Freemium In Education

We’ve been investing in the education sector for a few years now. We started our exploration of online education in early 2009 with an event called Hacking Education. The takeaways from that event have informed a lot of what we’ve invested in since then.

One of the key takeaways was that learning could and should become free. Our friend Bing Gordon said this at Hacking Education:

From an economic point of view, I would say the goal… is to figure out how to get education down to a marginal cost of zero. 

We have invested with Bing in online education. Bing and his partners at KP led the most recent round in our portfolio company DuoLingo. DuoLingo is the most popular language learning mobile app in the world. And one of the reasons for that is that DuoLingo is free.

So you might ask “how can you make money giving away a learning app?” This past week DuoLingo answered that question with the commercial release of the DuoLingo Test Center.

The DuoLingo Test Center is currently free but it won’t be for long. Give it a try if you’d like to see how it works. Once the DuoLingo Test is accepted at schools and employers, the company plans to charge $20 to take its test.

There’s an established incumbent (monopoly) in this market called TOEFL. If you’ve come to the US to study, you’ve probably taken this test. It’s a lot more expensive than $20 per test and DuoLingo is out to prove it can do this testing less expensively and better.

But what this example shows is something more than how one company plans to monetize its free app. It’s a model for freemium in online eduction. Provide the education for free but charge for the certification (testing). This is a very elegant implementation of freemium as its an easy on ramp and the customers who get the most value are the ones who pay.

I am pretty sure this will become the dominant monetization model in online education. We are already seeing it emerge in other sectors. A number of the attendees at Hacking Education predicted this over five years ago. It made sense to me then and it makes even more sense to me now.

#hacking education

Comments (Archived):

  1. JimHirshfield

    Schools and employers pay the $20, or they pass it on to the student? Or case by case?Either way, A+ strategy.

    1. fredwilson

      the student pays

      1. JimHirshfield

        I could see companies paying for their employees if the education is relevant to their job.

        1. fredwilson

          yup. i can see that too.

          1. William Mougayar

            There’s a difference between paying for education and paying for a certification test. Sometimes they are related, but not always.

          2. awaldstein

            How are they ever different in intent?If certs are the gateway to anything then they are by definition connected to the education that lead up to them.

          3. William Mougayar

            yes, but a different entity can administer the test. it’s not always the same entity that does both. granted, universities do both.

          4. LE

            One problem with charging just for testing is that it’s much easier to create a test than it is to create a learning platform. I could probably go to any university and get phd’s to create a test but I couldn’t easily create or complete with a well done interface and learning platform.

          5. awaldstein

            Not really. The test is not the issue, the value of the score or cert is and that is an expensive process. Create standards is tough, expensive and iffy.

          6. LE

            The difficulty of something and whether it is actually appears difficult are two different things. “Shock and awe” prevents many people from starting down a particular path. Something that appears easy-er will invite more people to try. Doesn’t matter if they are right. [1]On the surface making up a test seems easier.Having to get the test accepted of course is an entirely different process.Getting a good piece of real estate in NYC to open a wine store (knowing nothing about wine!) is probably very difficult. But to me it certainly seems easier and on the dartboard of trying vs.,say, competing in the space program (I’m using a wild example to illustrate a point).[1] You could apply the same logic to creating the actual site of course I just think that the test appears to be lower hanging fruit.

          7. awaldstein

            i have absolutely no idea what you are saying here LE 😉

          8. LE

            You said “The test is not the issue, the value of the score or cert is and that is an expensive process” which to me roughly translates to “it’s not easy it’s actually hard”.My reply said “The difficulty of something and whether it is actually appears difficult are two different things.” and as such something that appears easy will invite competition while something that appears hard (shock and awe at the complexity) will not. No question that people get turned off by perceived difficulty of success.

        2. William Mougayar

          I think that’s a slightly different segment than the one that Dulingo is targeting.Employers who want to improve their employees’ English language so they can improve their performance in their current jobs will probably use services like from (I know a bit about that because they used to be an Eqentia customer)

          1. awaldstein

            If you don’t know Openenglish for Spanish speakers wanting to learn English, you should.Founder is taking video coaching developed for that and moving the paradigm into wellness.

          2. Richard

            Looks like they are doing well. Is there anything unique about their teaching methods?

          3. awaldstein

            I know less about their language teaching methods and more about how they mashed up existing tech to create a vast community of tutors and students with good results.Huge business success.Know a lot more about how one of the founders, LA based btw, is moving remote coaching into the wellness space. Can’t reveal any details.

          4. William Mougayar

            Interesting model. Thanks for pointing it out.

          5. JimHirshfield

            I was referring in general terms to the edu-freemium model. So, as an example, think basic accounting course instead of language or Duolingo.But I agree with you more than you know @jlm agrees with you.

          6. LE

            Makes a lot of sense. You offer companies both a learning platform (for free) but more importantly a standardized way to see if their employee actually have the skills in a variety of areas.So if I want to take a person who is an admin assistant and have her learn bookeeping/accounting I point her to the site which both teaches her and validates that she has a basic skill set.Even better yet, you can slice and dice thatdown to all sorts of granularity each time picking up a few bucks for a particular skill set. [1] So instead of saying “accounting” you focus on a slice of accounting but you also lower the test fee for that slice and you make it up on a larger volume of potential knowledge areas. And there could be 20 slices per subject. Once General Electric accepts your test platform their employees can run loose and learn all sorts of things that are validated by the test provider/site.[1] And making it a much easier mountain to climb with reinforcement for the employee.

          7. JimHirshfield


        3. LE

          My wife just bought (two weeks ago) Rosetta Stone [1] to learn spanish because she said she needs it for her job (some poor patients speak spanish and they normally use translators). First question I asked her was “will work pay for this”? She said they won’t and she didn’t want to rock the boat (negotiating a pay raise currently).An owner of a small company could easily make a decision to pay for this on a case by case basis. A larger organization not the same (right?).[1] Wow I’m seeing that on the website rosetta stone is $274. Reduced from $499.

          1. fredwilson

            it would be interesting if she tried DuoLingo for her spanish. i have heard many say its way better than Rosetta Stone and the price is right too!

          2. LE

            What I am going to do is send her the link.First thing I will see is how she responds to that (after paying the money).She might feel that she needs to justify that she just paid for something (and won’t want to feel stupid).The next thing I will do is see her reaction to the website. I sent her another website mentioned on AVC (in the comments) and it failed the “puny brain” test. (We discussed that last night at dinner.) (I don’t think that is going to happen with duolingo btw.)Next I will see if she actually tries duolingo.I don’t want to tell her why I want her to I want observe how she reacts to the suggestion because most people just get things suggested casually. And they either take to it like a fish to water or they don’t. Because I’m always curious about how people make decisions. If any interesting happens worth noting I will report back,

          3. fredwilson

            The perfect test

          4. JimHirshfield

            Tax write off maybe?

          5. LE

            Hmm. Excellent point. I will get the receipt and write it off on my business expenses. (We also have spanish customers..)

      2. LE

        Student paying $20 means they have other options if other options exists. Schools paying $20 (actually less) means that a competitor has to sleep with the purchasing agent in order to get their test accepted. (Where PA = “decision maker”). Schools paying is a bigger barrier to competition where a contract can be signed (to the person with the first foot in the door) based on some kind of exclusivity. (Think cable monopoly..) [1][1] Thought offered w/o knowing much about why schools would or would not pay obviously…

  2. pointsnfigures

    Chicago LEAP Innovations is starting up:… I agree, education needs to be unbundled. We have to ignore the rules like Rob Kalin did. I foresee a huge impact being made with co-working spaces like Combine those, driverless cars, mobile, and you can change the entire rhythm of human life. We have been bound to 9 to 5. What if we change that too?

    1. awaldstein

      Unbundling unnatural rhythms of life.I’ll wear that t shirt.

      1. pointsnfigures

        Trying to live that life….

    2. LE

      We have been bound to 9 to 5. What if we change that too?I think we have moved away from that in many areas already. And I think in more than low hanging fruit situations. [1]There was a State Farm commercial that I saw that is based upon the fact that someone is talking to State Farm at 3am and his wife thinks he is talking to a woman:…It’s funny and creative. I remembered it.Anyway what in particular do you see that is bound to 9 to 5 that can be moved away from that (and specifically hasn’t been moved away already?). Time is a demarcation point for things it even occurs on Fred’s blog. No point in making a comment at 3am most activity occurs in the AM or a bit later based on when Fred makes a post.[1] Would also offer that this gives a huge advantage to larger organizations and disadvantages smaller companies, startups or otherwise or can’t staff 24/7 and/or allow employee flexibility. A big “sucks to be them” situation.

      1. pointsnfigures

        right now, the shift is happening in a trickle. I am thinking more about two career families and childcare. Daycare can cost $17k per yr or more. Splitting the daylight hours, working out of a co-work space allows one parent to be with their children. Or, have day care combined with co-working like they do at Nextspace Potero Hill. Lots of permutations-but by 2020, 45% of the world wide work force will be independent and unattached.

        1. LE

          I know people do this apparently but I can’t wrap my head around doing work and having kids around. It’s a big distraction.Impossible for me to believe also that the impact on productivity doesn’t approach the cost of child care. Further (and once again I know this is being done as you mentioned) having onsite childcare in any type of close proximity seems to be a big distraction as well. (Of course from an employees point of view it’s a bonus. ) A total zone killer. Now of course if you work in a job at the local postal office where you don’t have to think and/or are an automaton it’s fine and probably a good thing, right?I can tell you that I wouldn’t even come close to wanting something like that and neither would my wife at her job.Anyway the purpose of work is work. If you can’t do work because you have kids then maybe you have to reevaluate the job you have chosen to do if you can’t afford childcare.Curious, back when you were trading how would it be if your kids were around? I would imagine it would probably mess up your thought process a bit.

        2. LE

          Splitting the daylight hours, working out of a co-work space allows one parent to be with their children.By the way I’m a little unclear here what you were saying. Are you saying that because child care costs 17k one parent stays home and works from home or the parent stays home and doesn’t work from home?And if a parent is in a cowork space how does that change things or matter? Are you saying there is childcare at the cowork space?

  3. takingpitches

    Curious as to where the 30 million test takers a year for toefl comes from? At least based on the toefl site, 27 million people have taken the test in its entire history. Does each person have to take it multiple times because of high failure rates or something else I am missing?

    1. fredwilson

      where do you see that number?

        1. fredwilson

          hmm. great question

    2. Saikiran Yerram

      TOEFL is usually taken by foreigners who apply for schools in the US. I am from Mumbai and I took the test at a test center in Mumbai. This was 8 years ago, so I am assuming today one can take it online.One other thing to note that when you apply for student visa (for all english speaking countries), you’re required to show the TOEFL scores.

  4. William Mougayar

    With video, this should make remote testing possible and easy.I guess the trick will be to have the thousands of institutions accept the Duolingo test as being equivalent to the TOEFL.

    1. fredwilson

      that’s what they have to accomplish

      1. pointsnfigures

        wonder if there is a way for them to work with It’s their target market.

    2. Saikiran Yerram

      I know Verificient tech (Kaplan Techstars alum) provides remote proctoring services for this type of scenarios. Its pretty cost effective way of administering tests.

  5. awaldstein

    Whenever freemium finds a model that adds value to the offering rather than default to advertising, which adds little to anything, I’m all in.I applaud this both in intent and in a model that in theory to me, could work and make sense.

  6. Richard

    I like the approach (and the MOOCs are using it presumably effectively). One of the problems for duolingo is that unlike Academia has it does not have the built in legacy honor code, I could see this being a problem.

  7. Samujjal Purkayastha

    Freemium only works for consumer education companies, when you are targeting students or their parents. But here’s the problem, parents and students in US don’t like to spend money on education, except for higher ed costs. So your market potential is quite small. More money is in selling directly to the institutions, where freemium doesn’t work as evident from the failures of the likes of Edmodo (K12) and Coursekit (higher ed), to name just two. And folks like Duolongo, like the MOOC providers will struggle for credibility against the incumbents. And convincing consumers isn’t good enough, if universities don’t accept scores from Duolingo, which has actually led to the high barrier to entry in the market and led to the monopoly or rather duopoly of TOEFL and IELTS. Same goes for MOOCs. Unless employers (beyond the tech world) and post-graduate institutions accept the credentials, consumer interest will eventually die down.

    1. Richard

      There is one other weakness, unlike STEM skills, language skills are pretty easy to evaluate without a written test. Moreover, Duolingo has to first show that its testing correlates with some skill set.

    2. Jessie Arora

      @samujjal:disqus- Completely agree and think when talking about startups + education it is essential to make the distinction between K12 vs. consumer (or rather informal learning.) It often all gets lumped together creating a false sense of ‘success’ or ‘effectiveness’ for edtech when in reality the impact of tech on K12 schools will continue to be painfully slow.

  8. Elisha Tan

    Accessibility and monetization have fairly obvious solutions, I’m curious to find out what solutions are there for tougher problems like:1. How do we gauge/evaluate if someone has learned something, other than using tests?2. MOOC has solved the problem of accessibility but it has a terribly low completion rate. How can we make people stick?

    1. Jessie Arora

      @elishatan:disqus “MOOC has solved the problem of accessibility ” – this is a gross over simplification. MOOCs can be helpful to super-motivated students, but it’s important to not they are one modality that serves one learning style for the self-starter, connected student.

  9. Itamar Cohen

    It’s ironic (And somewhat discouraging) to see that a revolutionary approach to education, such as the one presented above, focuses its monetization strategy on “the certificate” – The root of all that’s backward in education today.

  10. Guy Friedman

    Great in theory – but very hard to execute “free ed/cheap certification” – it’s doable but will require really targeted investment in educating companiesIn other words it can’t just be consumer driven (let me put this on my resume!) the TOEFL is a really strong brand.I think Duolingo has great technology and team looks really good – but it will be a big lift to get this idea to scale…

    1. Vasudev Ram

      >In other words it can’t just be consumer driven (let me put this on my resume!) the TOEFL is a really strong brand.Yes. So is IELTS.

  11. kirklove

    DuoLingo is a flat-out beautiful app/experience. Want to say they nailed it, yet they continue to improve it 😉

  12. Andrew Kennedy

    Great post. I love it.

  13. Elia Freedman

    Would DuoLingo’s model be possible without venture funding? Could they have afforded to go so free for so long without it?

    1. LE

      Unless money grows on trees most certainly not.

    2. fredwilson

      i doubt it. but that’s good news for VCs. entrepreneurs still need us!!

      1. Elia Freedman

        That’s what I thought. So… can freemium really be done without VC funding? I’m worried about the time to revenue and if we aren’t funded then our window is relatively small unless we use consulting or some other revenue stream to keep us going.

        1. LE

          You can’t compete in a market where competitors are funded and you aren’t funded. To much of a handicap.unless we use consulting or some other revenue stream to keep us goingForget that. That’s fine if that is the business you want to be in. But you need to be firing on all cylinders (with plenty of fuel (money)) in order to make something work. You can’t supplement money that is growing on trees for others with lesser money that you have to actually earn (which takes time, right?) in order to compete with them. No funny money no dice.

          1. Elia Freedman

            Exactly what I was thinking.

          1. Elia Freedman

            Shareware is not freemium. They are two different models.

          2. Vasudev Ram

            I know they are not, but aren’t they somewhat similar? What do you see as the difference?

          3. Elia Freedman

            Similar from the perspective that something is free. Traditional shareware was completely free and pay if you had a heart. Trial software was known as shareware, too, but was time limited. Freemium isn’t time limited but feature limited instead.

          4. Vasudev Ram

            Not sure about that. Shareware had/has various models (I’ve used many shareware apps over the years) – not just time limited; it also had feature limited, nagware, etc. In fact many popular shareware apps were not time limited.

    3. Vasudev Ram

      A wise old economist once said “There is no free lunch.” – So said my high school economics teacher in one of his classes. IIRC he was quoting Galbraith or Keynes or one of the other famous economists of the time.Alternative source: TANSTAAFL:…So someone or the other pays for it, in this case, VC, until it makes money.

  14. LE

    Once the DuoLingo Test is accepted at schools and employers, the company plans to charge $20 to take its test.It may have to be cheaper. Not because $20 is a lot of money but because at $20 price it will invite competition. (You can argue that that is good or bad of course but in this case I think the bad outweighs the good.)

  15. LE

    If you’ve come to the US to study, you’ve probably taken this test. It’s a lot more expensive than $20 per test and DuoLingo is out to prove it can do this testing less expensively and better.The point I was making in my other comment “it may have to be cheaper”.

  16. kevrmoore

    It will be interesting to see if lower quality, freemium users are monetizable en masse. This is the problem that Live Mocha had. Online, asynchronous lessons are inferior as a methodology for achieving fluency in a foreign language. Thus, the value prop to this user base remains a question, and monetizing them could prove to be quite a hurdle. Fluency in language learning for the middle class and business professionals is most effectively taught through immersive, interactive methodologies. Can an online lesson school with a low quality, freemium user base really create a certification standard in this space and migrate its user base to pay? Will be a big challenge.

    1. sigmaalgebra

      I’ve got to agree with you, strongly.I did too much in education and am awash in experience with classrooms of wide variety, self learning of content of wide variety, etc. There’s an old saying: “Learning is not a spectator sport.” Instead, for nearly anything significantly useful to know, learning is hard work, at least for the student, maybe also for some teachers.In my learning, foreign language was a special case: In high school, I tried to learn French but could not. The teacher taught grammar first, reading second, and speaking as a distant third. For the vocabulary, apparently were somehow to get that by, what, sleeping with the book under a pillow?But there was good news: In college I knew I was supposed to know about two foreign languages, had done poorly with French, so tried German. The German teacher turned out to be by far the best teacher, just for the teaching, I ever had in school, by a wide margin. The guy was unreal: He taught me two years of college German in half time in one semester. So, he was ballpark 8 times more effective than expected. Yup, he started with reading and speaking, kept any grammar simple, and did writing only third. He did a lot with rote memorization. For vocabulary, we did some reading and, especially, speaking. He also had us use a language lab that was run by a native speaker.When I came out, a German exchange student said my accent could fool a German. Of course: All I’d heard of German was just the language lab voice of the native German speaker.Later I took a graduate course in differential geometry from a Czech mathematician, right, who’d worked with Uncle Albert, to help him clean up some of his math, wrote a book, in Czech, and had it translated into German. The math was trivial. The German was grim, and a German math exchange student said that the German was in an archaic style difficult even for him. Still I could read the German.For my Ph.D. language requirement, my advisor said for me to go to the library, find a paper in math in German, and translate it. So, got a paper, with some math I mostly knew, dusted off my German text, reviewed a little, and translated the paper. Piece of cake. A fellow student had been a math, German double major in college and had spent a year studying math in Germany, and my advisor had him grade my translation! He found an error in one word; his meaning for the word differed from what was in the English/German dictionary I used!But, for good vocabulary for fluent oral communications in a foreign language, tough to believe just a book and self study would work very well. For math? Okay; my view is that mostly people who learn math, even if they are taking a class, learn it mostly just by self study anyway. But for oral fluency in a foreign language? I’d need access to native speakers and some teaching.More generally, if people want to learn stuff, for a lot of stuff, fine. Often it will be reasonable for people to learn on their own and then take a test. But there are some obstacles: The core obstacle is what I mentioned, the learning can be hard work.For more, I doubt that anyone will want to go for abdominal surgery or a root canal procedure from someone who learned just via self teaching just on-line, but for some fields, especially computer programming, on-line instructions with some testing should work well. Still, with some irony, at least for relatively low level positions, the computer industry has found ways to ‘filter’ and ‘certify’ what they really want already via, say, some work history, some code samples, and some interview questions. Academic records can help but are not necessary. Other means of certification appear not to be very important. But, of course, for higher level positions, some $20 certification test won’t be very relevant, and often some high end academic work, maybe graduate school at MIT, CMU, or Stanford, will remain crucial.Altogether I’m wondering really how much of current classroom and university based education could be ‘disrupted’ by some on-line content and a $20 certification test.

  17. Saikiran Yerram

    I love Duolingo for its simplicity and ease of use. But I am not entirely sure about competing with TOEFL.TOEFL is usually taken by foreign students when applying for US (or any english speaking countries) schools and is required by the immigration department. In my opinion it will be a long road ahead for them to replace TOEFL.

    1. Humberto

      I don’t about how hard it’ll be for it to replace TOEFL in terms of actually getting the US government to accept it.. I’m betting on super hard.Most US universities don’t even accept The British Council Proficiency in English – and that is a much better, renowned and harder test which is approved by the British government.In terms of quality, that’ll be super easy. TOEFL is a very basic test.

      1. Drew Meyers

        hard. but not impossible.

        1. Saikiran Yerram

          Nothing is, but time matters. The question is whether the investors can wait out that long. To rephrase, Its implausible but not impossible.Nevertheless, it is a moot point if they have some serious pull across government and universities in making Duolingo a standard test.

  18. LE

    Provide the education for free but charge for the certification (testing). This is a very elegant implementation of freemium as its an easy on ramp and the customers who get the most value are the ones who pay.I’m guessing that job assistance and possibly green card help could be down the road.

  19. LE

    Couldn’t resist (after a chat window pop’d up) asking Rosetta Stone why I should pay $274 vs. Free for duolingo.Short answer to why I should pay $274 was “Duolingo is free because the content of the application is limited”.[Screen grab attached.]

  20. george

    Super impressed with this approach. Online education is truly the most effective way to monetize learning, mobile apps is taking it to a new level.From my view, this eye to eye strategy will win if the visionaries don’t give up or sell out.

  21. Steven Kuyan

    Great post, education is in need of a significant reform, including the tenure process.I think it would be interesting to explore a freemium process for college education, get the core classes — condensed into the first year of school — for free and pay for the specialty that you want to pursue, ie. CS, Physics, etc. This way if you can learn it on your own, its entirely for free but you still get some sort of degree. If not, you can pay to learn as much or as little as you want.If school’s can figure out how to give out a degree under some sort of freemium model, I just don’t know how much longer the “value” of the degree will hold.

  22. Humberto

    There are plenty of models out there like that, where money is not the currency charged. For example a lot of article writers online do it because not only do they like it, but it also creates prestige and influence among their target audience.. and they can also charge in the end.For example, AVC is an extreme case of freemium model. There’s a a million readers who learn from FredW’s posts for free, and he gets to learn from them and create a community, but if you really want to learn hands on and get the certification for entrepreneurship, you most likely will want to hire his professional investment services (aka funding, counselling…) for which the fair share is charged.

    1. sigmaalgebra

      So, now we have that the main source of freemium is itself a case of freemium!

  23. MFishbein

    The challenge with “certification (testing)” is getting employers to see it as viable. Most people go to school so they can get a job. Right now employers see a degree from a traditional university as a viable credential. Once employers start seeing other credentials as viable, education will be changed for the better.

  24. Dave Pinsen

    App Academy (… ) has an interesting model for paying for education:App Academy does not charge any tuition. Instead, you pay us a placement fee only if you find a job as a developer after the program. In that case, the fee is 18% of your first year salary, payable over the first 6 months after you start working.At the end of the program we run a hiring day. If you choose to work at any of our partner companies, we will deduct $5,000 from your placement fee. Of course, you are free to accept a position with any employer, whether we introduce you to them or not.According to the company, 98% of their alumni get jobs in the field. Only 5% of applicants to App Academy get accepted though.This is essentially the inverse of the model of most for-profit and lower tier not-for-profit colleges, which let nearly anyone enroll and rack up debt before most end up in low-wage jobs that don’t require college degrees.

  25. David Semeria

    Reputation on the web can frequently be an empty concept. Who cares how many badges you have on Foursquare? Or how many followers you have on Twitter? (What counts on Twitter, by the way, is the followers/following ratio, not the absolute number of followers)Then again, not all reputations are created equal. A good Stackoverflow score is something to be proud of, to be sure.Thiis is where I think Duolingo has a huge opportunity. Unlike app localization, for which there are many companies who will translate your UI into Cantonese, there exists — as far as I’m aware — no equivalent service for UGC.The web is becoming ever more global, but language translation (especially for UGC) remains an unsolved problem. Duolingo has a clear opportunity to do for translations what Stackoverflow did for for technical support.And it all hinges on a respected measure of reputation.

  26. Alex Wolf

    @fredwilson:disqus While I agree with free content/learning as the wave of the future, I am troubled by our obsessive need for testing. Yes, it’s very important to see what people know, and yes, quantifying that knowledge/competency is needed in some form. But I am weary of testing as this drive we have for some kind of end result to learning. And I make ed products which could fit this model.Teaching to the test has failed so many scenarios. Does the testing aspect bother you in any way?

  27. ChuckEats

    really surprised 2/3 more important languages aren’t covered – english & mandarin

  28. Kumara S Raghavendra

    Absolutely! Coursera has the same model as well where the courses themselves are free but a certificate from the Institute offering the course is charged.

  29. Josh Bailey

    This is an elegant solution. I’m currently thinking through a freemium model for a memorization app and it sometimes feels like I’m grasping for straws in terms of finding a way to give a strong value-add (the paid part) to a customer beyond the core set of features that the app itself provides (free part). Thanks for sharing this Fred. It’s helpful. (New to the blog by the way.)

  30. Bobur

    A nice move and the one that makes total sense. I guess the same can apply for Codecademy. I am glad that online education is on the road to be taken more seriously.Just two setbacks for me. First, this looks like a costly way of checking for DuoLingo. This could change with better algorithms in the future. Second, and more importantly, no one wants to be recorded! No matter how much you trust DuoLingo (which I do.)Look forward how this will work out. Great job!

  31. TaylorMiles

    Education should be Free.

    1. Lilia Tovbin

      Perhaps self-education, but guided instruction can’t be both free and high-quality, at least not on a large scale. Somebody has to pick up the cost, even if not a direct consumer.

  32. Simone Brunozzi

    Check out – freemium, a-la duolingo, and specific to cloud computing education.

  33. Chris Peterson

    @fredwilson -I’m wondering if you see any tension (as a matter of business strategy or personal philosophy) between entrepreneurial ideal of meritocratic can-do capacity and the business model of credentialed education.The TOEFL idea sounds great (go ahead and kill it), but if we switched the language example from English to programming, it seems that most entrepreneurs (and people in the entrepreneurial world) would prefer someone who ‘can code’ over someone who, say, as a Microsoft certification.Now maybe that means the existing credentials in programming are actually poor signals, poorer than no signal at all in fact. But it does seem to me like there is some fascinating friction between those two things, and I’m wondering whether a) you agree and b) if it has any implications for your investment(s).

  34. 21edu

    @fredwilson:disqus The idea behind offering assessments to monetize is a good one for a freemium model. However, taking on the TOFEL test is a hard and expensive sale. TOFEL is the standard for all foreign students planning to study in the US, it is part of the admissions process. In order for this montization strategy to be successful, all US institutes of Higher Education would need to accept this as a standard and create a systemic change to their admissions process. Systemic change is very expensive to create and market. Further, the pain point that you have identified for the student (cost) is not a Higher Education pain point. Part of the admission requirements for foreign students is to submit a bank reserve letter stating that you have enough money to cover the cost of to live and study in the US while attending a practical school. In Higher Education, foreign students are looked as immediate cashflow, the cost for them to attend is not a concern. DueLingo needs to focus on businesses, where I see a definite pain point and are cost sensitive.

  35. Abhishek Rathi

    Is the freemium model only applicable to Online Education/Learning? or Does it also includes EdTech products & platforms that creates a seamless experience around Education (such as P2P College review, P2P textbook rental, Classroom management, Alumni Networking etc.)?

  36. Tian Yang

    I can’t agree more as I’ve long held the view money in MOOCs is not in learning but the verification of learning. After all every MOOC is competing for people’s limited time/attention so offering learning for free makes sense. Perhaps under certain circumstances employers should be the ones paying for the tests because they are the ones with the most value at stake to verify someone’s knowledge level.

  37. Lilia Tovbin

    Lets try to not generalize and be a bit more specific when we refer to education or online education – K-12, college/professional, and continuing self-education are dramatically different spaces. On K-5 levels guided instruction is most effective and studies show that kids learn better through interaction with people than by using devices+software – we (humans) are wired to learn from people. Technology is great for reducing costs in all industries, education including, but it doesn’t eliminate the need for 1) high quality content 2) guided instruction methodologies which can differ from subject to subject. Neither content development nor guided instruction come cheap.Freemium model is and has been quite common in education space for both self-funded and VC-funded companies, so there is plenty of interesting cases to observe. On one side you see sites that grow very fast by giving away their product and can’t expand product development and on the other extreme you have hot ed tech startups raising funds in the millions on a promise to develop something.As many already commented earlier – developing a great product and giving it away relies on funding, so somebody (VCs) picks up the cost of the product being given away.- Founder of a self-funded Ed Tech company

  38. rmukeshgupta

    I have written about how to hack higher ed in a way that not only benefits the students, but also the teaching institutions and the corporates who hire folks from these institutions. The jist of the post is that teaching institutions should adopt a “Membership model” where they sign-up students for life membership, with the membership fee being at the lowest at the start and increasing annually (by a fixed %), which can then be sponsored by their employers if needed. In return the students get access to do a course from the university every year going forward. The more details in the post @

  39. LydiaSellerofPurple

    My husband has been student teaching physics at an underserved (i.e., poor) school in Central Calif. coast. Students need devices, Internet connectivity at home. Some work 3 jobs. He saw the need for a teachers’ physics app that is working on with a team (unfunded as it will be free for schools, so no $$ in it for an investor). He finished his MA in Science Ed. Teacher salaries and job security are appalling. Free sounds nice until you add humans and the infrastructure needed to develop, market, sell, download & support free apps. 2nd career for him (retired tech VP). I worry about the new generation who are leaving teaching in droves – and, even if online, someone needs to teach, etc. And don’t look to discovery teaching! Studies and real-life have shown it’s a disaster.

  40. LydiaSellerofPurple

    VCs and wealthy entrepreneurs need to fund K-12 education, including sciences, especially in California. And, let’s fund the arts, including music, painting, sculpture & writing, please.

  41. swagshirts99

    wish you luck

  42. Uday enterprises

    what will exactly pay by students?

  43. Guest

    small help required..i downloaded the duolingo app in india and the default language is hindi..whereas comfortable language for me will be english…default language being hindi is a little put off…and i am not finding a way to change it too…written to duolingo support too …felt should write it out here also…thanks a ton..

  44. diymanik

    Most ed tech startups have one fatal flaw, they keep trying to make online learning more like schools. The real opportunity is for startups to take the way online learning is done online and make it more accessible to the average person. The startup that finally cracks the code on internet learning will do it not by trying to run through the gatekeepers we call the education system, but by going over their heads. I believe that startup will do it by focusing on letting people learn what interests them the most. Giving individuals unparalleled choice in the 5 W’s will give that startup the advantage it needs to become the de facto for online learning. Watch out, at first that startup will look like a tool for hobbyists only but we know that’s where the big things start out.