Unbabel’s CEO Vasco Pedro was recently a guest on The App Guy Podcast hosted by Paul Kemp. He spoke about going through Y Combinator, growing a startup and finding balance between private and professional life. Here you can read the full interview:
Paul: Welcome to another episode of the App Guy podcast. I’m your host, that’s Paul Kemp. This is the show where we get inspiring app entrepreneurs, we’ve had a lot of founders going through what we call incubators, accelerator programs. And we’ve got a great episode where we’ll deconstruct the journey of a particular CEO founder who’s gone through a very famous incubator program called Y Combinator.
Let me introduce to you Vasco Pedro, he is the CEO of Unbabel a Y Combinator backed startup that combines crowdsourced human translation with machine learning. It’s going to be a fascinating chat. Vasco, welcome to the App Guy podcast.
Vasco: Thank you, Paul. Thank you for inviting me. Happy to be here.
Paul: We’re going to go through your product. But I’d love to know like what the journey was like for you going through Y Combinator. We hear so much about it. There’s a lot of people trying to get in. How was your journey going through Y Combinator?
Vasco: It was great, it was challenging, but it was definitely worth it. So us for example, we were literally about to sign a term sheet the day that we got the email saying that we would be subjects for an interview.
So it changed our course from the beginning. We’re five co-founders, which is a little bit unusual for Y Combinator, and the five of us had to fly in for the interview and it was ten minutes, highly stressful. Obviously things went well. Then throughout the whole three months we rented a two bedroom apartment. The four guys were in one bedroom, Sofia, the girl, was in another bedroom which doubled as an office – a sales room, and then we worked in the living room. I think one of the things that sometimes people think about YC is that it’s very glamorous. It’s mostly just a lot of hard work – staying in one place for a long period of time, just focusing on what you’re building. And it’s really about removing distractions. A friend of mine tends to describe it as a relentless personal growth. And I think that’s really the beauty of Y Combinator – in the way that it helps you focus on the fundamentals of the company that you’re building. And then it gives you the space and the motivation to work as hard as possible towards getting those goals.
Paul: So for everyone listening who is tempted to apply for this year, what one thing could they be doing now to improve their chances of success?
Vasco: I think the best thing is to build something people want. There’s no magic bullet. If you’re pre-launch, get to launch as soon as possible. Demonstrate that you have either a social traction so that there’s people that are very interested in what you’re doing. If you can launch it and show that people care and that you’re growing on your consumer app, that you’re growing on users and engagement.
If you’re an enterprise app, ideally start selling, and show that you have a ton of people that want to use what you’re building. I think that’s the basic thing, right? And then, of course, it’s good to have strong team, it’s very important. YC is really good at understanding the relationship between founders and company, and usually that’s at least 50% of the factor. I think one of the most common causes of start up failure is actually founder disagreement. So having a relationship with your founders, and having done it before together, those are all positive things.
Paul: It sounds great. So let’s talk about what you’re doing then. We may return to the journey of Y Combinator, because I think that’s fascinating. But let’s educate the listeners to what it is you’re actually doing with Unbabel?
Vasco: If you think about it, we kind of live in this fantasy that there is this one Internet – that was kind of true in 1998. So in 1998, English was representing about 80% of the Internet. There was this feeling the most important tool in global communication, which is still true. But what’s happened since then is that if you look now, English actually represents 35% of the Internet.
I think there’s a fundamental phenomenon that’s going on, the same one that lead to the creation of multiple languages – as soon as there’s critical mass of a culture in some place, they start creating their own content, their own culture, even online. And that’s been happening with languages too.
So realistically, there’s an English Internet, there’s a Chinese Internet, there’s an Arabic Internet, a Brazilian, Russian, etc., where the vast majority of content that people interact with and consume on a daily basis is all in their native language. 75% of the world doesn’t even speak English as a second language.
And so this is becoming actually a bigger and bigger problem. There’s a reason why all successful global companies are companies that operate in multiple markets. There’s a huge advantage in being able to be global and the fact that the Internet is actually diverging in terms of language.
It creates additional challenges for any company that wants to be global. The problem is that, if you look at possible solutions like machine translation – the quality is not there. I’d love to tell you that we’re going to solve the machine translation quality problem in the next 10 years. My background is in AI and natural language processing so it’s something I would love to see but I don’t think that’s going to be the case. There’s just something fundamentally hard about the language. Language is probably the most obvious expression of our intelligence. For example, when you’re speaking with someone who has a strong accent, it’s just very hard immediately to not be biased toward their cultural capabilities or their intellect.
So language really is a hard nut to crack and if you’re going to solve language, you’re going to end up solving strong AI, and nobody really has a clue how to do that. Machine translation doesn’t really work for this purpose. The other possibility is human translation but that’s really expensive and slow, and really hard to scale.
In a way, Unbabel is the layer that unifies this Internets. And what we do is we combine both: we use AI, and when I say AI, I mean machine learning, deep learning, machine translation in different forms, all sorts of different components of AI, to reduce the amount of human effort required to translate with high quality.
This means, on one hand, automating parts that are easy to automate, on the other automatically understanding who are the best people to translate something and also making each of those people really efficient. So AI helps humans become better by augmenting their capabilities. The results of this is that we already reduced the cost of translation 10 times and the goal is to get to a point where you can provide unlimited translation to the entire world. That’s what Unbabel is about.
Paul: Actually that is quite relevant to a lot of us listening because we are always translating apps and translating various things and we need to do more of it. Over the period I was involved with the languages app, it did really well, got up to the top 100 overall in the app store.
Was there any particular moment that you’ve had of breakout success that you can attribute that to something you’ve done? You’ve been growing and you’ve been, obviously, focusing on user interface aspects, what about some spike in downloads or attention that has driven your success?
Vasco: Traditional market already exists, right? There’s a part of what we’re trying to do that is a well understood problem. People have been translating since forever,Egyptians had translators and it was something that was always very, very important. In a way this area of business is well understood and there’s a strong demand. So when you go to a customer and you say „hey, do you want to do what you do but in a much more affordable way, a much more scalable, faster” – this is very compelling, it makes it easier to attract attention.
One of the things that we’re trying to figure out is how the fundamental process of translation. So if you think about it this way is, if you look at the Industrial Revolution, for example, when shoe factories really started scaling up, they didn’t just hire a bunch of shoemakers and said, okay, you do one pair and you do this other pair. What they did was they deconstructed the process of fabricating shoes and then they were able to optimize it and scale it. And this doesn’t happen yet with translation. I think being very successful means making a customer happy, creating this magical moment.
For example, we have this Zendesk integration. That means that you can do customer service in any language in a seamless way. So if you speak English, you get a ticket, let’s say, in Chinese, Unbabel will translate it to English, you write your reply in English, and we translate it back to Chinese. Your user basically doesn’t even realize their not speaking to someone that can’t speak their language. And it creates a very magical moment.
Paul: So you’re hoping for the fact that you’re offering something so magical that it’s going to get a lot of word of mouth, almost like viral kind of marketing. Or are you hoping that some PR press release marketing is going to attract attention because one of the big problems you see we have Vasco with everyone listening is, it’s really hard to get your head stuck above the crowd and get noticed with such a noisy market.
How are you coping with it?
Vasco: We definitely have that problem. So, like I said, translation is well-understood, it also means it’s a big market and there’s a lot of players. And of course, I can tell you the hundred different ways that we differentiate ourselves from our competition.
But the competition is there nonetheless. Figuring out a way to get attention is always hard. I think it also depends a lot on what kind of company you are, right? An insurer, consumer company or enterprise company. You have very different tactics. It’s usually said that your first growth should come out of the product, right?
You should have a product that is so appealing that customers, be it early adopters, will flock to it, will see the value. A lot of your growth initially should be organic, should be word of mouth. That being said, that’s kinda the theoretical, this is the ideal situation.
I see a lot of overnight successes, eight years in the making. And I think there’s a lot of engineering behind having success and in our case we use a mix of things. Sofia, one of Unbabel’s co-founders usually says she’d totally write a blog post about surviving an MVP because that’s what most startups do, right?
You come up with a basic version of your product which is, by definition, incomplete and you launch your MVP. And you don’t really know what you’re doing, you don’t really know who your customers are and how are you actually going to market to them so you try a bunch of different things. In our case we tried PR, we’re trying to do partnerships, we’re trying direct emails, we’re trying word of mouth, we’re trying conferences and then slowly you start understanding. There’s two schools of thought here, one is that you can’t really innovate in all aspects of your business. So if you’re already innovating your product, don’t try to also innovate on sales and marketing and all of these other areas, focus on what you’re innovating and go with proven techniques on others. Here’s the possible sales channels – just explore those and optimize for those.
The other school of thought is, especially for consumer apps, that no app has been dominant without exploring a new channel that somehow wasn’t explored before. YouTube with MySpace, some people with YouTube. As soon as something becomes dominant something can start exploring that to grow.
The fact is that even if you’re core product is really good. To get to a point where you have the feature set that appeals to your customer to the point where they’re really promoting it to other people takes time, innovation, focus, understanding who your customer is. That is very hard to do overnight and also creates this virtuous cycle. The more you understand about your customer, the better you can do at both your product and marketing to them which means you have more people recommending. I think that question that a lot of startups have is „how do I grow?” How do I get in front of people, potential customers? I think a lot of it is hard work. It’s not just expecting that you’re going to get there in one week, but being able to recognize the signs. That’s another thing that I think YC helps you a lot.
You have so many different things to optimize. I think the best way is to understand what is your core metric. Especially at the beginning where you can do a number of things. Just pick one that if you grow in this metric, whether it is number of users or amount of money you’re making, that you can honestly say that if you’re growing in that metric it is an actual sign that you are doing well.
And then just focus on growing that one number every week, like 10% a week, which is really hard.
Paul: You have gone through YC and we do hear that they expect extremely successful companies on the outcome and so they put huge expectations on growth, on valuations.
These guys want unicorns to pay for all the ones that don’t work. Are you able to just focus on one metric given there must be such huge pressure to deliver massive growth?
Vasco: One way of thinking about it is, you don’t really control how you can grow. I mean, you have to be able to grow a number that is convincing. The thing you can control is what is the metric that you choose, right? So it should be something that actually represents something that you’re comfortable with. But then, if you really focus on that, especially in the beginning, you shouldn’t get high growth.
It’s kind of a self fulfilled prophecy, right? If you’re not growing then either you have the wrong metric, or you’re really not creating a product that people want. We got in and we said look, we had launched the first week in terms of selling and called [Paul] Graham – at the time he told us look „if you’re selling nothing beats growing in revenue”.
That is the best graph you can have. That’s the best proof you can have that what you’re doing people are willing to pay for. And so the first week we had just launched, we had sold $20 in translation. And we thought great, 10 percent a week, it’s gonna to be easy.
$22 next week, I’ll have it right here in my pocket, we’re done. Especially in our case we have this really strong technology and all this stuff, and we were really thinking that we were going to get to really focus on the technology. Paul Graham said „guys it’s awesome that you have that technology, but right now focus on growth, focus on understanding what is the value your customers see in your product. Afterwards you get back to improving the technology”. So everybody can grow 10% can go ahead and implement whatever features they want, but what you’ll find out, that’s exactly what we found out, is that when you focus on growth, especially in the beginning, you got to have a couple of weeks that you’re going to grow like, a 1000% right?
Because you start very small. But then the hard part is not to grow 10% in one week, it’s growing 12 weeks 10%, right? At least 10% because it keeps getting harder and harder and what it forces you is to do things that don’t scale in the beginning, which also seems kinda retentive, but it helps so much.
For example for us, one week, in order for us to grow we decided to list ourselves in Elance to apply to translation jobs and then put those translation jobs in the platform. And we thought, okay, maybe this doesn’t apply, but what it did show us is how much a translator life sucks.
50% of your time is spent applying to jobs you don’t get and the other 50% of your time, most of it is actually spent dealing with weird formats and scan PDFs and all sorts of other crap – it’s not really about translation. Doing that gave us such an insight on what translators are going through, that we’re able to make translator’s job better. Actually that was one of our things that lead us to change our entire way we pay translators and I think is much for the better.
Paul: And you’ve got a lot of co-founders. From your perspective, do you have any advice on how people can meet interesting co-founders?
Vasco: I have advice on how to maintain the relationship between co-founders.
Paul: Yeah. Yeah, we did say that’s incredibly important.
Vasco: Unbabel has five co-founders. Four of us had worked together in a previous startup. Two of us actually did a third startup together. And Sofia and I met a few years ago when I got back to Portugal. There was already a relationship. We weren’t childhood friends, there was no relationship beyond that, but it was a work relationship that grew into friendship and I think that’s usually a good start.
I think that in startups that have more than two co-founders usually, 95% of the time, one of the co-founders is not there after a year. And the fact that the five of us are still together after two and half years and going strong is a very good sign.
I have kind of a theory on that, the way I like to think about this is in terms of emotional debt. So, for startups it’s okay to have technical debt, everyone understands that and it’s fine, but I think startups have other types of debt.
You have marketing debt and sales debt. In the beginning you don’t know what you’re doing, you’re still trying to figure out, you don’t have the resources to hire someone that already knows how to do it. But I think the one debt you cannot accumulate is emotional debt between founders.
And that’s usually, it’s a little bit like, marriage or any kind of relationship. On a daily basis you have little things that you end up not discussing and then if you don’t deal with those things, once you get to big issues you, end up discussing everything else, right?
So for us, what it works is we surf. Every two weeks we go surfing on Friday afternoon. We started this right in the beginning, we suck at surfing, we’re really bad, but the thing we noticed is that there’s something about doing this strenuous activity followed by a period of talking that makes it very clear that you’re not only doing this with someone that you respect, but someone that you enjoy, that you like spending time with.
And somehow those moments, after you’re kind of tired, and you’re having a bite to eat together, are the times you can have very honest conversations about things that might be directly related to work. But it might be things that have been annoying you or on your mind, and that helps pay off the emotional debt.
It obviously doesn’t have to be surfing, but I really encourage people to find what is the one way they can pay their emotional debt, and there will be plenty of time for you to think „I don’t have time for that, I need to focus on work, but I really think it pays off to not postpone it”.
Make sure that you find the time to spend with your co-founders, reminding yourselves that they’re people you enjoy doing things with. Otherwise it’s very easy to get too stressed.
Paul: And finally, then, the last thing is, one final question. How do you actually organize your day to be most productive? Do you have any tips on what you’re doing for productivity?
Vasco: It depends on the phases, right? So, it also varies with different startups. If you’re a two-person startup and you have kind of the hustler and the hacker, you end up having to do a bunch of different roles.
In my case, we’re fortunate that the fact that we have five co-founders means that we can all focus on different things – there’s enough tasks that I can be focused mostly on one hand investor relationships and on the other on vision and helping the team. The way I think about myself is when I’m not fundraising, which fortunately is not that often, I’m trying to understand where can I block things. I have a vision of the whole company so I have the best opportunity to work both in sales, marketing and technology and all of the other areas to bring knowledge from this other areas and then block situations. I have four kids. In the morning I take a couple of my kids to school then I usually get to the office about a quarter to nine. And in the morning there’s more about scrum and understanding of where things are. And I like to tackle big issues in the morning while they’re still fresh. Then depending on the day, lunch, sometimes we have late lunch catering in battle, we have discussion around it. Some days I try to have one day a week where I go out to lunch with my wife, we live close to where the office is.
When I’m in Portugal, when I’m in San Francisco it’s kind of harder to do. But I try to have meetings also during lunch when I can.
Paul: Can I just jump in, because I actually have twins as well. I think a lot of people listening to this do have kids. Is it beneficial to have a start up or be an entrepreneur when you’re with kids? Does it give you an amount of freedom that you wouldn’t normally get if you had a corporate career?
Vasco: I think there is advantages and disadvantages. So I think the advantage for me is two-fold.
One is in some ways, you can have more flexibility. The way that we set up things is I live close to the office, kid’s schools are nearby. If I need to go home very quickly and come back, that’s easy, right? I can go have dinner at home, come back, keep working, that’s not a problem.
The other thing is that you really feel like you’re not playing startups anymore. You know? My first startup when I was twenty-something, it felt different. It felt like you had more time to waste. You have less time to waste now. And you also feel like it makes things matter more.
The time that you’re spending here is time that you’re not spending with your kids, and you see more of how much you’re sacrificing, so it better be worth your time and you really want to make sure that it’s efficient. On the other hand, of course, its more, well it’s less stable.
I do think that there is a little bit of reality distortion in a lot of founders and, you know, it’s probably the case with me. It’s that we’re less able to see the risk, because in your mind you kind of feel like you can control the outcome a lot. You think this is an awesome opportunity, this is great prompt to work.
The fact is that I’m happy. I talk to a lot of my friends that have families, and at the same time they’re doing something that a, they’re not 100% happy with, right? And that also has an effect on their relationship with their kids. I think if you’re doing something that you’re passionate about that’s also a great example for your kids. See that you go out and you work hard but you do it in something that you think makes a difference. Then there’s also an upside. I know plenty of friends who had successful startups and very significantly changed their lives and the possibilities that they could give to their families.
I think it also depends on the area that you are. In my case, I did my PhD at Carnegie Melon. I was fortunate enough that it never felt that even if a startup fails I would be out of a job. I always felt that my skill set was valuable enough that there would be something next if this didn’t work.
That’s something that a lot of people think – that working in a startup is a detractor for working in a large company. That has been my experience also. In the end, if you’re doing something that you’re passionate about and you become good at it, even if you’re not successful you learn so much. You’ve acquired so many skills that probably wouldn’t have otherwise. You actually became more valuable for the job market.
Paul: Ido have to confess that I do another hobby podcast called The Entrepreneur Dad Podcast.
If anyone who’s interested in that debate, just go to theentrepreneurdadpodcast.com
Vasco: Definitely, I think the hard part is the hours. In our case we have offices in San Francisco and here in Lisbon, I spent about a third of my time in the US, so that part is tough. Or during YC my youngest daughter turned one. And my co-founder’s daughter at the time was two months old and he moved to California to do YC for three months away from family. So that part’s hard.
Paul: It is. Well, we’ve run out of time unfortunately, Vasco, but it’s been a wonderful chat with you.
You can listen to the podcast here.
The post The App Guy Podcast: Vasco Pedro shares his advice on how to successfully grow your startup appeared first on Unbabel.