Lessons from 'Duolingo': How to Make Learning Hard Things Easy
ฝัง
- เผยแพร่เมื่อ 24 ก.ค. 2024
- In this 2019 GDC session, senior director of engineering Karin Tsai explains how Duolingo created an experience that balances teaching difficult subjects and keeping users engaged.
Register for GDC 2022, in-person in San Francisco, March 21-25, 2022: gdconf.com/passes-prices/?_mc...
Join the GDC mailing list: www.gdconf.com/subscribe
Follow GDC on Twitter: / official_gdc
GDC talks cover a range of developmental topics including game design, programming, audio, visual arts, business management, production, online games, and much more. We post a fresh GDC video every day. Subscribe to the channel to stay on top of regular updates, and check out GDC Vault for thousands of more in-depth talks from our archives.
Duolingo is a prime example of "what you measure is what you'll get". They're not measuring "learning" and "language skills", but rather optimise for "engagement". Speaking from my point of view, the end result is that me and everyone I know who's used it at best only learned some vocabulary and a few phrases, after using it daily for months. We stopped learning the content, and instead only memorized whatever was necessary to finish a level. Of course as a business, engagement trumps almost anything else...
I realised that my comment sounded very negative, but I do think this talk was really great. There are many learnings for how and what to measure, and pitfalls and unexpected consequences, both good and bad. I just think the fundamental concept of Duolingo is probably flawed. It might be something that can be pivoted, but only if they really re-examine their core metrics.
It's a bit like Google trying to optimise metrics but ended up ruining both the web and TH-cam.
@@Ermude10 Your comment was critical but constructive. What you said is true, though. I also think the main issue is the core design of the application does not facilitate language learning very well because the scope/scale that they are aiming for is just too large.
The "streak" feature relies on players "addictive" nature and a fear of missing out (even if its meaningless).
Its a social pressure or a self-stressing addition thats ultimately just added stress and players are willing to pay to not feel that stress.
The more games and systems do that, the worse it gets, as a player cannot do everything every time and its a problematic feature to have in the global sphere of programs.
I'm okay to feel pressure if it motivates me to study
The streak is actually for them, not for you. Kinda how Facebook or Instagram keeps feeding you with notifications to keep you on the platform for as long as possible. Making it easy to connect to other people is not the primary goal for Facebook. Their main goal is to maximise revenue, and they try to keep you on the platform for as long as possible in order to do that.
You seem to have a good grasp of it, now please use it for your advantage :3
They will not give up their atomic weapons
So go ahead,
m a k e a b o m b y o u r s e l f
_b l o w t h e w o r l d u p_
Duolingo detaches language from culture and social interaction, making its learning a chore (and not "fun"). I had never seen it working well and I doubt I'll ever see.
I worked in a school that gamified book reading. In didn't work also.
I personally ended up realizing that it doesn't work. I tried to learn german and did it religiously on a daily basis for a year. I didn't expect to become a master or anything, just being able to have a simple conversation. You end up playing a game. The sentences are often the same, the patterns are often the same. You might get 100% of things correct, but that doesn't mean you're learning the language. You're learning how to play duolingo in that language. That's pretty much it. Learning a language requires a level of practice and immersion that a few games simply will not suffice. At least for me, and I believe, for most people.
you started from the wrong assumption that duolingo alone would teach you everything and you wouldnt need to practice in other contexts ever. I doubt they state that to be true.
Duolingo is best as an introduction to learning a language. You should eventually move on to better language learning services, or pick up books or go to classes and learn through there.
Duolingo mimics the way children learn-repetition, repetition, repetition with minor variation. It makes it easy to incorporate new vocabulary in the language you are learning. Think of how long it takes a child to master a language. I have done Duolingo for a year and now intuitively know the correct answer-mostly with a 90% average. I do it because I am 77 yr. old and want to keep my brain sharp and find this takes only a small amount of time while being fun, too.
I think everyone is right, but for me, I’ve learned more words the past six months than I have the last 20 years. It gives me examples of sentences, verb tenses, etc. We can probably argue about the memorization, but as far as I’m concerned, I memorized English. I’ll say the phrase.” the bus will pick us up at the corner.”, which is incorrect it won’t pick us up. The better phrase would be, “the bus will collect us at the corner”. So did I learn the incorrect phrase or did I memorize it? I’m gonna give Duolingo a plus at this time, it is a very good start for me.
Duolingo was personally not for me, because I prefer a more structured type of learning. Despite that, the talk did make some good points about how to increase player retention, and how to present large amounts of information in a more easily digestible form. I could see how this could be useful when designing, e.g. survival-craft genre games, in which the initial learning curve is often very steep indeed.
So the lesson here is that is you actually make a useful product it will drive away the causal learners and learners who might not have used the product if it wasn't for all the dumbing down. Duolingo markets themselves as the best tool to learn a language, but it's really just a tool to learn to play Duolingo.
Fantastic talk! I also signed up for a year as a result, you got me! :D
TIL: Streaks are stonks
Considering that Duolingo is hilariously awful at teaching you anything we can learn a lot from it, since one can always learn more from mistakes than successes :v
The rules are simple: Spanish or vanish.
It's an interesting talk, with some good ideas.
Having said that, the banging on about "fundamental core values" really doesn't mesh well with the hyper-detailed, pervasive use of metrics and design-by-KPIs.
It feels like marketing speak; window-dressing for a business that wants to be seen as an "education for free" organisation. It sounds like you want to be seen as a charity or noble endeavour.
But the truth is you're running a big business, with over 400 employees. You have a very large wages bill, so you're not doing it all for free. There's no shame or harm in that; more power to you. You help employ people and you help people learn languages. But you're not a charity.
Just tell the truth, and speak plainly. That way, you won't have to use marketing speak to cover the real truth of commercial decision-making.
all companies these days and maybe for a long time make an effort to drill into their employees that there is some higher cause they're serving. So you can ask them to work unpaid overtime with less complaints, they're less likely to band together over unfair practices and oncall. .etc.
"just be honest" goes away in all but the tiniest companies.
"...Because Users Don''t Read" ......so reading isn't a thing anymore?
"What the community asked for and what their behavior showed wasn't aligned" *Insert any long standing series that makes change the fans have been asking for years* Fans: Why would they do that? It doesn't even feel like (Title) anymore.
Something that is unclear is what's their main goal, it seems like their main goal is to get people to use the app instead of having them actually learn, or enjoy the experience.
Plus the fact that they explicitly talk about treating users like children is pretty shitty
Yea, dumbing something down is aiming at the lowest denominator.
High archives are basically always distracted or get outright offended for being treated like a child and stop using a program.
If your customer has a broader understanding already, its a completely different customer base than one that wants as many people as possible (and the most people are relatively stupid and have weaknesses to abuse, like addictive features and anything with a fear of missing out, or social pressure features to force them into a behavior that is beneficial for the company to make money of).
These dumbing down features and abusive mechanics often creep into a software with innocent or clueless understanding, while they produce more customers, the effect on society as a whole can be disastrous.
They are a company, company want money. It's a bit cynical, but this talk exemplifies it.
Main takeaway: don’t do data driven design if you don’t even remotely understand why the data happens. If you base your decisions on -0,5% metric which you can’t even explain, that’s just absurd. 0.5% is nothing, and also you’re looking at a local minimum, and taking the opposite decision might increase by loads after a second change.
This is just micro optimizations gone mad, something programmers are taught to avoid like a plague. I have no idea why people are still going this obviously flawed way, rather than actually having a concept and goal and improving that.
"People are not, usually, super motivated to work hard at something, voluntarily, especially when they aren't forced to."?
OK, sure, but let's see you do a 1 hour talk on education without stumbling slightly or repeating yourself.
@@w0mblemania think the person is highlighting the content of the sentence instead of how she's speaking.
Pretty interesting, but the corporate speak used all throughout the talk is very unsettling
That’s how the company wants us to speak and that’s how they taught us in college. My college English teacher hated my essays they were to clear into the point. She said that I need to add a lot more words.
I found this talk exceptionally dry (the intro was written like an academic paper rather than a talk seeking to engage an audience), and inapplicable to the average indie. If you consider a change of 0.5% to be massive, then your project is big enough to have people dedicated to stuff like this.
Huge miss.