In February 2014, Unbabel was just getting started. The other four founders and myself were halfway through YCombinator early stage venture funding, an intensive three month program that leads up to a demo day, when all the startups pitch their products or services to an audience of investors and press.

We were living in Mountain View, just a short bike ride away from NASA and Google, cramped up together in a small two-bedroom apartment. It wasn’t by accident either; you were supposed to be spending a lot of time together in close quarters. We woke up, worked, had lunch while talking about work, worked again, talked about work all the way through dinner, and worked some more. There hardly is time for anything else when you’re trying to put together a working prototype in just ten weeks.

But that wasn’t the case that night. It was February the 13th, Valentine’s day eve. André Silva, a developer and our first ever employee, was bunking with us for a couple of weeks. He thought it would be cool to create a simple landing page where people could translate a love message to send their loved ones. Love conquering borders, and all that.

André and Bruno Silva, our founder and Head of Design, quickly jumped on the idea. Sadly, it never saw the light of day. André ended up falling asleep on his keyboard — lovers around the world would have to find another way to express their love.

Innovation meets pizza night

Fast forward a few months, and we were no less overwhelmed with work than we were during that winter in Silicon Valley. We were building the product, signing customers, fixing things as fast as they would crash. Our jobs were mostly reacting to whatever was going on — something that happens in every early stage startup.

But what if we set aside some time every week to take Unbabel’s technology to build something interesting, useful, or both? Projects we had in mind but never had time to implement, tools that could help us somehow, completely spur-of-the-moment ideas, even if they were unreasonable, unfeasible, and unpractical. We wanted a space to be creative, not reactive.

And so, the Valentine’s Eve saga gave way to Wednesday pizza night. Anyone who wanted to work on a side project could stick around on Wednesday after office hours to eat some delicious pizza and work on a prototype that they could then present to the company.

As the idea of a Night Lab grew, and more people joined and participated in it, it no longer made sense to restrict it to Wednesday nights — a lot of people were interested in working on their own projects, so we decided that a Night Lab was something you could do whichever day of the week you wanted. And so it continued all these years. Stay late, eat some pizza, and then at the end of the quarter we’ll vote on the best initiative, which will be recognized with a $1000 prize.

Don’t get me wrong, I have two daughters whom I very much love to spend time with. And all of our coworkers have their families, friends, pets, and hobbies to go to at the end of the day. Even if it’s just watching Naruto for the hundredth time. Night Lab isn’t meant to derail anyone’s work-life balance, just provide a space where people can join their peers to build whatever’s popping on their minds.

Night Lab throughout the years

5 years and many pizzas later, we’ve seen amazing projects brought to life. There was Salvadorable, a video player that translated the lyrics of Salvador Sobral’s Amar Pelos Dois, the Portuguese Eurovision song that went on to win the contest, into 17 languages.

There was Unbabel Voice, an automatic answer system that used our transcription and translation integration to enable anyone to call that number and translate any voice message into any of our supported languages.

There was Unbabylon, a landing page celebrating our nationality diversity, literally putting every employee on the map. Even the short-lived yet extremely popular Paper2Pdf tool that got us through many months with a shortage of designers, where anyone could create one pagers from any Dropbox Paper document.

And yet, a few months ago, the tradition sort of fizzled. We were going through a high-growth phase — there were a lot of growing pains, team restructuring, new hires — and Night Lab was the last thing on our minds. Despite the fact that the previous presentation wasn’t that long ago, in the end of May 2019, over half of the people working at Unbabel had never seen one. Suddenly, I had people slacking me because they had overheard someone talking about “Night Labs” and didn’t know what that was all about.

And so, we decided to bring it back.

There and back again

Usually, software companies have specialized teams whose main job is to innovate. At Unbabel, that team is called Labs — they’re responsible for coming up with new and exciting applications for our technology. But innovation shouldn’t be cordoned off in one team. Paulo Dimas, VP of Product Innovation at Unbabel, believes that’s one of the main challenges in any company.

Night Lab offered some relief — it always opened innovation to everyone, even in the most unexpected places. There was rarely an initiative that wasn’t cross-functional. We would get developers working with sales people, linguists working with marketers — every project was a product of the collective.

But Night Lab was facing a problem. What do you do with all of these great ideas? Some would stay live for people to use them as they saw fit, but many would just be abandoned after the presentation. Unbabel.org, for example, was a very ambitious Night Lab idea that, inspired by the great work being done by organizations like Translators Without Borders, proposed the creation of a non-profit arm of Unbabel that would provide our AI-powered, human-refined translations for free in response to the unfolding humanitarian crisis in recent years. But at the time, we were about 50 employees. We didn’t have the people, skills, or legal resources to take on such a challenging project.

Closing the innovation loop

But at this stage, we can have a more hands-on approach and see these prototypes all the way through. And we can do this by creating three different paths:

  1. If it’s a simple tool or landing page, just give it the bandwidth it needs to stay live.
  2. If it’s something practical that could be integrated into a product roadmap, the product manager should have ownership of it.
  3. If it’s something a bit more ambitious, with clear business potential to become a viable product or service, the project should go to Labs, where that team would incubate it and develop it into a full fledged product or service.

In an interview about how Labs works, Paulo explains this process:

One of the things we learned is the importance of incubating a team inside labs. When we decided our team has all the disciplines, when we have that product launch on the roadmap, when we feel like this is going to be really big, then we start incubating a team inside Labs, a team that will afterwards become autonomous.

And in fact, it has happened before. Unbabel Cast, a mobile app prototype that allowed near real time transcription and translation, introduced in a Night Lab presentation in April 2017, eventually evolved into an actual product.

If you let innovation come from anyone inside your company, if you encourage them to work on new ideas and give them the space to see where they go, you close the loop. That’s the idea, anyway. We’ll see where it takes us.