How A Free Language Learning App 'Duolingo' Scaled to a $11.68 Billion Valuation
Today's newsletter is about how two immigrants made one of the best and highest-rated educational applications which makes $42 million per month.
The Founder
Luis Von Ahn was born on 19th August 1978 in Guatemala City, Guatemala. He was raised by a single mother, who was the first woman in Guatemala to complete medical school.
When Luis was eight years old, he wanted the original, old-school, first version of the Nintendo Entertainment System. But her mother gifted him a Commodore 64 computer. He didn’t know anything about the system, but he figured out how to load games into it, from then on, he started using more and more of the computer, and it fascinated him.
He studied at the American School of Guatemala (Colegio Americano de Guatemala). The school provided an international curriculum and gave students early exposure to technology and the English language, which, according to Luis was a great privilege.
In 1996, an international recruiter from Duke University, USA, came across Luis as he scored very high points on the Scholastic Assessment Test (SAT). That same year, Luis moved to the United States to complete his higher studies at Duke University.
He got a Bachelor of Science degree in mathematics in 2000 and was awarded Summa Cum Laude honors which means “with highest praise” or “with highest honors”.
From PhD to Millions
After completing his undergraduate studies, Luis was accepted for a PhD program in Computer Science at Carnegie Mellon University.
During his doctoral studies, Luis along with his mentor, Manuel Blum, created the ESP game (Extrasensory Perception Game), which was designed for labeling images on the internet.
In the game, two players who couldn’t communicate directly were told to describe the image. If both players described the image similarly, they would score points. This helped in generating the exact and correct label for the image, which in turn could be used to enhance the search engine of the internet.
In 2003, the ESP game was sold to Google for a couple of million dollars. Google later named it Google Image Labeler, which helped in proving Google’s search algorithms and image recognition capabilities.
Inventing CAPTCHA
In 2000, Luis attended a talk at Carnegie Mellon University, which was given by one of Yahoo’s chief scientists. One major issue discussed was about the spammers creating millions of fake email accounts by writing a computer program. These emails were then used to send spam emails, which crushed Yahoo’s system.
This made Luis realize that if he could make something that could differentiate between humans and automated programs or bots, he would overcome the spam problem.
So, Luis and Manuel Blum created CAPTCHA - Completely Automated Public Turing Test to tell Computers and Humans Apart, which basically showed deformed text images that must be typed correctly to prove you’re a human. Computers are unable to detect these distorted texts or images.
The Hidden Purpose of CAPTCHAs
By 2007, two hundred million people gave 10 seconds of their day for typing the correct text for captcha. If you multiply these 10 seconds by 200 million it would be equal to half a million hours per day.
Luis wanted to make use of these half a million hours. Simply getting rid of the captcha was not possible because the security of the web was dependent on it. So, they came up with the idea of Re-CAPTCHA (Redoing the CAPTCHA).
By typing the correct text for re-captcha, people were not only authenticating themselves as humans, but they were also helping in digitizing the books.
The re-captcha system showed two images of distorted texts - one was a known word that was used for verification, and the other was an unknown word, scanned from a book that couldn’t be transcribed by computer. The user had to type both texts correctly which in turn helped in interpreting and transcribing the words from the old book.
By 2009, the number of words that were digitized through re-captcha was 80 million per day, which was equivalent to 4 million books per year. According to Luis, about 10% of the world’s population helped to digitize books through re-captcha.
In September 2009, re-captcha was sold to Google. The amount was not disclosed to the public.
Forming Duolingo
When Luis Von Ahn was a professor at Carnegie Mellon University, he had a PhD student named Severin Hacker. Both of them were searching for a PhD thesis topic for Hacker. They were looking for something that could be turned into a company and could be used as a thesis topic as well.
Luis and Hacker both agreed that they wanted to start something related to education that would be accessible to everyone and not just biased toward the rich class of society.
Luis believed that high-quality education, especially learning English, was not accessible to those who were financially challenged. He believed that learning English could improve people’s economic opportunities.
Initially, they both thought of teaching math or computer science, but eventually, they settled on teaching English and other languages.
On November 27, 2011, they launched Duolingo with the idea of making people learn new languages for free.
Strategies Behind the Success of Duolingo
Luis and his team created a platform where learning a language was free. In 2012 there were 1.2 billion people around the world who were learning a foreign language; people desperately wanted to learn a new language.
The user interface of Duolingo is gamified. Features like rewards, points, and progress bars are added to make language learning more fun.
The feature of maintaining streaks made people come back and learn through the app so that they can be persistent.
Duolingo was not making any money till 2017. Then, they thought of monetizing the platform by providing premium subscriptions to the users. Users can learn through the app for free but have limited features like lives or hearts.
By taking a premium subscription called Duolingo Plus, people can now have unlimited lives; they can track their progress, take offline lessons, and use the application ad-free.
By 2021, Duolingo claimed that 73% of its revenue came from premium subscriptions
The notifications sent by Duolingo are simple and easy, which motivates the user. They are designed in such a way that people are tricked into feeling guilty for not completing the lesson or the exercise. For example - “Hey! These notifications don’t seem to be working now. We’ll stop sending you the notifications from now”
According to Luis, these notifications made a lot of people to come back and it resulted in a 3% rise in retention.
From 2015, Duolingo started showing ads after the end of the lessons. This allowed the company to maintain free access for users and generate money by placing ads in between.
During the pandemic, the company went public and its valuation increased to $6.5 billion in 2021.
Duolingo, as a company, has a huge database. This database is sent to AI, which in turn uses this data to personalize the user experience.
The Company runs more than 10k A/B tests to get more people to pay them, to get more people to use Duolingo, more people to recommend it to their friends, etc.
And that’s how Duolingo became the third-largest educational company in the world with an approximate valuation of $11.68 billion.
Thanks for reading!
Great story and great narration!