in Japanese?

Silvia: I cannot tell you. I was in Germany …

Vasco: How do you do that in Japan? Like, would you address … ? What would you address?

Silvia: Basically, you don’t animate the object.

Vasco: Okay.

Silvia: So, of course, it’s notifying you but you make the connection of, yeah, service is notifying me, sure, I’ll be biased, but in some cases you cannot just do that. So it’s these cultural things that make it harder to translate services and you have to put a lot of thought behind how you want to approach the problem because then there is tons of theories of translations and there’s conscious brand decisions of how you want to approach that. Do you want to be a foreign company? Because that may actually work well.

Vasco: I see.

Silvia: That may work in your favor. If it was …

Vasco: If you position yourself as a clearly foreign company, you have a foreign tone. But that might work because it’s a deluxe brand of like, “Oh … ”

Silvia: This German linguist called [?? 18:37] defends the foreignness of translation. Like translation should actually be something foreign and it should struck the reader as, I’m reading something that is not common for my culture, it comes from somewhere else.

Vasco: Right.

Silvia: It’s like that can be a branding decision, like we actually want to sound foreign and we want to … I don’t know if I personally condone that, but that’s definitely, if you want to take that approach, it may work. Like I’ve seen it work for a lot of French brands. Like the French Schick and a lot of make-up and cosmetics brands use that like, “We’re French and we’re going to use this as a marketing tactic.”

Vasco: Right. It feels like consumer brands in technology, for example, they try to really speak your language and they try to feel like, “Yeah, we’re just there with you, we’re in the market.” But I can totally see how fashion or certain times, like German engineering. I mean, the whole idea of German engineering, if it sounds German, so like, “Oh, there’s quality.” Right? Like there’s a certain quality stamp to it. That’s very interesting. Do you feel it’s the same in China?

Silvia: I haven’t worked with China or Chinese projects in Chinese for five years, I think. China is a very specific market but there’s a lot of protectiveness of their own language and services and then how these two map together. But at the same time, there’s a lot of interest in the Western culture and sometimes I feel that U.S. services translate better into Chinese, from what I have seen in the past …

Vasco: Than in Japanese.

Silvia: Than in Japanese. I think there’s more receptiveness, in certain areas, of course. You cannot generalize in China. But it feels there’s more of a connection and it probably has a lot to do with like all the history and the past of China in general, that there is more …

Vasco: Or maybe sort of the recent trend towards capitalism in China kind of puts it more in par with the way that the U.S., for example, tends to see things. And especially products, right? It’s not so much politics, but the products and the interaction between products and consumers. And it feels like in Japan there’s much more of a hierarchy that is explicit in the relationships you have with everything around you. And so that is more expressed through language, perhaps.

Silvia: And I think something else about Japan is that it has also been a technology hub. Sometimes when we think about technology innovation, it’s like lately you think about Silicon Valley, though they’ve been doing a ton of technology services.

Vasco: That’s true. I mean, in GAs that was …

Silvia: Innovation and …

Vasco: Japan was going to invade everything.

Silvia: I’m pretty sure that that has also shaped how they constructed their own object or technological object to human relationship through language.

Vasco: It’s very interesting because Japan clearly has the most advanced conceptualization of that interaction between humans and robots, for example. It’s much more pervasive and a lot of their culture in terms of anime movies, daily interactions, robots are much bigger in Japan than everywhere else and they tend to animate those relationships a lot. But, as you said, in objects like phones, like there’s clearly a distinction between something that is trying to be human in some way and something that is not trying to be human. And I think that gets reflected in huge parts of language, right? I mean, I had one semester of Japanese, I know very, very little …

Silvia: You know more than me then.

Vasco: I know just enough to say I’m a student of Carnegie Mellon and that was pretty much it. I hope my teacher isn’t listening. I learned a lot at the time, but I always got the feeling that the subtleties of how you interact are much more explicit in the language. For example, my teachers used to say that because there are females teaching and that I would learn how to speak like a woman and that I would need to spend time with male teachers, because otherwise it would be very noticeable that I was speaking some inappropriate way. And the words would be the same. Just the way, the emphasis in the tone we’d use. So I imagine that as a machine, you are going to have a hard time in Japan. So one last question. Are there any resources that you would recommend in companies or people thinking of how to approach this problem in internationalization that you think are interesting for people to … ? Either books that you think are useful or authors that you think are interesting reading or, I don’t know, translation manuals.

Silvia: I think, as you mentioned before, there’s not like a, “Here’s your localization guide. Like, here you go and now run.”

Vasco: You should write that. You know? Write that book.

Silvia: And that was actually my 2015 resolution.

Vasco: To write that book.

Silvia: Yeah. Write the book. And then all of you listeners, you can buy it and we can call the listeners.

Vasco: “Silvia’s Ultimate Guide to Localization”

Silvia: “How to Survive in the Localization World.” No, but what I would say is that all I have learned is through people in the industry … Something that I feel really, really lucky about is how open everyone is about their localization processes, things that work, things that don’t work. Of course there’s always staff that is arbitrary but everyone’s willing to talk about how they do it because, as I imagined before, I feel that we are kind of not as innovative as other industries and we should do more of that. Like we should open source more of our processes and what’s working, what’s not working. So what I would say is just feel free to reach out, pick people’s brains, we’re usually very willing to do it. And also I think conferences tend to be a good place to start. Sorry that I’m not being specific.

Vasco: No, no, that’s okay.

Silvia: It’s like I …

Vasco: No, but you’re right, there’s not the Ultimate Guide for Localization.

Silvia: Sorry if there’s an awesome book that I’m missing. I’m generalizing, so that’s …

Vasco: There will be an awesome book in 2015: Silvia’s Ultimate Guide to Surviving the Localization World.

Silvia: I would say that, for example, I read that resources on internationalization is all of the open source work that Mozilla has been doing. They have a fantastic corpus of internationalization resources, the best practices, et cetera, and you can see them kind of in progress. I would say that’s a great place. There’s a little on Github.

Vasco: The IMUG group are here in Silicon Valley. I think you can also use the C webcast, maybe that could be interesting. I just remembered that some of their talks on i18n are kind of … Sometimes they talk about very specific things. Sorry, now I’m giving suggestions.

Silvia: No, go for it. And then, yeah, let’s say, thinking about what companies you think are doing a big job in reaching out, asking for advice. I think everyone’s always willing to have a cup a coffee and talk about this part as it is. And the most important thing is trying and going for it and trying. The thing about localization is that it’s not open-heart surgery. So I never have the feeling that, “Oh, my gosh”, if we feel that someone’s going to die, fortunately.

Vasco: Unless you’re localizing open-heart surgery.

Silvia: Yeah, and that’s why I never did any medical translation. That’s actually a conscious decision. I don’t want to translate anything medical because I don’t want to …

Vasco: Be responsible for them.

Silvia: Yeah. But just try to … So, yeah, if you work for a company that makes open-heart surgery devices just listen to my advice.

Vasco: Well, thank you very much. Thank you for joining us and for sharing all these thoughts.

Silvia: Thank you for having me.

Vasco: Yeah, it’s been a pleasure.

Silvia: Thank you.

Vasco (outro): That’s it for the second part of Episode 2. Thanks for listening. I’m your host, Vasco Pedro, and this week’s episode was produced by Drake Ballew. Be sure to tune in next time, when we’ll be speaking with Anna Schlegel of NetApp. See you next time on Reach.


To learn more about Unbabel, visit us at Unbabel.com. To become an editor with Unbabel, sign up at our Editors page.