How I Built A $6.5 Billion App Called Duolingo | Founder Effect

0:00 carnegie mellon is an excellent university but it is a stressful place people

0:05 aren't happy they're just they kind of just it's not good it's not

0:10 a happy place one of the things that severan and i decided early

0:13 on when we were starting a company is like look whatever happens you

0:17 know our company should be a happy place by the time luis van

0:20 on turned 24 he was already a millionaire several times over the 43

0:25 year old may not be a household name but i'm willing to bet

0:27 you're one of the hundreds of millions of people who use this technology every

0:30 day lewis isn't your average unicorn tech founder he actually pays his drivers

0:35 to give feedback on their interactions

0:37 with potential executive level hires on their way

0:39 to and from the airport to weed out toxic personalities alright luis

0:44 thank you so much for taking some time out i appreciate it yeah

0:47 of course there are three numbers to look out for in lewis's story

0:50 42 000 the amount he made each week digitizing copies of the new york

0:54 times 183 million the total amount of outside investment he raised and six

0:59 and a half billion the total valuation after duolingo went public in june 2021.

1:05 here's how louis van on built duolingo one of the most popular educational apps

1:09 in the world while managing to keep it free for almost everyone who uses it

1:13 for cnbc make it i'm nate skid this is founder effect lewis grew up far

1:20 from the ivy leagues in guatemala his mother

1:22 was a doctor and made sure he learned

1:24 english at a younger age for many people that can be the difference between

1:27 a life of struggle and one of opportunity how big of a deal do you think

1:31 it was to go to an english program in terms of setting on the path

1:34 that you ended up that you ended up on i think it was a huge deal

1:38 in guatemala for example you could probably

1:40 double your income potential by just the fact

1:42 that you know english but you don't have to need it no anything else there

1:44 were two formative moments in lewis's young

1:47 life the first was witnessing the tension

1:49 in his family's candy business between the owners

1:52 and their workers different people in my family would

1:55 you know a lot of times they just have this vision that it's like kind

1:58 of us versus them the second was

2:00 a visit from a recruiter from duke university who

2:02 was scouring central america for undiscovered academic

2:05 talent and she essentially kind of she didn't

2:08 quite fill out the application for me

2:10 but she almost filled out the application for me

2:12 in 1996 lewis moved to the united states to attend duke university with no money

2:16 to his name yet he still managed to graduate at the top of his class

2:20 with the goal of becoming a math professor but that dream didn't last long i

2:24 realized that all the professors that were

2:26 in math were doing research on problems that hadn't

2:29 been solved for 500 years or whatever lewis wanted to spend his time and energy

2:33 tackling new challenges in 2000 he was accepted

2:36 to a computer science phd program at carnegie

2:38 mellon but it didn't take long for him

2:40 to develop a knack for creating profitable

2:42 businesses in 2003 he created a simple

2:44 game pairing two players and showed them each

2:46 the same image if their descriptions matched they moved on to the next one what

2:50 they were doing is basically just telling

2:52 google what's in these images um and so

2:55 that that really you know kind of improved

2:57 image search etc lewis says google bought

2:59 the game in 2003 for a couple million dollars in 2006 lewis landed on his next

3:04 big idea after listening to a talk

3:06 by yahoo's chief scientist the problem was that spammers

3:09 were writing code to steal millions of email

3:12 addresses and flood those inboxes with junk

3:14 mail lewis's answer was called this thing

3:17 called the captcha which is these distorted characters

3:19 that you have to type um you know all over the internet whenever you're buying

3:23 tickets from ticketmaster or whatever you just

3:26 you know you get this image of messed

3:28 up characters um so we came up with that that was our idea about 200

3:33 million people take 10 seconds out of their day to fill out a captcha

3:36 and while some would sit in amazement

3:38 at their impact on humanity lewis suffered from pains

3:40 of guilt which led to his next big idea and so if you multiply 10

3:44 seconds by 200 million i started thinking okay that's that turns out to be 500

3:48 000 hours every day started thinking okay can we can we make good use

3:51 of these 500 000 hours gave to this rise to this kind of next project which was

3:55 called recaptcha so this is like a redoing of captcha um where the idea was

4:01 that as people were going to they were typing these you know over the internet

4:06 not only would they be authenticating themselves

4:08 human but they were helping us to digitize

4:11 books where did the new tech reach the new york times which was in the process

4:15 of digitizing about 150 years worth of old

4:17 newspapers lewis charged the times 42 thousand

4:20 dollars for every year of content he digitized

4:23 we could digitize an entire year of content

4:25 in about a week so pretty quickly we started getting checks for 42 000 bucks

4:29 like you know about one a week lewis founded recaptcha in 2006 and sold it

4:34 to google in 2009 for an undisclosed sum but he said it was in the tens

4:38 of millions of dollars in 2006 lewis

4:41 was awarded the macarthur fellowship also known

4:43 as the genius grant that came with five

4:45 hundred thousand dollars and no strings attached

4:47 it's not like you apply for it or anything just one day you get a phone

4:51 call and they just ask if they're fortunately i picked up the phone because

4:57 you know nowadays if i get a random phone call i do not pick up

5:00 the phone so what did you do with the five hundred thousand dollars put

5:03 in the bank account uh uh honestly i

5:06 probably spent it mostly on a little seed funding

5:12 for this recapture so um where did the aha moment for a language service come

5:22 about where did this happen yeah that was so i was that was around 2009 2010

5:28 um i had sold recapture to google i had a phd student named severin hacker who

5:34 is my co-founder at duolingo at the time

5:36 we hadn't started anything one of the insights

5:38 was you know computers are getting much smarter and we could make it so

5:42 that computers really could teach everybody as opposed

5:44 to teachers having teach everybody that was kind

5:46 of the idea now that they knew they wanted to teach they just needed to agree

5:50 on a subject eventually we settled on teaching

5:52 languages and the reason for that was

5:55 because both of us have you know both of us learned english so we

6:00 thought okay let's do something to teach

6:02 english the other thing that we really wanted

6:04 to do was we really sought technology as a way to to be able to really

6:10 democratize education beautiful thing with technology is

6:13 that it doesn't cost you that much more

6:14 to teach more people than just to teach one person um so we thought okay

6:18 well we teach everybody and and we can teach them for free and just like

6:22 that duolingo was born well sort of so um how do you come up

6:26 with the name duolingo we looked at a lot of names one of the ones um was

6:34 f-l-o-o-n-t which should sound kind of like

6:36 fluent but it more sounded like fluent

6:39 and then you know my friend said oh that sounds like i flunked it all over

6:42 the floor like it's like not good um so we had we had a bunch

6:46 of names eventually we came up with not

6:48 duolingo but monolingual and and that sounded like

6:52 a like an illness like you have monolingual and at some point just duolingo made

6:57 a lot of sense now that they had a name and a mascot it was

6:59 time to turn their idea into a business instead of applying for a grant through

7:03 carnegie mellon in 2012 lewis reached out

7:06 to union square investors and secured three million dollars

7:09 in seed funding they had just invested in like twitter and tumblr and they were

7:13 like the biggest thing out there and foursquare

7:15 was also the biggest thing out there

7:16 and so unesco advantages was like oh my god like amazing okay so um can

7:21 you tell me the amount that series a from union square was yeah i mean one

7:25 thing that is important to mention is series a back then this is the year

7:28 2012 we're very different than csa's today

7:30 which today is just massive so in 2012

7:33 a very nice series a that you were happy with was three million dollars three

7:39 million today that's not even called a series

7:41 eight today's like seed funding around that time

7:44 lewis gave a ted talk and at the end he made mention of this really

7:48 cool new application focusing on language that he

7:50 was working on well that talk went

7:52 viral and soon duolingo which at the time was just a landing page with a place

7:58 to put an email address had a waiting list with over 300 000 names

8:02 on it at the time the other thing that was going on at the time

8:04 was uh there wasn't really a good way to learn a language on the computer i

8:08 mean the thing that there was was rosetta stone and it was like super expensive

8:11 it was like a thousand bucks and so there was this thing that just you can

8:14 learn a language here it's entirely for free and so a lot of people were

8:17 like yeah sure i'll give you my um my email um and so you know

8:21 that worked out pretty well the instantaneous interest

8:24 in duolingo and his proven track record helped

8:27 lewis raise even more capital 183 million

8:30 in all he used almost all of that early

8:32 investment money to build out a team and for the next three years he focused

8:36 solely on growing his user base he

8:38 didn't even think about monetization up until 2017-ish

8:44 duolingo was making no money this was uh not it was our finances were very

8:48 simple simply we spent money on mainly

8:51 people's salaries and that was at that point

8:55 though we decided it's probably time to make

8:57 duolingo be a self-sustaining business and we

8:59 started actually monetizing and it has worked

9:02 out very well by now lewis says duolingo

9:04 had about 10 million active users and was

9:06 the number one education app in the world

9:09 now they just had to figure out how to make money while keeping the app

9:12 free we didn't just want to say you know turn around and say oh just

9:15 kidding um now you gotta pay so what we ended up doing is we ended

9:20 up coming up with a business model that ends up being pretty similar to say

9:23 what spotify does or what the dating apps do which is um you can use

9:27 duolingo as much as you want for free but uh um if you don't pay

9:31 us you have to see some ads at the end of a lesson and then

9:35 if you want to turn off the ads you can pay us to subscribe and then

9:38 we turn off the ass and we may give you other kind of premium features

9:41 so that that combo of ads and subscription worked out really well um and so

9:46 we ended up making a you know every year since then we've made more and more

9:50 money a full 94 of duolingo's active

9:52 monthly users opt for the free version which

9:55 includes some ads but the company makes most of its revenue from the other six

9:59 percent of its users who were paying

10:01 subscribers from six percent of our users give

10:04 us the majority of our money by now there are more people in the u.s

10:07 learning languages on duolingo than there are

10:09 students learning languages in all u.s high schools

10:11 combined and one of the reasons for duolingo

10:13 success is that it feels like a game

10:15 in fact the app keeps track of how many consecutive days a user logs

10:19 in skip a day and it goes to zero we have over by by now i

10:24 mean we haven't quite released the the figure

10:27 but we have released this one which

10:28 is we have over a million daily active users who have a streak longer than 365.

10:36 so we have more than they haven't missed a single day in the last year

10:43 what day did you ipo what was it like for you personally were you nervous do

10:47 you remember the moment it was extremely exciting

10:49 and it's a big milestone for the company and for everybody who has been working

10:52 on this i mean duolingo has really good employee

10:55 retention as in like people really rarely leave

10:57 duolingo so most of the original team is

10:59 still here and so there's been all these people that have been at this for you

11:02 know the last i don't know eight nine years um so it was pretty transformative

11:06 what happened to the share price our share price was 102 um it went it went

11:12 really high i mean uh uh the first trade was 140 i don't know maybe 141

11:18 or something like that some i don't know the exact number is around 140 and then

11:22 it just kept going up et cetera you don't know like that would be like

11:25 plastered on my wall is like a big memory you know share price is uh i'm

11:31 i i i was told by a lot of ceos of public traded companies not

11:36 to pay too much attention to share price

11:38 and i've been doing that and it's actually

11:39 really good feedback here's basically what your share

11:42 price moves randomly with like basically no connection

11:47 to what's going on with the company what

11:48 was the biggest money mistake you've made along

11:50 the way with duolingo i don't feel bad about anything we've done by the time

11:55 we went public we still had a hundred

11:57 and some million dollars in the bank account

11:59 meaning we had only really spent 80 million of course we had been making some

12:03 revenue the last few years etc but basically

12:05 we could have raised a lot less money

12:08 and by raising a lot less money you know i think um boss employees me

12:12 and the rest of the employees would have

12:14 owned a larger fraction of the company when i

12:22 was talking to the management team about uh

12:24 interviewing you uh one of the other senior

12:26 producers said that she had gotten very far

12:29 um in the in the interview process at duolingo

12:31 you guys flew her out to pittsburgh you put her up and she said

12:35 that even though she didn't get the position

12:37 the culture and the vibe at duolingo stayed

12:40 with her and she was like nate you have to ask about that culture and so it

12:46 dawned on me when you were telling me

12:48 about this candy factory that you watched and then

12:52 hearing about the culture that you created

12:53 and i'm wondering if you can kind of sew

12:55 that up for me like what you learned there and what you apply now most

12:58 tech companies there's a lot of employee churn

13:00 meaning like people leave the company etc very

13:03 few people leave dual language because it's

13:04 a good workplace and i think that there's two

13:06 reasons for that i think the you know what i what i saw with with my family

13:10 and this is not it's not that my family was doing anything bad or anything

13:14 it just in in a in a country like guatemala uh there really is a kind

13:20 of a boss versus employee like us versus them kind of thing and and i

13:24 saw that that really didn't work i mean

13:26 it's much better when when everybody is much

13:28 more egalitarian culture so duolingo has a you

13:31 know in as much as possible a very

13:32 egalitarian culture that's one thing the other

13:34 thing is you know when when we were

13:36 at carnegie mellon carnegie mellon is an excellent

13:38 university i have nothing bad to say about

13:40 it it is really an excellent university

13:42 um excellent for artificial intelligence for all kinds

13:44 of things but it is a stressful place

13:49 and people when you enter the buildings there

13:52 you people aren't happy they're just they're kind of just it's not good it's not

13:57 a happy place um and uh you know one of the things that severan and i decided

14:02 early on when we were starting a company

14:03 is like look whatever happens uh you know

14:06 our company should be a happy place lewis

14:08 has a unique way of weeding out potentially

14:11 toxic employees i'll tell you some of the things we've done by the way even

14:14 for executives um whenever we fly an executive

14:18 for an interview or or not just accept it

14:21 for a lot of people whenever we fly them uh um we have a driver

14:24 go pick them up in the airport and we have a set of drivers that are

14:27 the same everywhere we actually that's part

14:30 of your interview and people don't know it

14:32 um it is how you treat them and so we we get feedback from the drivers

14:37 about how well they were treated and so now normally that most people are just

14:42 perfectly fine like just like that but we have we have not made offers to very

14:46 very um qualified competent people because they were

14:49 nasty to our driver and well uh we

14:51 don't like that because that just means you're

14:53 gonna um you know you're gonna be nasty

14:55 to the little people and we don't want that and so so yeah i think

14:59 that that type of stuff has really helped

15:02 lewis businesses have been incredibly successful and yet

15:06 they all seem to serve a greater purpose

15:08 the picture matching game had the added benefit

15:10 of generating seo terms captcha helped yahoo

15:13 and many many other digital businesses decipher between humans

15:16 and robots and recaptcha is helping to digitize

15:19 the world's books these ideas have made

15:22 lewis incredibly wealthy but he's most proud

15:25 of the culture he created at duolingo a lot

15:27 of the people that come work at duolingo

15:29 do so because they love this mission of you

15:31 know developing the best education in the world

15:33 and making it uh you know universally available you