Game of Tongues: How Duolingo Built A $700 Million Business With Its Addictive Language-Learning App

Beg for your life in Spanish,” demands Duo, a plump green owl. The viral meme circulating on Twitter stars the mascot for Duolingo, the hugely popular language-learning app, downloaded by more than 300 million people worldwide. While the meme is a joke, the real app is plenty insistent.

Slack off on your daily Spanish lesson and a cellphone alert will auto-generate at the same time you used the app the day before. Duo appears on your screen as an emoji, nagging you to keep up with your drills.

Millions obey Duo’s commands. Created seven years ago by MacArthur “genius” computer geek Luis von Ahn, 40, Duolingo has hooked everyone from Bill Gates, Khloe Kardashian and Jack Dorsey to Syrian refugees in Turkey.

“The moment I felt proudest was when I realized, Wow, the richest man in the world is using the same system as the lowest people on the economic scale,” says Von Ahn. “That is for me really special and pretty big.”

Duolingo has lots of room to grow bigger still. More than 2 billion people worldwide are studying a foreign language, and increasingly they are doing it online. Digital language-learning generates $6 billion in revenue, and that number is projected to rise to $8.7 billion by 2025. But it’s a highly fragmented market with dozens of players scattered around the globe, and it’s begging for a dominant one.

Von Ahn certainly knows how to execute on a grand scale. When he was a 21-year-old grad student he gave the world CAPTCHA—better known to billions of internet users as those annoying squiggly letters that you have to transcribe to prove that you are not a bot. Later he sold two inventions to Google, netting more than $20 million.

He splurged on a Lamborghini and a Tesla Model S but otherwise leads a modest life, remaining in the six-bedroom house he bought with his ex-wife in Pittsburgh’s Point Breeze neighborhood, close to Duolingo’s offices. “I could have gone to Guatemala to live in a villa, but I didn’t want to,” he says. Instead, he has thrown himself into building the world’s most downloaded education app.

Duolingo already offers more languages than its competitors do, 36 at last count. They include little-spoken tongues like Hawaiian, Navajo and Gaelic and the fictional language High Valyrian from the HBO blockbuster Game of Thrones (1.2 million users are studying it). À la Wikipedia, Duolingo enlists volunteers to help create its more obscure courses.

Seven years after its launch, it has nearly 30 million active monthly users, according to its own figures. Venture capitalists have pumped in $108 million, ballooning Duolingo’s valuation to $700 million in 2017, $150 million more than the market capitalization of Rosetta Stone, its 27-year-old publicly traded rival.

Along with teaching more languages (Rosetta Stone offers 25), Duolingo attracts users because its basic ad-supported version is free, compared with the $120 a year Rosetta Stone charges its 500,000 subscribers. Another competitor, Berlin-based Babbel, says its revenue of more than $115 million comes from the $85-a-year subscription fee it charges its million-plus subscribers.

Only 1.75% of Duolingo’s users pay for its ad-free version ($84 a year), but because its base is so huge, revenue hit $36 million last year. Von Ahn says that will climb to $86 million in 2019 and $160 million in 2020 as more users sign on and pay for the premium app, which will attract them with new features. Employee head count will rise from 170 to 200 by year’s end.



Duolingo’s headquarters, in a converted furniture store in East Liberty, a gentrifying neighborhood not far from Google’s Pittsburgh office, is set to expand to a second floor. The company isn’t profitable, but Von Ahn says it will be cash-flow positive this year. He is planning an IPO by 2021.

He likes to say that the money he forfeits giving away the app is equivalent to the cost of his rivals’ bloated marketing budgets. He also boasts that users are less likely to quit Duolingo than they are to bail out of competing language apps. “Our retention is comparable to games,” he says.

The comparison is apt. Duolingo grabs users with gamification tricks like points, treasure chests and “streaks” for continuous use. The app’s three-minute lessons are designed with a simple interface. In a typical exercise, the sentence “I eat bread” appears above seven bubbled words. Drag the words onto a line, hit “check” and “You are correct” appears at the bottom of the screen with a satisfying chime. Fans tweet about funny sentences generated by the software, like “I am selling my mother-in-law for a euro.”

Duolingo has gotten bad press from writers who try the app and don’t learn much. But Von Ahn promises only to get users to a level between advanced beginner and early intermediate. “A significant portion of our users use it because it’s fun and it’s not a complete waste of time,” he says.

He’s been logging 15 to 20 minutes of French every day since November, and when asked to describe what he did the previous weekend he says, “Je fais du sport. Je suis mange avec mes amis. Je suis boire du biere en un bar,” mangling his tenses. (Rough translation: I play sports, I am eat with my friends. I am drink beer in a bar.)

Bob Meese, Duolingo’s 42-year-old chief revenue officer, has been studying Duolingo Spanish for more than six months. In response to the question, “¿Hablas español?” he freezes, then says, “Could you repeat that?”

Von Ahn learned English the old-fashioned way at the American School in Guatemala City. His mother, a physician who had him when she was 42 and unmarried, raised him on her own (his surname comes from his father’s German family). He excelled in math, and by the time he was 12 he wanted to be a professor.

He went to America for college and graduate school, earning a bachelor’s in math at Duke and a Ph.D. in computer science at Carnegie Mellon (CMU), where he studied under Manuel Blum, 1995 winner of the Turing Award, known within the field as the Nobel Prize of computing.

Together they created CAPTCHA, an academic project they gave away for free. While he was still a grad student, Von Ahn built a tool that used crowdsourcing to identify image files. He sold it to Google in 2004 for an undisclosed sum. “I didn’t really need to worry about money after that,” he says.

Shortly after earning his Ph.D. in 2005, he got an unexpected call. Bill Gates wanted him to head up a team at Microsoft’s research lab. Gates kept him on the phone for an hour and a half, but Von Ahn turned Gates down.

Instead he took a job as a computer science professor at CMU. Three weeks in, the MacArthur Foundation awarded him a $500,000 “genius” grant for his pioneering work in computer security and crowdsourcing. He used the funds to create a second version of CAPTCHA called reCAPTCHA. The squiggly words it displayed were drawn from old books and newspapers that scanners couldn’t read. In typing the words, billions of people gave away the free labor needed to digitize massive amounts of text. In 2009, Google bought reCAPTCHA for more than $25 million (Von Ahn owned more than 50%), and he spent the next two years as a Google employee on leave from CMU.

“A significant portion of our users use it because it’s fun and it’s not a complete waste of time.”

Luis von Ahn

Back at CMU in 2011, he and Swiss-born grad student Severin Hacker (yes, really) focused on an idea the two had been mulling while Von Ahn was at Google, a free digital language-learning tool. Von Ahn and Hacker named their venture Duolingo because they wanted it to serve two functions. It would teach languages to users for free. It would also rely on those same users to do translations.

First, Duolingo would find customers who needed texts translated. Then it would get crowds of users who were studying English on Duolingo to translate English passages into their native language. If enough users worked on the same passages, the translations came out well.

Raising money was easy. “We were very impressed with what he had done in the past and with him personally,” says Brad Burnham, a Duolingo board member and a partner at Union Square Ventures, the company’s first investor. “We didn’t really try to model the business.”

To create the initial courses, Hacker and Von Ahn read a pile of books on language instruction, including Spanish for Dummies, and built crude basic courses for English speakers to learn Spanish and German. They googled the 3,000 most commonly used words in each language, translated them into English and used the words to compose simple sentences. Then they wrote an algorithm that would spit out lessons that included prompts for sentence translation, listening, writing and speaking.

“We didn’t want the lessons to be handmade,” Von Ahn says. Software could be programmed to respond to users, directing them back to easier lessons if they made mistakes. Early on, they added fun stuff like points and Duo the owl.

Duolingo launched in 2012, and the app started to catch on. But only two customers, CNN and Buzzfeed, signed on for translation services. “It was like a Rube Goldberg machine,” Von Ahn says, because the translations required too many steps to get to the final product.

In 2014 he ditched the translation business, and for the next three years the company had no revenue. But after raising $38 million in VC money, he had more than enough cash on hand to keep plugging away. He hired linguists and second-language-acquisition researchers who added instruction extras like grammar tips and conjugation tables.

Von Ahn says he turned investors away while pitching those he thought would make the best partners. Ashton Kutcher, Tim Ferriss and heavy hitters from Kleiner Perkins and CapitalG (formerly Google Capital) are all Duolingo backers. In 2017 he introduced Google and Facebook ads, followed by ad-free subscriptions. Duolingo ended the year with revenue of $13 million.

Meanwhile Duolingo’s popularity was growing and it was adding new languages at a fast clip. Drawing on his crowdsourcing background, Von Ahn developed a system for vetting volunteer native speakers who contribute vocabulary words and sentences to build new courses.

In 2016, Duolingo started working on another potential revenue generator, the Duolingo English Test (DET), to compete with the TOEFL (Test of English as a Foreign Language), the dominant English proficiency exam for foreign students applying to American universities. Owned by the nonprofit Educational Testing Service, the TOEFL costs $215 and requires students to spend three hours at a proctored testing site.

Von Ahn had his own harrowing TOEFL experience back in 1995. All the seats in Guatemala City were filled, so he had to travel to San Salvador. “El Salvador was a war zone,” he says. “I spent $1,200 to take the test. It was crazy.”

The Duolingo English Test (DET) costs $49, lasts 45 minutes or less and can be taken remotely provided a student’s computer has a working speaker and camera to prevent cheating. More than 180 schools, including Yale, Columbia and Duke, already accept it as a substitute for the TOEFL.

But the DET may be doing foreign students a grave disservice, says Elvis Wagner, a professor at Temple University whose area of expertise is second-language assessment. In 2015 he coauthored an article in Language Assessment Quarterly that says, “DET seems woefully inadequate as a measure of a test taker’s academic English proficiency or for high-stakes university admissions purposes.” He says he was sorry to reach that conclusion, because the test increases accessibility to students. “But there is virtually no comprehension involved and no language production,” he says.

Duolingo vs. The Competition


The test has since improved, Von Ahn says. It now includes open-ended questions that students must answer in writing and a new speaking prompt that requires comprehension. But one section appears to have been inspired by the puzzles Wheel of Fortune contestants solve by filling in missing letters in a written passage. In another section, students must distinguish fake English words like “groose” from real words like “skate.” None of the exercises appear to require college-level English proficiency.

Von Ahn counters that scores on the DET are “highly correlated” with TOEFL scores and “that’s all that matters.” His staff compared results from 2,300 students who took both tests and found that those who did well on the DET were likely to get high TOEFL scores. But TOEFL executive director Srikant Gopal says the correlation is meaningless because the DET has only “rudimentary exercises that bear little resemblance to how English is actually used in academic settings.” Nevertheless, Von Ahn hopes the test will account for 20% of revenue by 2021.

Beyond the English test, he says he is just getting started toward realizing Duolingo’s promise. Next up: adding more challenging lessons to the language courses. “We want to get you to a level where you can get a job, even a high-paying job, using a language you learn on Duolingo,” he says.

But Diane Larsen-Freeman, a linguist at the University of Michigan who is considered one of the nation’s leading experts on second-language acquisition, says most learners can’t get very far without conversing with other humans. “The interpersonal connection really is key to any kind of language learning,” she says.

Von Ahn doesn’t argue. Two years ago, Duolingo created a website where users can organize events at cafes and restaurants, so learners can practice with other people. But with 28 million users spread throughout 180 countries, the current roster of 2,000 events a month, with an average attendance of 10, touches a minuscule fraction of learners.

Von Ahn is unfazed, pointing to the snowball effect of the app’s popularity to date: “We are going to continue with languages until most everybody in the world who is trying to learn a language is using one of our products.”

Duolingo is on Forbes’ 2019 list of Next Billion-Dollar Startups. For the complete list, click here.

Follow me on Twitter. Send me a secure tip.

In February 2018, I took on a new job managing and writing Forbes’ education coverage. I’d spent the previous two years on the Entrepreneurs team

Source: Game of Tongues: How Duolingo Built A $700 Million Business With Its Addictive Language-Learning App

Why Singapore Is So Good At English – Isabella Steger


Singapore keeps getting better at English. The city-state made the top three of an annual ranking in English proficiency conducted by English education company EF Education First (EF), the highest-ever ranking for an Asian nation. Though Singapore has for years ranked near the top of the list, this year, it leapfrogged Norway and Denmark to place behind Sweden and the Netherlands. Minh Tran, the Hong Kong-based co-author of the report who frequently consults on English education for foreign companies……..

Read more:






Your kindly Donations would be so effective in order to fulfill our future research and endeavors – Thank you

The Surprising Link Between Language and Depression – Emily Petsko


Skim through the poems of Sylvia Plath, the lyrics of Kurt Cobain, or posts on an internet forum dedicated to depression, and you’ll probably start to see some commonalities. That’s because there’s a particular way that people with clinical depression communicate, whether they’re speaking or writing, and psychologists believe they now understand the link between the two. According to a recent study published in Clinical Psychological Science, there are certain “markers” in a person’s parlance that may point to symptoms of clinical depression……

Read more:


Your kindly Donations would be so effective in order to fulfill our future research and endeavors – Thank you


Ask a Radical CopyEditor: Are There Limits to Self-Identity Language? — Radical Copyeditor

Every person gets to have full agency over the language they use to describe themself, even if those words are experienced by others as oppressive.

via Ask a Radical Copyeditor: Are There Limits to Self-Identity Language? — Radical Copyeditor




Your kindly Donations would be so effective in order to fulfill our future research and endeavors – Thank you



A Sneaky Theory of Where Language Came From – Ben James


I’m sitting in the sun on one of the first mild days of the spring, talking with a modern-day flintknapper about the origins of human language. His name is Neill Bovaird, and he’s neither an archaeologist nor a linguist, just a 38-year-old bearded guy with a smartphone in his pocket who uses Stone Age technology to produce Stone Age tools. Bovaird has been flintknapping for a couple decades, and as we talk, the gok gok gok of him striking a smaller rock against a larger one punctuates our conversation. Every now and then the gokking stops: A new flake, sharper than a razor blade, breaks off in his palm.

I’ve come to see Bovaird, who teaches wilderness-survival skills in western Massachusetts, because I want to better understand the latest theories on the emergence of language—particularly a new body of research arguing that if not for our hominin ancestors’ hard-earned ability to produce complex tools, language as we know it might not have evolved at all. The research is occurring at the cutting-edge intersections of evolutionary biology, experimental archaeology, neuroscience, and linguistics, but much of it is driven by a very old question: Where did language come from?

Oren Kolodny, a biologist at Stanford University, puts the question in more scientific terms: “What kind of evolutionary pressures could have given rise to this really weird and surprising phenomenon that is so critical to the essence of being human?” And he has proposed a provocative answer. In a recent paper in the journal Philosophical Transactions of the Royal Society B, Kolodny argues that early humans—while teaching their kin how to make complex tools—hijacked the capacity for language from themselves.

To understand what Kolodny’s getting at, I ask Bovaird to walk me through the history of Stone Age technologies. He starts by smashing an irregular, grapefruit-size stone between two larger rocks. He picks through the resulting fragments, looking for a shard with an excellent cutting edge. This is simple Oldowan technology, he tells me—the first stone tools, used by our hominin ancestors as far back as 2.5 million years ago.

Next, he flashes forward a million years to the technological revolutions of Homo ergaster. No longer did toolmakers simply knock stones together to see what they got; now they aimed for symmetry. Bovaird holds up his work in progress, a late Acheulean hand ax—the multi-tool of the middle-to-lower Paleolithic, good for cutting meat, digging dirt, smashing bone. The blade of this ax has a zigzag edge, with tiny, alternating flakes removed from each side of the cutting surface. To achieve this level of serration, Bovaird explains, he needs a precise understanding of how the stone works, as well as the ability to plan his work many steps in advance.

Somewhere on the timeline between the long run of the Oldowan and the more rapid rise of Acheulean technologies, language (or what’s often called protolanguage) likely made its first appearance. Oren Kolodny and his co-author, Shimon Edelman, a professor of psychology at Cornell University, say the overlap is not a coincidence. Rather, they theorize, the emergence of language was predicated on our ancestors’ ability to perform sequence-dependent processes, including the production of complex tools.

Kolodny’s arguments build off the groundbreaking experiments of Dietrich Stout, an anthropologist at Emory University. A flintknapper himself, Stout has taught hundreds of students how to make Acheulean-era tools, and he’s tracked their brain activity during the learning process. Stout found that his students’ white matter—or the neural connectivity in their brains—increased as they gained competence in flintknapping. His research suggests that producing complex tools spurred an increase in brain size and other aspects of hominin evolution, including—perhaps—the emergence of language.




But language couldn’t just pop out fully formed, like Athena from the head of Zeus. “Every evolutionary process, including the evolution of language, has to be incremental and composed of small steps, each of which independently needs to be beneficial,” Kolodny explains. Teaching, he says, was a crucial part of the process. When hominins like Homo ergaster and Homo erectus taught their close relatives how to make complex tools, they worked their way into an ever more specialized cultural niche, with evolutionary advantage going to those individuals who were not only adept at making and using complex tools, but who were also able—at the same time—to communicate in more and more sophisticated ways.

Kolodny points out what might seem like a contradiction in this notion: Many species of ape use tools in sequence-dependent ways and also have highly developed levels of communication. But the order in which those apes produce their utterances doesn’t make much difference to their meaning, Kolodny explains. “The question becomes not ‘How did language arise only in humans?but ‘Why did it not arise in other apes as well?’ And the answer is, the qualitative difference between us and other apes is they don’t have the communication system coupled to those temporal sequencing structural capabilities.”

That “coupling” is where the hijacking comes in. The technical term is exaptation, a word coined by the late paleontologist Stephen Jay Gould to describe an evolutionary event in which a biological function is repurposed for an alternate use. Kolodny and Edelman argue that the neural networks required for complex, hierarchical, sequence-dependent tool production were exapted by our brain’s communicative apparatus, which is why word order and sentence structure make such a difference to meaning. Rudimentary language, which evolved in the context of toolmaking and teaching, was ultimately able to break away from its immediate contexts—this is the hijacking part—eventually employing those original cognitive pathways for its own unique purposes. The result, as Monty Python viewers have appreciated for decades, was our modern, turbo-driven faculty for language.

In order to go along with Kolodny and Edelman’s theory of linguistic hijacking, you have to agree that language-related structures (like syntax) and action-related structures (like the sequential steps to making a tool) are similar enough to be driven by the same neurological mechanisms. But many scientists—including Noam Chomsky, the most influential of all modern linguists—aren’t willing to swallow that pill.

Chomsky has been notably reticent on the subject of language evolution. On numerous occasions, he’s called the question either irrelevant, unsolvable, or both. A surprise came in 2014, when Chomsky, Robert Berwick, and other titans in the field weighed in substantively on the topic of evolution for the first time, arguing in a series of jaw-dropping papers that language basically did show up on the scene like a fully formed Athena, syntax-driven shield in hand. “The language faculty is an extremely recent acquisition in our lineage,” these authors wrote, “and it was acquired not in the context of slow, gradual modification of preexisting systems under natural selection but in a single, rapid, emergent event.”

Berwick, a professor of computational linguistics at MIT and the co-author with Chomsky of Why Only Us: Language and Evolution, sees little merit in the stone-tool-based theories proposed by researchers like Stout and Kolodny. “The Stout business doesn’t work,” Berwick wrote to me in an email. “The experimental findings show that, to the contrary, verbal language DOES NOT FACILITATE toolmaking.” Berwick calls the purported connection between toolmaking and language a useful metaphor, at best.

Chomsky’s position is such a brazen refutation of known evolutionary processes that Kolodny, Stout, and many of their colleagues aren’t sure how to engage with it. Stout claims that Berwick’s refutation of his research misses the mark entirely. Ultimately, Stout tells me, he expects the positions of Berwick, Chomsky and other formalist linguists to find a sort of synthesis with his own views on language, but he doesn’t see agreement occurring anytime soon.

Luckily, the flintknapper Bovaird and I don’t need a synthesis of theories of language evolution in order to sit in the sun and talk. Birds are singing, a woodpecker raps its head against a dead trunk, and Bovaird is “pressure-flaking” a fine, serrated edge on a symmetrical obsidian blade. He’s way past the Acheulean now, demonstrating the refined techniques used by the essentially modern Homo sapiens of the upper Paleolithic. Picking up a rounded tool made of antler—he’ll use it as a hammer—he compares flintknapping to a sophisticated game of chess, in which the order of every move is of supreme importance. It’s an observation that either has everything to do with how language got here, or—if you’re in a Chomskyan frame of mind—has nothing to do with language evolution at all.

“The stone has rules, and you can’t break the rules,” Bovaird says. Then he leans over the blade, flexes his wrist, and strikes.


If everyone who read the articles and like it, that would be favorable to have your donations – Thank you.



%d bloggers like this: