Localization Academy

Data Driven Decision Making – Zachary Haitkin From Netflix

•      Was this content helpful?

Was this content helpful?

How do you incorporate data-driven decision making in localization? Find out in this interview with Zachary Haitkin – the Localization Project Manager at Netflix.


Andrej Zito 

Zach. Welcome to the podcast.

Zachary Haitkin 

Thank you for having me.

Andrej Zito 

Where are you located? Let’s start with that.

Zachary Haitkin 

I’m currently in San Francisco, California. I’ve been here for the last 10 years, but I grew up in Southern California and in Los Angeles County. So I’ve spent my whole life in, in California grew up in Southern California, went to college way up in Northern California, almost to Oregon, in Humboldt County, and then now settled here in the Bay Area for the last decade.

Andrej Zito 

What did you study?

Zachary Haitkin 

I studied Spanish language. Okay. And international studies with an emphasis in Latin America. I’m not a native speaker of Spanish, but I am fluent Spanish. So…

Andrej Zito 

Fluent, fluent in tequila.

Zachary Haitkin 

Very fluent tequila. Too too fluent. But, yeah, it wasn’t the first thing I studied in college out of high school. I studied business because I don’t know business. Right? Money. Yeah. Money, right. Yeah. So I and the thing is, I didn’t care at all. And actually didn’t do very well in college at all. Had to leave a couple of times, because I wasn’t interested in what I was studying. Then finally, when I realized that I could study Spanish, it was something that I was good at and hadn’t really thought about doing it as a career or as a major. I went back to school, and what do you know, everything went smoothly the rest of the way. So yeah, I studied Spanish.

Andrej Zito 

So that sort of leads to localization. But I know that you didn’t start in localization right away, right?

Zachary Haitkin 

No, no, I’m out of college. I actually, in high school, and a little bit after high school I had done the only work experience for the most part I had was in rest in restaurants and catering. So I was actually I was able to use a decent amount of Spanish. In theback. There’s definitely quite a bit of Spanish being spoken. But yeah, after I graduated, I went to work at a bar as a bar manager, or a bartender worked my way up to bar manager and wasn’t really sure what, what to do. And then Lyft came along. And I started driving, because and I know if you’re in Vancouver, you just got Lyft not too long ago.

Andrej Zito 

Yeah, yeah. Yeah, I saw it a billboards everywhere. But let me pause right here. Because before we go into Lyft, why did you start as a bartender? Like, didn’t you think of like pursuing career and what you just graduated from?

Zachary Haitkin 

Yeah, you know, I, I thought about being a translator. But I didn’t really know the right people or the right avenues to go down. And to be honest, I wasn’t really super driven to like, start my career after after college, I really didn’t know what to do a little bit lost. Still trying to figure things out at the time. So, you know, I probably could have dug in and really done some work there. But, you know, I actually, I didn’t even think about localization. I didn’t. I know that sounds strange. But I didn’t know localization was like a thing. Like I understood, I understood that. Companies offered their services or their website or their content in other languages, but that to me just was was translation. I didn’t really understand what the localization industry was at the time. So I think I was more just trying to make some money trying to get on my feet. And then after after that, then try and figure things out. But yeah, no, it’s I was more just, for lack of a better word lazy. And wasn’t, wasn’t trying to get off the ground just yet. So I went back to what I knew, and that’s why I went into the the restaurant industry.

Andrej Zito 

Okay, so Lyft. He’s trying to driving. Yes. How do you go from a driver to a localization manager?

Zachary Haitkin 

Alright, buckle up. Okay. It’s no pun intended. So, while I was working as a as a bar manager, bartender bar manager, I started driving on the side. And this was very early. This was probably mid 2013, late 2013. And a friend of mine was telling me that, you know, you got to you got to get on this. It’s like a taxi, but it’s not. You get this giant pink fuzzy mustache, you put it on the front of your car, and you make $45 an hour picking people up and I’m like, What? Alright, I’ll give it a shot.

Andrej Zito 

Was it a startup at that time?

Zachary Haitkin 

It was. It was when I started driving in mid July 2013. There were probably at headquarters, I don’t know for sure. But I’d say maybe less than 100 people probably working in it just launched as a service in August of 2012. So it was less than a year old. They were only operating in half dozen markets in the US, so it was still very, very small. And so I would, I would, I finished my shift at the bar, I’d hop in the car, and I’d go out and drive and pick people up. So one night, I was driving late and picked up someone who was a Lyft employee, a support team manager at the time, he and I had a good conversation. He’s like, Hey, man, you seem like a normal person. Do you? What do you want? Do you want to come work for us Do you want to do we’re having a recruiting event for their support team at the time. And he invited me to that, recruiting events. And just like that, and now he didn’t offer me the job on the spot. But he offered me an invitation to the event where they were hiring people like crazy, because they were seeing some pretty amazing growth at the time. And so yeah, I interviewed a couple of times, and and got the job as an entry level support associate.

Andrej Zito 

What can you even remember? Like, what was your discussion when you were driving that guy? Like, what were you talking about?

Zachary Haitkin 

So he was my neighbor at the time. So I remember picking him up somewhere downtown at a bar and driving him home, which was only a few blocks from my house. And so that was kind of the first thing we started talking about. But it was at the time that it was much more how I put this, there was people were fist bumping, and it was very, like cool and avant garde to to be in this Lyft. It was like better than a taxi and there was like real people. And so, you know, I think the conversation was just like, what do you do? Like, what do you into? Like, what do you like, and, and so I was very much like your friend with the car, which was one of the Lyft slogans very early on. So I think it was one of those things where we just had had that conversation. I’m just like, here’s what I do. And then I assume, you know, what do you do and inevitably, Lyft came up and, and that’s why I think, you know, my, my skills as being in the restaurant industry, being a bartender, you know, you want to talk to the bartender, like you want to, you want to have that conversation. So I was pretty good conversationalist. And that’s how I was able to endear myself enough to be able to get the invitation to the event.

Andrej Zito 

I usually, I mean, not usually but, you know, like, you see the questions, like the first part is like, like, hey, let’s get practical. Let’s talk about localization. And then let’s talk about your personal life. But since we’re already talking about is like, I don’t know, I’ll figure out something in the post that I think or maybe I’ll just release it at once. But you know, I also had one friend who worked as a bartender, and then he worked at global mean, like a smaller agency, where he was working as a project manager on data collection. So he told me that that he is like, a lot of the people come to bar just to talk to the bartender. Yeah. So do you think like these these skills of communication and like making people feel comfortable and opening up to you? Is that something that you develop during the bartending? Or did you have like these good social skills before?

Zachary Haitkin  

I think I had some of it in the before, you know, coming into it, I definitely am somewhat outgoing. definitely like to have a conversation with people. But having so many conversations while working at the bar, I think I was I was I had a lot of practice, right? And if you think about it, what are what are two professions that you talk to, you talk to the bartender and you talk to the taxi driver, right. And so it was kind of a an interesting transition going from bar to taxi, more or less of the, you know, Lyft driver. So I think those skills just transferred right over. And you know, like, at the end of the day, you know, I ended up giving almost 4000 Lyft rides. By the time I finished, I don’t have a car anymore, because in San Francisco, you don’t really need a car. So I got rid of my car few years ago, but by the by the end of it, and I given 4000 rides, and so I picked up at least 4000 people actually a lot more than that, you know, there were two or three people they got in the car. So I definitely had plenty of time to practice those, those conversations skills.

Andrej Zito 

Okay, let’s tie it back to Lyft. So you went to the to the recruiting fair, you had the interviews, you were hired, but you were not hired as anything to do with localization. Right? Because at that time, they didn’t know anything about localization.

Zachary Haitkin 

That’s right. The way I like to put it Lyft was originally built for one language, one country, one currency, which had its advantages. Certainly a lot. Lyft to grow really quickly, because there wasn’t a lot of attention being paid to that, which was fine, you know, for the time being. So being in the United States, they definitely grew a lot faster for a significant period of time than Uber, here in the United States, then eventually, think either late 2017, or 2018, I can’t remember, Lyft launched in Toronto was the first Canadian city. And so that was a lot of the work that was done to solve the country and the currency issue to be able to operate in Canada.

Andrej Zito 

So when you mentioned the first Canadian city was the first city abroad?

Zachary Haitkin 

Yes, it was. Okay. Yeah, yeah. So, and Lyft still only operates in the US in Canada. At the moment, so, right. But yeah, so so when I got hired, it was just support associate answering phones and emails from riders and drivers, trying to resolve their issues, anywhere from you know, a driver got into an accident on the more serious side to, I didn’t get my $5 credit, give me free money. And I just saved sure here, here’s some, here’s some credit, you know, and move on. But I was able to incorporate some of my Spanish skills, because there was only me and one of the person on the team at the time, that had had had the skills in Spanish to be able to answer emails and phone calls in Spanish. And so we didn’t get too many messages, I’d say about 80 emails a week in Spanish, but you know, it was pretty much me and just one other person that were handling all of that. So I always tried to find ways to use the skills and my passion, which is, which is really Spanish language, to try and apply that to whatever the role I was I was currently at. Okay. So after spending a year on support, answering Spanish emails, there, I was approached by someone on the operations team to help manage markets, basically, this person wanted to move to another team, but they needed to find a replacement. And so they were managing Chicago, Miami, Atlanta. And again, this is very early at Lyft. When I joined as an employee, I was probably employee like 250, maybe between 200 250. So it’s still pretty early considering the company’s you know, 1000s of people now. So when I was on support, someone approached me to join operations to manage these, these markets remotely. And seeing that it was Miami again, that’s another opportunity to use Spanish language skills, considering it’s a heavily Spanish speaking market. So moved on operations for a year, help manage those markets, grow those markets, in terms of rides, get more drivers on the road, more people taking rides. And so after about a year of that, I got pulled onto the competitive intelligence team, because not only was I a Lyft driver, I was also an Uber driver. So people would ask me, Hey, what’s what’s going on with with Uber? What are they doing with this? What promotion? What coupon are they dropping, and so eventually, I was I was pulled onto the, onto that team to be able to grow a network to collect competitive intelligence and do analysis and pass it along to the team that was looking for free insight into different parts of the of the competitive platform, which, you know, at the time being on the in the US was just Uber.

Andrej Zito 

Okay, how it did lead to localization? Were you driving Uber while you were an employee of Lyft?

Zachary Haitkin 

I was really, both Oh, yeah. Yeah, I mean, there was nothing against it. And it was really valuable. You know, I don’t I don’t mind saying it now. Now that I’m, I’m done with it. But you know, at the time, there was nothing against it. So I was able to gather some some valuable information from from being a driver for both.

Andrej Zito 

So what was it like, you know, like, like, driving for the competition, like, Did you feel like they do a lot of things differently, better, worse?

Zachary Haitkin 

Honestly. And I’llsay it’s, with love for Lyft in my heart, I will always be part of the Lyft family, even though I don’t work there anymore. Uber did have a lot of parts of their app and product and experience that were superior. And I just think that I think that came from being around longer. Even though Lyft was first to do peer to peer, so actual people driving their own cars picking people up. Uber did have three years on them in terms of app development and platform developments because they had the professional drivers like the black car drivers. And then once Lyft came on the scene, they lost Stuber x, and that was the direct competitor. So, you know, I think there was a lot of things that Uber was better at. But I think that’s why they asked me to join their competitive team to find out what those things are, pass them along. So so we can learn from them. And Lyft could do it better. So yeah, no, it was it was an interesting experience. And then, of course, talking to writers as well, about, you know, what, why do you choose Uber? Like, why? Why do you? Why do you choose Uber instead of Lyft? and gathering that information, and then sending it back to, you know, the consumer insights team, stuff like that? So yeah, it was it was a valuable, a valuable way for, for me to contribute to the team that I was on at the time.

Andrej Zito 

When we had our intro call we talked about one of the topics that we wanted to talk about was data, and how to utilize data. So what you were doing at that time was basically bringing the data right about competition. So I’m wondering, did you see any action taking on the data that you brought in? Like, what was better about Uber versus Lyft? Or was it like a black hole, which sometimes the users feel like that like when they submit a feedback that you’d never gets them?

Zachary Haitkin 

Yeah, definitely not a black hole? I think, from an external perspective, it may feel like that. But I think the the feedback and pieces of information and intelligence that I was able to gather, were definitely taken into consideration. Now, were all of them implemented? No, because there’s certainly amount of prioritization that happens. And the idea that Lyft wants to do their own thing they want to, they want to forge their own trail, and that’s fine. I’m totally fine with that. Because at the end of the day, it was up to me to provide the information, it wasn’t up to me to actually build the product itself. But I will say one, one thing that was very actionable, and that I was able to have direct influence was collecting driver incentives. So part of my role on competitive intelligence was to build a network locally, so not just in San Francisco, but across all cities that that Lyft and Uber operated to collect driver incentives. So for example, in Chicago, there might be a if a driver gives 20 rides, they get a $25 bonus, something like that. So collecting that information from as many sources as possible, is synthesizing that data into saying, okay, we had 10 drivers that got this offer and 10 drivers that got that offer, and then passing that along to the local team, or the people that are making the decisions on what incentives Lyft is going to offer, knowing what the competition is doing, we can match or beat that. Depending on whether or not we need more drivers out there. So that was that was something that was I brought to the table and was absolutely able to take action on even so much as down to the specific hour, like knowing which hours incentives were available, and then having having lifts offer a similar or greater incentive, if we decided that’s the strategy we wanted to use at the time. So I’d say that was definitely a way that some competitive data we collected and then implementing it almost immediately.

Andrej Zito 

Okay, so there you are driving for the competition, giving the information to Lyft is the next step localization.

Zachary Haitkin 

In an indirect sense, yes. While while I was working on on the competitive team, almost the entire time I had spent at at Lyft, I had been advocating for a Spanish language experience ever since I was on support. People would a lot of times write in because they didn’t understand what the experience was. Now, there’s a lot of other languages I think, would have been helpful, but my strong suit is Spanish and in the United States, Spanish is the is the second most spoken language.

Andrej Zito 

Besides when you mentioned that the people didn’t understand the experience. Are you talking about like new customers, or maybe the drivers?

Zachary Haitkin 

I’m mainly talking about drivers because if you think about it from a passenger or rider perspective, you open up the Lyft app, you put where you’re going, you press a button, someone comes and picks you up and you really are only interacting with the app for a few minutes. on the driver side, sometimes they’re driving for 810 12 hours at a time. So it’s a much more in depth experience for a longer period of time that the driver is having with the app than the rider. So that’s I think there was more emphasis on on the driver experience. So yeah, there was a lot of times drivers wouldn’t understand, like I mentioned with incentives, you know, like a bonus or something, they get an email, it’d be in English, and they wouldn’t understand it, because they didn’t understand English as as, as much as they did, say Spanish. So almost the entire time I’ve been advocating for getting a Spanish language experience. And there were, it was a long process of talking to the right people. And because I started the company, so early, I was able to interact with people that were pretty high up like directors and VPS. And even the even the co founders, I got a good chance to interact with them. There was a shark tank type event that we had, where we each, like there was teams that pitched ideas, and that the idea that I pitched was Lyft Latino, and this was about six to eight months before we actually launched a Spanish language experience. But that’s what I pitched. And and we want our team one, because the executives realized that it was absolutely something that that needed to happen in order to facilitate growth at the company. So yeah, it was just I would say, for about five to six months, I was doing double work. Because once they decided to localize the app into Spanish, there was a huge effort that got kicked off that it touched almost every team at the company. And I think that’s that’s not really a secret. You know, if you’re going to go from one language, the hardest part is going from one to two languages, right? Then going from two to however many more is a lot easier. But the one to two was, what was the really hard one. So while I was working on the competitive team, on the side, I was doing extra work to help facilitate the creation of the localized experience in in in Spanish for lips. So when the time came to permanently put people on that team to maintain it. That’s when several people in the room when they were deciding, said, “Hey, this guy’s been doing it for the last six months, let’s give him a shot.” And yeah, that’s how I found myself. Finally, officially on the localization team. Finally, I know, after after four years, it would it had been I had been at the company for for a little over four years, when I finally was able to work officially in in localization. And so yeah, I joined as a project manager. So we had a program manager, and the Spanish language manager at the time in the beginning. And yeah, so it was really up to me to manage more internal stakeholders, and not so much with the vendors, translation vendors, or TMS vendors to start. But But definitely, if someone needed something translated, or they had questions about translation, I was really the the person that they would go to, to help answer their questions. When I when I first joined the localization team.

Andrej Zito 

Yes, I was so excited that we finally reached a localization point in your career. But I still want to go actually back because I have some questions here. You know, when you were talking about, like, the Spanish speaking users that they didn’t understand the app, did you have any data at a point like, or did Lyft have data about, like, how many people could be Spanish speaking?

Zachary Haitkin 

Yeah, so it was, it was pretty easy for us to tell, just based on device language. That’s how that’s how we could that we could tell whether or not the person would probably prefer Spanish. And, you know, it’s, it’s interesting, because I think there’s a little bit of a misconception about someone that maybe Can, can speak English pretty well, that they can understand written English without, without any issues. And so, you know, reading and comprehension is much different than than speaking. So we felt that if if someone’s phone was in Spanish, then they probably would have preferred that Spanish experience. So it was pretty easy for us to pull pull data, because we were already collecting a decent amount of data from the riders and the drivers phone. So we already knew what the percentage of users that were, that were using Lyft in English with the Spanish device.

Andrej Zito 

You mentioned that for you it was like a personal thing, because you were so good with Spanish, but were the Spanish speakers like on top of the list of using different language than English? Yes. In your data, right.

Zachary Haitkin 

Yeah. And I would I would say, it definitely varied by markets. different markets had different languages that were third or fourth or fifth. So yeah, the vast majority of of market In the US had Spanish as their second language behind English. But then you look at a city like Boston, where there’s a very large Portuguese speaking population that definitely popped up. Then you look at New York, you’ve got Chinese and Russian that had very significant portions of users that that had their, their device in those languages. So Spanish was by far number two behind English, but there were certainly some other languages behind Spanish that that had some significant percentage points of riders and drivers.

Andrej Zito 

If we can, by any chance, go back to the Shark Tank moment where you were pitching the idea, basically, the inception of localization at the Lyft, right? Yes. How did you convince them? Especially the managers, because like, yeah, you can say that, yeah, I understand Spanish, maybe my friends and drivers, their Spanish, and it would be nice if they understood the app in the language, but then the people with with the money or the managers, they can select? Okay, so what they’re already driving with us, like, how do you justify this? I would say sort of a big decision? Yeah.

Zachary Haitkin 

That’s a great question. So there were a handful of data points. And they were very rough. Because, you know, in the Shark Tank situation, it was almost like, I don’t want to call it a hackathon, because there was no coding involved. But it was very much get people in a room for a few hours, and then come up with something. So I don’t think we we weren’t expected to have, you know, really specific pieces of data to go on. It was a, it was more generalized. So we looked at the competition, I think was number one, Uber at the time was localized into 2530 languages. And it was very much preferred among non English speakers, because of the fact they offered a localized experience. So I think that was number one. Number two, I looked at pure population numbers in the United States. There was, at the time, there were 40,000,040 plus million Spanish speakers in the United States, which makes it the second largest Spanish speaking country in the world. Behind Mexico, Mexico is number one. And then there’s more Spanish speakers in the United States than there are in Spain, which is kind of crazy. Right? I know. Heavy. I know, right. And there was, there was another data point to looking at smartphone usage among Hispanic, the Hispanic population versus other other ethnic demographics, which actually skewed higher. So that that was another piece that we brought in that you’re more likely to be using a smartphone more frequently, if you are of Hispanic descent. Also, we looked at purchasing power. I mean, these were like Pew Research, public type, research numbers that we were pulling in, we looked at purchasing power. And we looked at the growth of the of the Hispanic population in the US, these were all data points that we, we came up with to pitch to the executives. And it was it was a powerhouse. I mean, the panel was the CMO, the CEO, the CIO, the creative director, I mean, that’s who was judging us. And so it was really, it was really a great opportunity to be able to be in front of those people and to have an audience. And then when we won, you know, I got a chance to have a 15 minute conversation about it with the with the, with the CEO of the company, and the and the president, there’s, there’s a lot of co founders that Lyft, but the two that are still there, john and Logan, you know, being able to have a conversation with them for 10 to 15 minutes about this. And then to get to know me, I was, you know, definitely one of the high points of my career.

Andrej Zito 

What I’m thinking right now is, you know, you mentioned What was interesting and surprising to me was that you mentioned that at the time, Uber was already localizing to 20 or 25 languages, you mentioned,

Zachary Haitkin 

I don’t know the exact number, but they I mean, they went international in 2011. I mean, they were they were already Paris, I think was there their first international market. So they I mean, before Lyft even launched, they were already International. I see they had a huge head start on their translation and localization program. And I think as far as brand awareness, you know, Lyft has definitely come around. But when you have people that, you know, maybe are immigrants or they’re not, they’re not as aware of other of other rideshares in the in the US. They’re gonna go if something’s in their language, they’re gonna go with that, and, and for a lot of a lot of languages that was over. So I think Uber has 40 languages now, at least on their website. So I think at the time it was it was it was quite a few a lot more than Lyft that’s for sure.

Andrej Zito 

Did you or the company Feel like the lack of localization is, let’s say hindering the growth of the company like it could be the key? Or was it more like making it a better experience for the people who are already using it?

Zachary Haitkin 

Well, I always felt that it was a growth lever. Absolutely. I, I firmly believe and continue to believe that having an experience that is able that is is able to be in that person’s language is going to be it’s going to be sticky. They’re going to keep coming back. And it’s going to get new people to join. Convincing other people of that. Not so easy. I think I think I think people did see it as something that was a nice to have maybe a little bit more than a nice to have just, I don’t know, it’s hard to explain. I think some people, yes, they saw it as a growth lever the way I did and other people were like, yes, but not to the extent that would be worth really moving forward in localizing into all of these languages. I think some people conflate international internationalization with localization. And whenever I started talking to a certain group of people, sometimes it would never be, well, we’re not going international. Why are we doing other languages? And, you know, I needed to remind them that we’re in the United States, there’s a significant population of people that speak other languages, millions of people. And so, you know, I could go, there’s a long list of data points that I tried to put together, or that I did put together to try and convince people to do quite a few languages, even putting dollars amount, you know, dollar amounts of on it, but at the end of the day, it just, it wasn’t, it wasn’t enough for them to prioritize it the way that I would have liked.

Andrej Zito 

How do you go from, from bartending and driving a car to presenting data points? Is it something that you like picked up? or How did you, I don’t know, get to the point where you knew, like, you need data to prove a point or to convince someone.

Zachary Haitkin 

That is definitely something I picked up guys, you know, I had never worked in tech. Before I worked at Lyft, I had worked at, I worked at an office as an office manager at a roofing company. Before I went to, to college, to finish my degree in Spanish, so I had, I had some experience working in an office. But as far as, as data and convincing people, I had to pick that up, that was definitely something that I learned almost completely at Lyft. One thing that I thought was, was really funny in the very beginning, you know, I joined and I was like, Oh, my God, like, I’m working at this company. It’s growing really fast. That’s pretty amazing. Like, you know, everyone must know how to do everything. And everyone must be just geniuses to be able to know. And the more I got to know people, and the more I started to understand, not to say everyone I worked with, for the most part, super smart, super talented, but I came to the realization of like, they are still also trying to figure it out. I’m trying to figure out some things they’re trying to figure out what they’re doing. And because they’re, you know, great at what they do, they figure it out. But at the same time, they don’t always have all the answers. And so once I figured that out, I think I felt okay with not knowing things, asking the questions, building the relationships with people, and then learning from them and asking them for for guidance and not feeling bad. That, that I didn’t know something. So yeah, as far as that learning how to convince people and that data is what drives decisions. That is definitely something that I had to learn on the fly.

Andrej Zito 

You know, you talk about learning and adjusting to things. So how did you learn localization? Because, you know, another thing that was surprising to me was that you were the guy who, let’s say, pitched the idea. And let’s say you convinced you and your team convinced the company to start exploring localization. But then you told me that they were like, outer like program managers or someone else. I thought it You were actually the one who was starting the whole thing. So they brought in some people with experience and then you joined the team and…

Zachary Haitkin 

Yes, yes. So I had been at the company a lot longer than than other people. So I think people knew me as someone who wanted to advocate for localization. But I have to credit the the program manager who came in who had a lot of external. His name’s Brian McConnell, really, really great guy, really knowledgeable, really knows his stuff about localization. And I credit where I am today because of the knowledge that he gave me. So even though I was the one that pitched that Shark Tank about having a localized experience, he was the one that did most of the work in the beginning to set it up higher TMS build that relationship, work with engineering to connect our code base to the TMS. Like, I could not, I didn’t know how to do things like that. And he did. And he taught me, you know, even though that wasn’t necessarily his, his role, his role was to build localization. But you know, I, I forced my way in, I continued to ask him and, and how to him and he was super nice about it, and really gave me the guidance that I needed to understand localization enough to be able to join the team is actually really funny. Going back to the Shark Tank, one thing I want to mention, one of the people was the it was the VP of rideshare, are the VP and he was one of the VPS of engineering was on that panel with the with the CEO and the CEO. And as I’m going through the presentation, you know, at the end, I’m like, okay, translate the app, and then do marketing, and then launch Spanish experience. And and in the question and answer, he’s like, “Well, do you think it’ll be difficult to translate the app? Do you think you’ll need engineering resources?” And I was like, “Nah, probably not that much. I think it’ll be fine. We can get like a vendor, and like, there’s a company, we could just hire to do that.” And he, like, he didn’t challenge me on it, he was just like, Okay. And I appreciate him not doing that. And, you know, after going through the process, and understanding that huge amount of effort that needed from the engineering standpoint, that I had no idea, as, you know, me being naive at the time, I just, it was, it’s a funny thing for me to look back on to, to know that I had to understand I had no idea what it actually took. But he was he was very nice about it. And at the same time, it wasn’t it wasn’t an engineering event, it was a bit more of an operation idea, type event. So I think it was, it was just, it was very nice of him not to challenge me on it. So anyways, just just the fun, a fun memory from that experience.

Andrej Zito 

Did the localization efforts start without you? Like, did they first hired the experienced people? And then you join the team? Or were you part of the team from the start?

Zachary Haitkin 

They hired other people. Before I officially joined the team, like I mentioned before, I was doing kind of double work. So I was unofficially working on the on the localization effort. So yeah, I was I was not technically I was not the first person to be on on the localization team, I joined close to when a lot of the work was finished. And they were preparing for launch was when I officially joined the team. But I had been working with them in an unofficial capacity for about six months

Andrej Zito 

Before that, when you joined the team, what was the state of localization at the time,

Zachary Haitkin 

Like when I officially joined the team, it was it was almost done. The translation pipeline had been built and was being tested. Most of the content across web and app had been had been translated into Spanish. So there was a lot of testing going on. We were probably two to three months out from launch when I officially joined the team. So a lot of the work that I had done unofficially was to kind of create workflows, and visualize workflows, on how things were going to work. So then what I when I joined the team, it was up to me to take those workflows that I had created and then implement them and maintain them. And you know, just very simple of like, an engineer checks in a new string. That string is then imported to the TMS. Then in a TMS, vendor number one does the translation. vendor number one does the editing vendor number two does the review. Then once the review it moves to published then the translation pipeline pulls it back into the code base, and it gets stored in a repository. And then when someone opens up the app with their phone in Spanish, they see the Spanish string instead of the English string, you know, creating and visualizing those workflows not actually doing the engineering work, but just showing how it works to other people at the company so that when they want to know how localization works, they have a visual guide to be able to do that. So. So that was a lot of the a lot of the work that I was doing before I joined. And when I actually joined, I took that and then and then ran with it and continued to show that to other people at the company, so that they could understand how how we were going to run things.

Andrej Zito 

You mentioned that you started as a project manager, right?

Zachary Haitkin 

Yes, yes. So I joined the team as a project manager, working under Brian, as the program manager. And then he there was also another Spanish language manager at the time, that that was overseeing the, the linguistic quality and working closely with vendors to make sure that the translations were right, as well as maintaining glossary, terms, editing glossary terms, and really finding our voice in Spanish. And making sure that was that was complete and inconsistent.

Andrej Zito 

Did you know anything about project management? Or was it another thing that you had to pick up on your own?

Zachary Haitkin 

So I had done a lot of it on the competitive team. Even though my, my, my I think my title was, God, I had this really cryptic title. I was like, strategy analyst or something or strategic, something, local insights, strategy specialists there. But I mean, I was I was a program and project manager, basically, I had the local competitive insights program that I was managing. And because I was the only person at the time, I was I was the program and the project manager. So So I took that experience from from the competitive team and was able to, to shift that over to localization.

Andrej Zito 

So you mentioned, if I’m not mistaking that you joined two months before the launch, right?

Zachary Haitkin 

Just about Yeah.

Andrej Zito 

To me, this is like, like, like a big thing? I would imagine. So what was like the atmosphere or the expectations from the company and from the team, like before the launch? And how did the launch? Go? And what did you or? Yes, okay, let’s start with those questions before I pull the ideas on.

Zachary Haitkin 

Yeah, I think, based on the way we we set up a test, like a holdout group, we were gonna launch to a percentage of of Spanish users, and hold out as percentage to be able to see if there was if there was a difference, right, like any test to be able to see if it was actually driving incrementality. Were our riders taking more rides? Were new users signing up more frequently? Were drivers driving more often? Were they earning more, were they getting higher ratings? And the answer was, was Yes. on all of those, you know, I can’t share specific numbers, but statistically significant numbers in several of those categories. That, that we saw a positive lift, pardon the pun in, in actual in the data. And so that was the expectation was that if we were going to, if we were going to spend the money and maintain a localized experience, we should be able to see some level of incrementality from our rider and driver population and and new user acquisition as well. So yeah, I think I think some of the key metrics that we saw increases where drivers were getting higher ratings, they were driving more frequently, new users were activating. So that’s that’s actually taking a ride. They were activating at a higher rate than the holdouts. Drivers were making more money, because they were driving more. So those are really some of the key metrics that we saw. When we launched a new language. What was even better is because I had that that data from Spanish. When I wanted to advocate for other languages farther down the line, I was able to use those numbers as a proxy to say, Okay, here’s the population of Spanish drivers, we saw this lift. Now if we did this for Portuguese, we’d expect to see the same lift. Here’s the dollar numbers that it means if we if we launched this language, so it was really important for for us to get that data in the very beginning to be able to have that in our pocket when we wanted to launch other languages.

Andrej Zito 

Did you also try to collect qualitative data like surveys from the users how they feel about it, or was it mostly quiet?

Zachary Haitkin 

It was mostly quiet. Mainly due to resourcing. I would have loved to have collected some some qualitative data. But the the, the consumer insights team, I think wasn’t really equipped or it wasn’t a priority for them to collect qualitative data around the experience. So we really did have to rely on on quantitative data there.

Andrej Zito 

You know, what I like when you mentioned that, when you launched, you already had the sort of plan of what kind of metrics you want to see impact on through the localization. Once you started, like one to one, once the localization program started to be more mature, did you add some more metrics into it? Or did you like handpick certain metrics? Among all those data, you know, like, you can collect a lot of data. But like for the localization program, what was what was the important data?

Zachary Haitkin 

I think we mainly stuck with the ones that we saw from the Spanish launch. It was tough, because the other languages that we ended up launching, like, right now Lyft is available in Spanish port, Brazilian Portuguese, and Canadian, French, it was it was difficult to get a large enough audience in those languages, to really get statistically significant data. With Spanish, it was easy, because there there were enough drivers and riders to be able to gather that data over a certain period of time. So for the other languages, at least in the United States, we just didn’t have enough enough of a sample size to be able to, to get a significant read on it. So we really relied on the same metrics for those for those languages. And just as opposed to coming up with new ones. So yeah, I, there was other. I tried to do some other comparisons as well. For example, I got my hands on a competitive data sets around different devices that have you looked at keyboard language on Android devices, primary, secondary, tertiary, keyboard languages. And then we could also see, it was all anonymized, like what apps they had installed. So you know, and there’s a lot of companies out there that do app, like app data analysis. So we looked at who has the Uber app and the Lyft app and the Uber driver versus Lyft. Driver, by by market, and you know, seeing what percentage of people based on keyboard language. And you could see, like, there was a lot of under indexing, of Lyft versus Uber and a lot of markets, even even for Spanish. So I think that was one thing I tried to, because, you know, as soon as we launched Spanish, I wanted to launch a bunch of other languages. So you know, I tried to try to make the case for that. Then looking at running Spanish language ads, seeing if the on Facebook, seeing if we could get Spanish users to activate at a higher rate or at a cheaper, cheaper acquisition cost. You know, I can’t really go into details on a numbers on that. But it was a fairly successful test. And so that was, you know, another piece of data there. And, you know, something close to home for you, at least in Vancouver, trying to make the case for for Chinese. You know, there’s a very large Chinese population in Vancouver. And trying to show that not only would we be behind Uber, if you because they basically launch at the same time, but there would be a significant lift in revenue. If we did launch with Chinese ended up not moving forward with it, unfortunately. But you know, those were the kinds of cases that we tried to build as we as we tried to push localization for what?

Andrej Zito 

That’s a good point that you mentioned that you were done testing the the localized ads. So what was the initial scope for the launch? Was it just the app?

Zachary Haitkin 

App and website. So the website, there’s a content management system, a third party content management system that integrated with the translation management system that made the website, pretty easy to localize. And then the app was, you know, using a REST API with with the TMS to hook up to GitHub.

Andrej Zito 

So when we’re talking about the data, and the matrix you mentioned. To me, it was mostly like business metrics, like the data that you would present to outsiders. But where did the data that you were tracking inside as the team, like when it comes to, for example, performance of the vendors, or like the translation turnaround time, the quality, maybe we can talk a little bit about those kinds of data.

Zachary Haitkin 

Yeah, um, you know, when we had a Spanish language manager, it was he was, he was the final say, for quality, he was doing the review, he was reviewing all of the Spanish translation. So it would go to the vendor and get translated, would get edited, if needed, and then he’d be the review. So it was very easy to track quality, because it was a person doing it, eventually, there was a shift in priorities, unfortunately, and they shifted the team to be just one person, just me. And, and that’s when I moved from, from project to program manager and ran the whole thing. But, you know, we relied very heavily on our vendors to manage the quality themselves. There was an internal ticketing system, where if someone saw an error, they would be able to report it. And if it was a localization error, there was a specific project in JIRA, the JIRA ticket would would come to me and I’d be able to be able to handle it, there was a handful of people using the Lyft beta app, that if they saw something, they, they’d be able to flag an issue pretty easily. So overall, I think there could have been more attention paid to quality but because of resourcing there just wasn’t, it just wasn’t there to have it be as good as it could be. It was fine. I mean, you know, that we were able to catch a significant amount of errors, but we really did rely heavily on the vendor for for quality assurance, and we worked with some some really great vendors. I was really thankful to have you know, I’ll name drop them, because I love them so much mother tongue, I don’t know if you’ve heard of them before. They’re part of the they’re part of a larger group called Omnicom, which is a big, you know, marketing conglomerate, but, but mother tongue was was really, really great. In, in their translation, transcreation, they really, they, they took the time to understand the Lyft voice and really went deep. In fact, so much so because we came up with this program where we actually brought the translators on site. And this is probably probably the thing that I’m maybe the most proud of, in terms of that the localization experience in program, we actually were able to get approval from left to bring the translators on site. And I mean, we brought them from around the world, from from Brazil, from Montreal from Mexico City. They they came to Lyft headquarters spent several days with me and the rest of the team in the very beginning. And were able to get their questions answered, and also just kind of integrate into the Lyft culture. And for them to see how the company operates. I gave him a tour of the office, we had Lyft heads, at the time had cars that you could take out and do like practice Lyft rides, if you wanted to actually go out and do rides in a car. So you know, we pretended I was driving, they would sit in the passenger seat, they would pretend to be the driver, another person, another linguist in the backseat would pretend to be the passenger, they’d switch. So they’d be able to see how the app actually functioned in English. And it wasn’t, you know, it was tough, like, especially being not in the US. There was no way for them to see how it actually worked unless they had been the US and taken taken ride. So giving them a chance to come and talk to us directly, and ask questions and get the experience of what it means to be a driver and a writer that really, really helped us in terms of setting them up for success with quality down the line. So that was definitely something that I recommend. It wasn’t wasn’t cheap, you know, but I absolutely think it was it was totally worth it. And I even came up with some metrics on, you know, based on the cost of travel and how much time it would be. Yeah, and I really broke it down. I said, Look, it’s going to cost them it’s going to cost us this much to bring him here. If we see this amount of incrementality from improved translations, it’s going to be worth it. And it was something ridiculously small. It was like a fraction of a percent because of the scale that Lyft was operating at all we needed to do was have this this amount of incrementality to make it worth it and not inevitable and that’s that’s absolutely got to see approval to To bring them on site. So that was just a really great experience in general. And that built the relationships with those linguists directly. Because you don’t see that a lot. Actually, you don’t see translation vendors letting clients have direct access to translators, that’s, that’s usually something where they separate mother tongue was was amazing in the fact that they let us have that direct contact, they came on site, and we brought mother tongue, our project managers on site to and so they were able to really embed with us at the company at Lyft. And, and just really understand what we wanted to do and how we wanted to do it. So yeah, that was a very long winded answer for how we dealt with quality we set we set them up in the very beginning with the tools that they needed, and built the relationships from the very beginning so that there were fewer quality issues down the line.

Andrej Zito 

I may be mistaken here, but going to where you are right now at Netflix, I think Netflix is one of the teams that still has some in house linguists right. I don’t know if it’s for all languages, or maybe just for some languages.

Zachary Haitkin 

I don’t think there’s in house linguists. Okay. I think so there’s there’s a team of language managers that have different regions or different locales that they oversee, that are a lot of them are speakers out of that of that language. I would say there’s not, there’s not a language manager for every single language. Some of the language managers oversee multiple regions, you know, in, in Europe or in APAC. So there is quite a large vendor pool that that Netflix does pull from. But the the language management team is is pretty large as well. But no, I don’t think there’s there’s in house thing was that that are being used?

Andrej Zito 

The reason why I asked this is I was thinking that if we can compare, like what you did for Lyft, you know, bringing the people on board with Netflix, and I thought like, maybe we don’t need to do it, because like Netflix already does it. But okay, I guess I was wrong. But the thing is, like, you mentioned that for live, like the people in outside of the countries, they couldn’t use Lyft app, but for Netflix, they can use it. So do you think like let’s say in Netflix, and Netflix, it will still be beneficial for people to spend some time with the team and to absorb the culture? Or you think like Netflix is such a global phenomenon that people already know, like, what Netflix is and how Netflix communicates?

Zachary Haitkin 

I think it’s the latter. You know, I don’t I don’t want to say that it wouldn’t be beneficial, because I think any relationship building at all, just in my in my career, and in general, that’s definitely a piece of advice I would give people is build the relationships with with people, and that’s what’s gonna lead to success. So sure, I think there would be some benefits for bringing linguists on site and meeting the actual engineers that they’re translating strings for, would it be as beneficial? No, I don’t think so. Because Netflix is so accessible globally, that it’s very easy to get the same experience on your computer anywhere that it is, if you’re if you’re in a specific location. So I don’t think it would be nearly as beneficial to to a service like like Netflix than it was for Lyft.

Andrej Zito 

I’m curious about the relationship that you mentioned, I would assume that in your experience, you had some underperformers? How do you handle that relationship? When you see that the people really don’t cut it for you?

Zachary Haitkin 

underperformers from a translation quality standpoint,

Andrej Zito 

for project managers or?

Zachary Haitkin 

Well, I don’t really have an experience like that in Netflix yet. I’m still very new, you know, I’ll definitely say that I’m coming up on three months at Netflix and at Lyft didn’t really see a lot of underperforming from our internal team at all. You know, the only reason why the team was was downsized was due to priorities certainly had nothing to do with performance. We were all pretty amazing at what we did. As far as far as linguistic quality and vendor quality. There was a time where we had some difficulty with Spanish language translation quality from a vendor, and we ended up having to having to move on from them. And that was a difficult conversation. I think I think we gave them you know, a decent amount of feedback and a chance to improve and we just weren’t seeing it. And we ended up having to having to switch content over and i think that’s that’s a good lesson for don’t put all your eggs in one basket, definitely have multiple vendors in your program so that if you have an issue with one, and yeah, this isn’t anything against that vendor. You know, they’re they’re pretty large vendor, they have a lot of clients, you know, I think it was more just an unfortunate situation. So I, you know, I don’t hold any any grudge or any ill will towards them, but just wasn’t working out for us. And we ended up having to move on. So you know, it’s it’s not an easy conversation to have. But it is important to to read the signs, read the signals. And if you need to make a switch, make a switch. There’s a lot of quality. Translate translation vendors out there. And if one’s not working for you, there’s there’s probably another one that will

Andrej Zito 

was that the decision of the language manager that you mentioned? Because you said that he was doing all the reviews? So was it? Was it feedback from him input from him or data from him? Or was it from some other vendors?

Zachary Haitkin 

It was mostly from our internal language manager at the time. It was ultimately the program manager who made the decision and had that conversation. But it was all of us together when we when we finally cut the cut the cord, but yes, it was our internal Spanish language manager who had been reporting feedback recording specific instances of lower quality. And that was, that was ultimately why we decided to move to a different vendor.

Andrej Zito 

Okay, so let me get this right. So you were part of a team? And then suddenly, it was just you? was it was it really like a sudden sudden thing? Or was it the plan? Or?

Zachary Haitkin 

No, it was very sudden, it was very, very sudden, it was almost overnight, pretty much. In my opinion, it was not executed very well, I think there could have been better communication. But that was the decision that was made at the company. It was very bittersweet for me. Because, you know, the people that I Well, the people that I had been working with for the last year, plus, were suddenly gone. And I was moved into a role that was I was promoted into a role to run the whole thing. So, you know, it was it was a step forward for me. But the people that I owed a lot of my success to, were not there. And it was very sad. It was I was not happy about it. Again, I think bittersweet is probably the best way to put it. But, you know, removing emotion from it, and understanding that it’s a business. And this is the business decision, that it was not a personal decision. I think that helps. But it was it was very sudden, to answer your question. Yeah, it happened. It happened really fast.

We’re always creating new localization content

Make sure you don’t miss anything. Join 3722 other professionals on our mailing list and be the first to get our upcoming newsletter. 

If you enjoyed that, you’ll love these…

Why hello there!

Enjoy 10% off

on your first course when you join our mailing list.

* All information collected will be used in accordance with our privacy policy. You may unsubscribe at any time.