| > The learning flywheel: The greater the scale of our learner base, the more we can use insights from data analytics to improve both engagement and efficacy. Iterating on education efficacy is valuable but it doesn't really very much matter if the material you're trying to teach is wrong (2+2=5) or if it doesn't translate (heh) to real life skills outside of the application. Particularly for language learning, where the goal is generally to find ways to consume or utilize the language outside of Duolingo, you are judging it based on how effectively Duolingo prepares you for that task. Tailoring lessons to specific learners and recognizing and understanding their learning patterns is a great way to enhance learning for just about anything. In that regard I think Duolingo is doing something interesting, if not incredibly obvious. I also like that the lesson paths give some flexibility on what sort of stuff you want to learn about while maintaining your 'streak.' Gamification, while not for everyone, clearly is a model that works to keep people engaged. The issue is that we're not just judging it based on engagement, at least if you buy the premise that they're trying to provide a high-quality education for free to anyone that wants it. The quality is sometimes questionable depending on the course you're doing and the pacing, even if you consider that they design it to adapt to different learners, can be remarkably slow. It also doesn't seem to really compare favorably against dedicated self learning or classroom instruction in terms of time spent versus how much you learn. How much does their engagement rate suffer if they tailor the course to increase fluency in a similar time frame alternative methods can offer? I suspect it'd reduce quite a bit, so it doesn't change too much. This is why if someone asks me about learning a language I would suggest perhaps to look at an app like Duolingo or Lingodeer or any other app that aims to compete with Duolingo as only a starting point to assess how committed you are to learning the language. If they can maintain a streak on Duolingo it at least shows that they may be committed enough to tackle a more challenging but more rewarding learning regimen that will increase their mastery of their target language in a much shorter time frame. I understand that people's commitment to language learning varies a lot. In fact I think Duolingo really understands this and depends on it for their monetization strategies. I think it's perfectly fine to use something like Duolingo if it fits your needs and I think that for the vast majority of people they'll find it easier to use and more engaging than alternative methods. I just want to keep people's expectations in check (you probably won't achieve fluency through Duolingo alone) and that alternatives will be a better use of your time if you are more committed to it. Language learning can be incredibly dull and frustrating at times which are really bad for engagement and as such Duolingo is going to try and design around that, but it's the reality of any foreign language you learn. There's some really basic stuff that you really don't care about and your brain simply doesn't want to remember but you have to force it into your head over and over until it sticks. Doing this gets you to your goal of enjoying consumption of the language through movies, tv shows, music, books, comics, and people much faster. > At the time of filing the S-1, Duolingo had a team of ~280 software engineers, product designers and product managers. Approximately 20% of them were on functional engineering teams, building and maintaining foundational infrastructure, while the remaining 80% were on product teams, iterating on features as well as researching and prototyping new product extensions. I'd be more interested in learning how many educators, linguists, and language teachers they have on staff or if they're still relying on purely crowd sourced language learning resources that have been pretty wrong in the past. Iterating new features to increase engagement or "learning efficacy" is great, but you need to have good curriculum first. If they're committed to providing "high-quality education" for anyone that wants it, it starts from that. |