In less than ten years, the language learning app Duolingo has become the world’s largest educational app. It has over 500 million registered learners, 40 million active users and clocked $190 million in revenues in 2020. As part of the company’s early team, Gina Gotthilf has played a critical role in growing it. The former VP of Marketing at Duolingo talks us about her failed experiments, her success, how she grew Duolingo in new markets and over 300 million users. She also talks to us about Latitud, her latest venture that aims to help founders in Latin America scale their startups.
Q. You’ve led growth at Duolingo as the company went from 3 million users to 300 million users. Can you talk to us about your playbook and what are some of the things that you did?
Gina: I knew that if people discovered Duolingo, they would want to use it. That’s the benefit of working for a company that you legitimately think is making a great product. People needed to discover Duolingo and going after people with large audiences who could get this message out to a lot of people was the goal.
I started by reaching out to the top editors and journalists across the country on LinkedIn and Twitter to get them interested in interviewing our Co-Founder and CEO Luis Von Ahn (the inventor of reCAPTCHA). I knew that if he told the story and people saw the genius of the app they would check it out, download it, and that’s how we would grow.
To get journalists interested in speaking with Luis, I had to find a way to make Luis and Duolingo relevant to them. Luis already had an audience because he is the inventor of the reCaptcha. I would find the top tech conferences and universities and organize talks at them. For example in India, he gave a talk in IIT Delhi and at the same time, we would organize a talk at a top tech conference. Then I would mail the journalist with these hooks that Luis Von Ahn has been invited to a talk at IIT Delhi and your top tech conference. This induces a FOMO into the journalist and they think that they should know what this is about and they’re missing out. That was the strategy we used over and over in multiple different countries along with getting partners on app stores and getting Duolingo featured.
Q. What kind of product lead growth experiments did you run and how do you deal with failures?
Gina: More than 50% of experiments fail. You need to tighten your seatbelts and be ready for that. Don’t allow your team to dwell too much on any specific experiment. Designers and engineers will want to perfect the experiment, but you want to create a Minimum Viable Product (MVP) of each experiment intelligently so that it’s simple and you’re able to move quickly. Your focus should be on looking at results regularly, analyzing those results, and then making decisions. You have to have a portfolio of experiments running at any given time. The most important thing for experiments that fail is being able to learn from them. Some experiments may have a negative impact on user growth. In many cases, people think that didn’t work out. But you actually hit a lever that impacts user growth now what happens if you push the lever the other way?
For example with badges, we launched an MVP of that, and then it didn’t work. It led to no change. We put it aside for a whole year, and then later decided to revisit it and realized that we hadn’t launched an MVP of the experiment, we had launched a minimum version of the experiment. We had not launched the viable version of the experiment in that it actually didn’t elicit the feeling or the experience that the actual feature of badges would.
If you sign up and you saw a girl with a balloon that’s not at all rewarding. That’s not what badges are about. It’s about collecting badges and showing them off. We had to figure out if we’re going to launch badges, what does it look like? How many badges do we have? Where does it live?
Q. Can you give us some examples of product lead growth experiments that worked for you?
Gina: Notifications for sure. Focus on onboarding. The first few moments that your user has with your product are the most important ones. That’s where the majority of your funnel is and when people are most excited about your product and most willing to try it out.
Any experiments you do there will have a large impact on your overall user growth. We did a lot of experiments there. Letting users use Duolingo before they sign up versus making them sign up immediately. A bunch of micro-experiments around how many times do we tell them to sign up. What we found was most effective was having two soft walls that allow users to use the app without signing up and then one hard wall which tells the user that they cannot use the app anymore without signing up.
We did a lot of experiments around onboarding. Initially, we asked many questions. But what we learned was that the more you ask, the more people tend to drop off. We made it super simple and removed anything that we didn’t absolutely need. Then we had a progress bar where users could see their onboarding progress. The progress bar was a great example of gamification in terms of how can we measure people’s learnings in a visual and compelling way. We had different team members play the week’s top games regularly and then explain some of the gamification mechanisms that we could use for Duolingo. Games are the number one apps in terms of retention and borrowing from their playbook only makes sense.
We would also email users reminding them they’re almost at the next level or showing them how much they needed to do to get to the next level. We would also run experiments on the app store with the images, headlines, screenshots, and copy.
Q. Duolingo is well-known for its micro-copy and its app notifications. How important is it to have a warm persona?
Gina: People underestimate the power of the brand in terms of growth. The brand personality, how users relate to the company and the app, and what it means to them is crucial. The copy being super friendly was always intentional.
What drove us was being intentional about our brand persona and thinking through if Duo, the character, was a person. What would he or she be like? What is Duolingo as a brand? What are the words that describe us as a brand? Friendly, funny, humorous, not serious all those words informed everything we did from a content, brand, design, copy, and even a platform perspective in general.
One thing that we did that was a little bit controversial. We always went a little edgy to define a unique persona and not just be another vanilla brand. We tried all kinds of things. We tried snarky comments. The fifth notification you get is a notification we called the passive-aggressive notification internally. Which said, ”These notifications don’t seem to be working anymore. We’ll stop sending them for now”.
People started creating memes about that, and it became a joke, and we ran with it. Allowing that brand voice to evolve, learning from what’s working, testing different things, using words that wouldn’t just be used by other brands are all important things. Make sure that you’re carving out something unique so that it’s memorable.
Q. Tell us about your latest project Latitud and what your vision is for it.
Gina: I’m excited about the next ten years for Latin America in terms of tech opportunities and growth. Companies are being valued at many billions of dollars, and companies from abroad are beginning to look at Latin America. There’s a lot of opportunities and a lot of people ready to go after these opportunities.
I met my Co-Founder Brian Requarth last year. He’s an American who has had a lot of success building for Latin America and just sold his third company for $600 million in Brazil. What we want to do list the tide of the entire ecosystem and increase the chances of top tier tech entrepreneurs succeeding from their first days to the investment process and growth.
We launched the Latitud fellow in August and are now in our fourth cohort. We select top tier tech entrepreneurs doing amazing things and do everything we can to make them succeed. There are three pillars that we identify. The first one is community. We are giving founders the opportunity to meet other tech entrepreneurs to ask questions and progress. The second is understanding capital and fundraising and how to play that game. How to meet the right people, pitch and so on. We’ve also started the first rolling fund in Latin America so we can start investing in some of these companies, and we’re also building products to help make it easier to launch tech companies in Latin America.
Q. What is your biggest advice to an early-stage startup founder?
Gina: I will throw out a couple. Dream big, go for the big opportunities, big markets and make a real impact.
The importance of a technical founder is something that I would never underestimate. No code is growing, and maybe that will change over time. But still, technical co-founders who can build but also who have that mathematical mindset. You don’t understand how valuable they are until you have one on your team.
Fail fast. Don’t make the mistake of trying to create the perfect thing and launching the perfect thing because you’ll just waste your time and fail anyway after having invested a lot more into it.