ReadLang bought by Duolingo

All about language programs, courses, websites and other learning resources
User avatar
Tofu
White Belt
Posts: 15
Joined: Wed Nov 08, 2017 7:21 pm
Location: United States
Languages: English (N), Spanish (Intermediate High), French (Intermediate Low), Japanese (Novice Mid)
Language Log: https://forum.language-learners.org/vie ... =15&t=7220
x 33

Re: ReadLang bought by Duolingo

Postby Tofu » Tue Nov 21, 2017 1:48 pm

Well I started familiarizing myself with anki yesterday, just in case.
3 x

User avatar
lavengro
Blue Belt
Posts: 728
Joined: Wed May 24, 2017 1:39 am
Location: Hiding in Vancouver. Tell no one.
Languages: English - finally getting a handle on this beast of a language. Also tinkering with a few other languages intermittently.
x 1994

Re: ReadLang bought by Duolingo

Postby lavengro » Tue Nov 21, 2017 7:00 pm

Edited for brevity.
Last edited by lavengro on Tue Nov 21, 2017 8:38 pm, edited 2 times in total.
0 x
This signature space now on loan to the mysterious and enigmatic Breakmaster Cylinder:
https://www.youtube.com/watch?v=hUGKldkiex4

User avatar
emk
Black Belt - 1st Dan
Posts: 1620
Joined: Sat Jul 18, 2015 12:07 pm
Location: Vermont, USA
Languages: English (N), French (B2+)
Badly neglected "just for fun" languages: Middle Egyptian, Spanish.
Language Log: viewtopic.php?f=15&t=723
x 6330
Contact:

Re: ReadLang bought by Duolingo

Postby emk » Tue Nov 21, 2017 8:22 pm

Cavesa wrote:I don't mind the market being split in two parts, a large toy part, and a smaller serious part. It has always been this way (all those "learn Chinese in a month and 100 pages" courses). What I do mind is the serious part being pushed away from its already gained position and damaged by the toys.

If "serious" language learners want a good selection of effective tools, then two things need to happen:

  1. There need to be enough serious language learners out there.
  2. Those language learners need to be willing to spend money on tools (or write those tools themselves).
Sadly, the vast majority of people buying language courses (at least in the US!) aren't very serious. If you asked them to put in 20 to 40 minutes a day for 6 months, they wouldn't be interested. To use the usual sports metaphors, it's like training to run a marathon or getting in really good shape—most people can do it, but the price is higher than they want to pay. And that's OK.

But (2) is an important point, too. The iOS version of Anki costs $25, and the money the author earns from that also pays for desktop Anki. (The marvelous AnkiDroid version was written by volunteers.) And yet, if you ask around on language-learning forums, you'd find plenty of people who say things like, "Anki is way too expensive at $25!" Many customers are offended by software that costs more than $3, no matter how niche the market.

OK, sure, maybe Anki isn't worth $25 to you. But it's a great product aimed a niche market of "serious" learners. And it's not a huge market, which means it's going to be more expensive per person.

Let's see if we can figure out how big the market for the iOS version of Anki might be. I can't get download numbers for iOS apps, but the most popular paid flashcard app on Android seems to be Flashcards Deluxe at $3.99 with 50,000 to 100,000 downloads. This gives us a maximum all-time revenue of $400,000. And it's been around since 2009, which would suggest a maximum of $50,000/year before advertising, servers and the costs of running a business. Not bad, but a competent programmer in the US could certainly do better at a day job. Now, Anki costs $25, it runs on iOS, and it has a different audience, so we don't know how much it makes—it might be peanuts, or it might be a surprisingly nice living. But Anki has never turned into a giant company with investors and a marketing budget, so it's probably not making giant sums of money.

If "serious" language learners want good tools, the two main options are (1) pay enough for them in great enough numbers for the author to make a nice living, or (2) write the tools as open source and share them. Any other tools are living on borrowed time.
12 x

User avatar
Adrianslont
Blue Belt
Posts: 827
Joined: Sun Aug 16, 2015 10:39 am
Location: Australia
Languages: English (N), Learning Indonesian and French
x 1936

Re: ReadLang bought by Duolingo

Postby Adrianslont » Tue Nov 21, 2017 10:31 pm

Emk, I (respectfully) think one of your two premises is wrong here ie no.1. And I think no.2 is not completely true - though it might seem to be.

Regarding no.1 - I think there are PLENTY of serious language learners in the world. Sure, there are many more learners who are less than serious hobbyists or high school students who are being forced to learn a language they have no interest in BUT there are a lot of people in the world learning languages and enough of them are serious about it to support a Readlang or anki style product - or other innovative product. Apart from language nerds such as found on this forum (a very small number) there are people who move to another country to live and/or work and there are university students. There are hundreds of millions of university students in the world and I think a large number of them have to be reasonably serious students. That’s a huge market. WE know how useful Readlang and similar tools are for reading and THEY have a lot of reading to do, which I am sure they would like to be easier! The thing is, they don’t even know Readlang exists - if they could be shown the benefits of Readlang - “Hey, if you use this you don’t have to thumb through a paper dictionary or juggle your attention between your book and your smartphone dictionary app.” I think enough would be convinced.

I think there is another part of the puzzle that is missing though - apart from the lack of awareness of such tools - the tools have to be RIDICULOUSLY easy to use and although Readlang almost fits that bill it is not completely there. It is ridiculously easy to use once you have your text uploaded but access to digital versions of texts is of course not ubiquitous. For users who don’t have a digital DRM-free copy of their texts this is a problem. Us more nerdy types will hunt for digital copies and use tools such as Calibre to get rid of DRM and convert formats but those are time consuming, too techy, steps too far for most people.

Anki is even less friendly for users. I have used anki for three years and think it’s pretty great. I happily paid for it and have encouraged others to do so over at Reddit. But it’s a bit scary. Thanks largely to you, emk, I have learned to make and use subs2srs cards so I’m not totally useless with tech but I still find anki intimidating. And I know many more people are more scared of it or distanced by the interface and card making process than me - I couldn’t sell it to my daughter who spent three years studying a language at university. She’s a serious student who got excellent grades but would have really benefited from Anki.

Related to this issue is what Gabe Wyner and Mike Campbell seem to be doing with their products - if I understand correctly. Apart from moving to subscription models and making their products unable to be pirated (two huge changes there), they are making their products EASIER to use. You will no longer have to own anki and download decks or make your own decks to use Fluent Forever. You will no longer have to store MP3 files on your perhaps overcrowded phone to use Glossika. These products will be that much easier to use. Hopefully they won’t be dumbed down in terms of functionality and there are of course concerns about pricing - there’s a lot of worrying about that here. It’s wait and see on those issues but the whole idea of tap on the app and start serious learning immediately is pretty attractive to me.

Anyway, that’s my thoughts.

TLDR: I think there are enough serious learners in the world to support serious tools and I think enough of those serious learners will pay for them BUT they have to know they exist and they have to be incredibly easy to use.

PS I’m not “blaming” Readlang for “failure” to market well - I think it’s a huge job and I think the whole digital texts/DRM thing is a huge hurdle.
4 x

Cavesa
Black Belt - 4th Dan
Posts: 4960
Joined: Mon Jul 20, 2015 9:46 am
Languages: Czech (N), French (C2) English (C1), Italian (C1), Spanish, German (C1)
x 17566

Re: ReadLang bought by Duolingo

Postby Cavesa » Wed Nov 22, 2017 1:18 am

Emk, I highly value your experience, but I think there is a huge flaw in your thought here. And it is one you probably share with many of the creators and marketers of language learning tools. You are looking at the issue only through the eyes of an English native. I don't mean to offend you. But in your priviledged part of the world, extremely few people NEED to learn a foreign language. They need to pass an obligatory class, or they are curious, or they just vaguely thing "it could bring me a better salary, if I learnt Spanish." Extremely few people need it, because the whole world is learning English and because those positions for English+Spanish (or something) speakers have lots of applicants native of the other language and with excellent command of English as their second language.

No wonder that there are very few serious learners in such a population. But it is different for the rest of the world. ESL is a huge monster with huge PR, it is extremely profitable business. And French or German or Spanish, while definitely being far behind, are very profitable too. Half the planet is learning foreign languages, an a big part of them takes it seriously. They are sometimes lazy, sometimes just confused about how to learn, but they are definitely willing to try and to spend money on something that would finally give them that important line in their CV.

Do you have any idea how much is a native of an internationally worthless language willing to pay to learn a language? There are many countries, which will be somewhat similar (but the numbers will vary of course), but I'll take the one I know the best for illustration.

The Czech Republic. The average salary is something like 25000 crowns (1 euro being between 26 and 27 crowns these days), and most people don't have the average (if I remember correctly, the median was around 22000). A small appartment (1+1) in Prague (definitely not the city center) is unlikely to come at a lower rent than 12000 (10000 already means there is a catch :-D ). Plus the bills like 1500 crowns, the public transport 300 (if you choose the cheapest variant and pay for the whole year at once) and many people need to feed a car instead (nope, the public transport is not as perfect as it looks to the tourists), and so on. Food and vast majority of goods people normally need comes at the same prices than in France, Spain, or Germany. Individual items may vary, but the same shopping cart will cost more or less the same money. Then there are other things than food that need to be paid etc. So you can imagine that an average czech doesn't have that much money to spare, and they need to put it in other things too, such as creating a financial reserve, yet they choose to pay a lot for language learning.

Language schools are thriving. Language classes in average language schools cost between 4500 and 8000 per semester, sometimes only trimester. Plus the textbook, which means additional 500-900 crowns. Language schools are thriving. Even lower middle class people are spending so much money for the classes. Families, who definitely need to think about their budget, are paying their teenage (or younger) kids language classes outside school, sometimes even for two languages. The individual teachers are becoming more and more common too, despite asking between 400-700 crowns per hour. People are paying more for natives, even the crappy teachers who just got a CELTA to travel the world and be treated like kings just because they are English natives. And don't forget the crowds taking the FCE/CAE exams, the fee being around 6000 crowns (perhaps more now, I took CAE in 2010), FCE being cheaper, preparatory books for those and preparatory classes for those (which tend to be more expensive than the usual ones, fast googling gave me prices like 8000 or 10000)

So, if we are optimistic and expect the learner to not get stuck at the False Beginner level (which is a very profitable invention of the language schools), if we are optimistic enough to believe the learner will truly get to the advertised level, even though slower than a serious independent learner: a stereotypical czech learner of ESL pays 12 semesters of classes to get to B2, and wants to pass FCE.
12 semesters per 6000 crowns
4 levels of coursebooks per 500 crowns (cheap coursebooks)
2 levels of Murphy grammars or similar, 600 crowns each
FCE preparatory coursebook 800 crowns
FCE preparatory classes 8000 crowns
FCE sign up fee 4500
(and they will keep complaining how hard it is to learn vocabulary the whole time, not knowing about Anki)

That is 88500 crowns for a normal serious conservative learner. For a teenager with responsible parents. For a secretary improving her qualification. For a student of economy. That's like 3300 euros. I don't think such a person would find 25 euros so expensive, if they were presented the product well.

And I definitely didn't count in the very popular classes abroad. Those are less widely spread, but still a lot and those are yet another level (cheaper two weeks are between 15000-20000, as far as I know)

And those were the middle class people (well, the czech society is much less stratified, so it is quite a wide term). Actually, as far as definitions of middle class go, majority of our middle class would be considered lower middle class elsewhere. When it comes to the richer people, they spend much more! The luxury classes for companies or for richer individual students, those are several times more expensive! And you would be surprised how many schools can sell such expensive services, people are willing to pay.

People feel the NEED to learn English (which is being promoted the most, even though I find this controversial), they NEED to learn German, and more and more are finding French, Spanish, and Russian to be reasonable and valuable choices. Vast majority of czechs (or poles, or japanese, or actually vast majority of the world) doesn't learn a language for fun, they want to seriously learn it. If an English natives wants to seriously learn a language, not much happens, if they fail. If a nonanglophone fails, they are bound to not reach their full potential in their career.

They pay such money despite the disadvantages of classes, which they are often even aware of. They pay, because they are convinced they are paying for efficient and great service. Or rather, they don't know any better alternative, and convincing them even of supplemental options is not that easy, as they trust the language school advertisements.

Compared to these sums, 25 euros per Anki are nothing. I gladly paid those 25 euros, even though I am a student. I am sure many people would, but AdrianSlont is right about the reasons why it doesn't happen, I can only add my view on them:
1.People don't know about Anki or Readlang. They see all the advertisements for language schools in real life and online. The internet has the potential to give space to great stuff like Anki, but crappy products are getting more light and not just in the paid sections. I am usually the first person who has ever mentioned Anki to them. So, I am the only user they know, and many simply don't take me seriously (I've encounterd all the "but you are really gifted" and "no, it must have been the Erasmus" or "you are surely lying about your skills" variations.) No product can get to such a market only through a few excited learners like us.
2.People distrust apps. And now it is not just because teachers tell them to, or because of being conservative, those days are gone. The avalanche of bad apps is destroying the reputation of all of them. I used to recommend people Memrise. Do you think I can recommend it now? With the new childish stupid kitsch design and the dubious official courses making the first impression? They won't give it a fair chance, despite the good functions and content still being there, and they will doubt my taste (and sanity). When we get so far that app=toy, in the eyes of the potential customer, we cannot expect people to pay more than a few dollars. They simply won't see it as a good investment, because the last four apps they tried were useless, how could this one be that different.
3.Anki or Readlang does not look easy to use. as AdrianSlont said well. I really dislike design being polished at the expense of the form, but there must be a central bliss point between nice+easy and efficient. Actually, both would profit tremendously from copyright that would make more sense. Right now, learners not living in the target country have no legal way to get ebooks for Readlang. And they have to make their own Anki lists out of their coursebooks, as sharing "Assimil wordlist" decks seems to be problematic as well, from what I've heard. (I'd be interested to know how has Scritter got the coursebook wordlists for their service. If they found a business model working for the publisher too, it would be awesome if it would spread).

So, both conditions are actually fulfilled. 1:enough serious learners-check. 2:learners willing to pay-check. The only problem: the sellers somehow cannot see this, as they are too focused on catching the attention of a usual English native vaguely interested in learning a language like "it would be so nice, if I could order my food in Spanish, when I visit Mexico."
Last edited by Cavesa on Wed Nov 22, 2017 1:43 am, edited 2 times in total.
7 x

User avatar
emk
Black Belt - 1st Dan
Posts: 1620
Joined: Sat Jul 18, 2015 12:07 pm
Location: Vermont, USA
Languages: English (N), French (B2+)
Badly neglected "just for fun" languages: Middle Egyptian, Spanish.
Language Log: viewtopic.php?f=15&t=723
x 6330
Contact:

Re: ReadLang bought by Duolingo

Postby emk » Wed Nov 22, 2017 1:40 am

Adrianslont wrote:Regarding no.1 - I think there are PLENTY of serious language learners in the world. Sure, there are many more learners who are less than serious hobbyists or high school students who are being forced to learn a language they have no interest in BUT there are a lot of people in the world learning languages and enough of them are serious about it to support a Readlang or anki style product - or other innovative product.

OK, let me actually break out the numbers, and work through them like I would for a startup. This stuff definitely isn't obvious (certainly not to me), and there are a few surprises.

The total world-wide market for language learning resources (apps, courses, schools, tutors, etc) is estimated to be about $35 billion/year. This is what an investor would refer to as the total addressable market (TAM). For the sake of discussion, let's divide this into three categories:

  1. Formal schooling. This belongs to textbook publishers and actual schools. Not even Rosetta Stone has enough marketing muscle to capture a significant portion of this market. Also, selling products to educational institutions is a special kind of hell for startups.
  2. Tools sold to beginners who think it would be nice to learn a language. Rosetta Stone is one famous example in this category, with a revenue of $264.6 million in 2013. So the most famous traditional course publisher has captured less than 1% of the total addressable market. A more modern example would be Duolingo, which has raised $108 million dollars in venture capital, and which costs US$40,000 per day to run! I imagine that they're still burning through investor money, and only time will tell if they've got a sustainable model.
  3. Courses and tools for independent learners who genuinely expect results from their tools (and who do their research). One nice example here is Assimil, which has an annual revenue of about €5 million, and 20 to 50 employees. They're genuinely popular in France, which helps. Another good example is Pimsleur, which is very aggressive about marketing, more like a company in group (2), except their courses work. But they're privately owned and I can't find numbers I trust. I imagine they make more than Assimil.
So, yeah, there's a ton of money there. But I think Assimil is a pretty good benchmark: They make highly effective courses, and they're strong in one major European market. And of course, using an Assimil course requires zero technical skills. But after three generations of working at it, they only earn €5 million/year, which would be enough to run Duolingo for maybe 5 months. So if we do a back-of-the-envelope calculation, and assume that we can build a company with 10% of the revenue of Assimil, then we're looking at US$590,000/year. The app store will take 30%, leaving us with $393,105. If we figure in the various costs of doing business, that corresponds to maybe a $200,000 salary. And I guarantee that somebody like Steve Rideout can get hired for $200,000, given the quality of his work and the wide range of his skills. That's well within the range of what Google pays their high-end talent.

So realistically, a company that's 10% as successful as Assimil can afford to bid against Google for one first-rate programmer-entrepreneur. Well, except I left out the marketing costs to acquire a new customer, which could easily eat up your profits. You'd need to know a few of things: your customer acquisition cost(CAC) (how much money you spend on marketing and advertising to get them to sign up) and your customer lifetime value (CLTV) (how much money customers give you before they stop paying). And you also need to know what your cash flow looks like—if it costs you $50 to acquire a customer, and if they spend $5/month for a year before "churning", then you only make a $10 profit in the final two months of the year. So once we take CAC and CLTV into account, we're definitely going to want to bigger than "10% as successful as Assimil."

(CAC and CLTV are probably two numbers that Duolingo is watching like a hawk. If they like what they see, they'll be able to get a big investment to grow their business rapidly. But their investment rounds aren't getting progressively bigger, which means that their investors aren't ready to hit the gas yet.)

Adrianslont wrote:WE know how useful Readlang and similar tools are for reading and THEY have a lot of reading to do, which I am sure they would like to be easier!

…I think there is another part of the puzzle that is missing though - apart from the lack of awareness of such tools - the tools have to be RIDICULOUSLY easy to use and although Readlang almost fits that bill it is not completely there. It is ridiculously easy to use once you have your text uploaded but access to digital versions of texts is of course not ubiquitous.

I agree that ease of use is essential to reach a mass market. But there's a nasty little wrinkle here:

  • If you're teaching beginners, you can provide your own curated content. Basically, you might wind up with something like an app that combines an Assimil course with a built-in spaced repetition system. (Assimil, if you're reading this, I could build this for you for less than you might think. Combine that with a good US marketing person, and you could triple your revenue in a few years. This is a serious offer.) This could be very easy, and more effective than anything Duolingo or Memrise are promoting.
  • If you're targeting intermediate or advanced learners, then you have a problem: They don't really need courses, because they're already reading books, watching TV, writing essays, etc. But they still run into dialog that they don't understand, and words that they can't define. In other words, they want to use their own content. And this makes the software harder to use, because you need to access the content somehow.
Actually, there is at least one popular tool for intermediate and advanced students who are reading their own content: Amazon's Word Wise for the Kindle, which includes definitions for hard words, plus a flashcard system. Basically, if you're reading English, Amazon has a built-in readlang-like tool!

Adrianslont wrote:Thanks largely to you, emk, I have learned to make and use subs2srs cards so I’m not totally useless with tech but I still find anki intimidating. And I know many more people are more scared of it or distanced by the interface and card making process than me - I couldn’t sell it to my daughter who spent three years studying a language at university. She’s a serious student who got excellent grades but would have really benefited from Anki.

There are several usability issues with substudy that I'd like to fix:

  1. Directly playing your videos for you. Unfortunately, I can't really improve this very much, because many video formats are only readable by officially-approved players.
  2. Converting subtitles to text automatically. You can download subtitles from various sites on the net, including some university research projects in certain countries. But these subtitles are often poorly aligned. A better choice would be to OCR the subtitles included in the video file. Existing tools for this are pretty clunky, but I'm slowly building a state-of-the-art subtitle OCR library that should be vastly easier than anything out there (and handle 50+ languages to boot). Unfortunately, this is slow work.
  3. Deciding which cards to make. For beginners, this is easy—just convert the entire episode and cull anything you don't need. For advanced students, you might only want to make 15 cards from an entire movie, which makes the whole process way too labor-intensive for what you get. One way to fix this would be to build a "readlang for video", where you could watch a video normally, and whenever you missed a line of dialog, you'd hit a key to make a card.
  4. Getting cards into Ank & reviewing them. The easy fix for this would be to build an Anki plugin that integrated with your video player. The good fix would be to build your own integrated SRS from the ground up, like readlang did, and tune it so that first-time SRS users didn't turn it into a torture machine by refusing to delete bad cards.
Here are some prototypes I've built at various times:

SRS Collector: Integrates with Kindle highlights, online dictionaries and Anki. Allows you to make good cards. NO LONGER SUPPORTED, may be broken.

Image

substudy player UI prototype (unreleased). This allowed me to watch a movie interactively, and mark only certain subtitles for later review.

lang-algunos-problemas.png

Adrianslont wrote:Anyway, that’s my thoughts.

You're perfectly right about ease of use being utterly essential. If I had to guess, I'd say that every time you make your program 10% easier, you reach a market 10 times larger than before. This might be an exaggeration, but it's a not a bad way to think about the challenge.

But the easiest way to simplify things is basically to package something like Assimil into a high-quality app, and aim solely at beginning students. Serious intermediate students might want to choose their own books and films or something unreasonable like that. ;-)
You do not have the required permissions to view the files attached to this post.
4 x

Cavesa
Black Belt - 4th Dan
Posts: 4960
Joined: Mon Jul 20, 2015 9:46 am
Languages: Czech (N), French (C2) English (C1), Italian (C1), Spanish, German (C1)
x 17566

Re: ReadLang bought by Duolingo

Postby Cavesa » Wed Nov 22, 2017 2:25 am

emk wrote:[list=]
[*]Formal schooling. This belongs to textbook publishers and actual schools. Not even Rosetta Stone has enough marketing muscle to capture a significant portion of this market. Also, selling products to educational institutions is a special kind of hell for startups.
[*]Tools sold to beginners who think it would be nice to learn a language. Rosetta Stone is one famous example in this category, with a revenue of $264.6 million in 2013. So the most famous traditional course publisher has captured less than 1% of the total addressable market. A more modern example would be Duolingo, which has raised $108 million dollars in venture capital, and which costs US$40,000 per day to run! I imagine that they're still burning through investor money, and only time will tell if they've got a sustainable model.
[*]Courses and tools for independent learners who genuinely expect results from their tools (and who do their research). One nice example here is Assimil, which has an annual revenue of about €5 million, and 20 to 50 employees. They're genuinely popular in France, which helps. Another good example is Pimsleur, which is very aggressive about marketing, more like a company in group (2), except their courses work. But they're privately owned and I can't find numbers I trust. I imagine they make more than Assimil.[/list]


1.Formal Schooling: Yes, you cannot replace it, people brave enough to go against the "everyone knows you need a teacher" idea are a minority.. But it leaves tons and tons of space for complements. Anki would be an ideal thing to use with any kind of coursebook. Yet, people don't know about it. Actually, some language schools are now offering a huge bonus: a vocab training program, that just looks like a much worse anki with fewer functions and little content. So, there would be demand, if only people knew they could want this. We didn't know we wanted smartphones, until we saw them. This is similar. Readlang would be ideal too, you would just need to make the content more accessible (the copyrights are the main problem here, and they are moving slowly). But there would be a lot of potential for Anki cooperating with the publishers of classroom aimed courses, that is a good example.

2.Tools for the players:yes, this is the market for dumbed down tools, for kitsch version of the future Memrise (I am not even curious about their next changes anymore), Duolingo with the false promises, for shiny stuff, for apps for 3 dollars. You are right

3.Independent serious learners: You are absolutely right, but I think you underestimate the opportunities created by the internet. Hey, our Pirate party has finally gotten into our parliament recently, with hypercheap campaign, as a result of 8 years of hard work in local politics and activism, and mostly internet based PR. They got into the parliament as the the 3rd strongest party. Other parties paid horrible amounts of money for PR but the Pirates proved that marketing is not the only deciding factor. Yes, they had a PR expert, but not a huuuuge marketing team. They had some posters in the city, but rather few and no billboards. They down own any media, unlike the head of the winning party. This is just a very recent example, but why couldn't it work for business and language learning resources too?

Just a high quality product, with good design that is not too hard to understand (but not hyperexpensive kitsch. I wonder how much did Memrise pay to their graphic for the recent creation), and clever use of the internet. Not without paying anything at all (I think that is one of the limits of the Language Transfer project), but with surprisingly modest investments.

In the products we are discussing, like Readlang and Anki, we usually see a great product in terms of functionality. Less good in term of looking user friendly. And definitely not that good at the marketing. I know it must be hard to manage the marketing well. But if you invest hundreds or thousands of hours into creating such a great thing, I can't see why you can't invest some more time and efforts (or get a colleague better at this). Is it so naive of me to see it as the the old joke about a prisoner climbing over 99 walls and giving up before the 100th?

So, yeah, there's a ton of money there. But I think Assimil is a pretty good benchmark: They make highly effective courses, and they're strong in one major European market. And of course, using an Assimil course requires zero technical skills. But after three generations of working at it, they only earn €5 million/year, which would be enough to run Duolingo for maybe 5 months. So if we do a back-of-the-envelope calculation, and assume that we can build a company with 10% of the revenue of Assimil, then we're looking at US$590,000/year. The app store will take 30%, leaving us with $393,105. If we figure in the various costs of doing business, that corresponds to maybe a $200,000 salary. And I guarantee that somebody like Steve Rideout can get hired for $200,000, given the quality of his work and the wide range of his skills. That's well within the range of what Google pays their high-end talent.

I think Assimil is doing a huge mistake in not getting to new and empty markets. German and English bookstores are filled with courses for learners of many languages. But Polish or Russian native get just a few courses and for two very popular languages (so the assimil can be overlooked among all the courses, while other shelves in the language section are empty).

I think Assimil could do surprisingly well, if they noticed half europe is struggling even with English, let alone the other big languages. And has absolutely nothing for the less huge ones. And I suppose the same could be said about other regions too. After all, they are making some Japanese based courses too, if I looked correctly. Sure, no such localisation and no single course would be such a huge gold mine like their Anglais sans peine. But I definitely think they could profit a lot from filling various niches.

It is the same problem over and over again. The people who could profit from such resources the most (and therefore would be more willing to pay for them) are being overlooked by the companies. Instead, the companies are fighting for an already conquisted land, for an already oversaturated market.



So realistically, a company that's 10% as successful as Assimil can afford to bid against Google for one first-rate programmer-entrepreneur. Well, except I left out the marketing costs to acquire a new customer, which could easily eat up your profits. You'd need to know a few of things: your customer acquisition cost(CAC) (how much money you spend on marketing and advertising to get them to sign up) and your customer lifetime value (CLTV) (how much money customers give you before they stop paying). And you also need to know what your cash flow looks like—if it costs you $50 to acquire a customer, and if they spend $5/month for a year before "churning", then you only make a $10 profit in the final two months of the year. So once we take CAC and CLTV into account, we're definitely going to want to bigger than "10% as successful as Assimil."

(CAC and CLTV are probably two numbers that Duolingo is watching like a hawk. If they like what they see, they'll be able to get a big investment to grow their business rapidly. But their investment rounds aren't getting progressively bigger, which means that their investors aren't ready to hit the gas yet.)

Thanks for this explanation!



Adrianslont wrote:WE know how useful Readlang and similar tools are for reading and THEY have a lot of reading to do, which I am sure they would like to be easier!

…I think there is another part of the puzzle that is missing though - apart from the lack of awareness of such tools - the tools have to be RIDICULOUSLY easy to use and although Readlang almost fits that bill it is not completely there. It is ridiculously easy to use once you have your text uploaded but access to digital versions of texts is of course not ubiquitous.

I agree that ease of use is essential to reach a mass market. But there's a nasty little wrinkle here:

  • If you're teaching beginners, you can provide your own curated content. Basically, you might wind up with something like an app that combines an Assimil course with a built-in spaced repetition system. (Assimil, if you're reading this, I could build this for you for less than you might think. Combine that with a good US marketing person, and you could triple your revenue in a few years. This is a serious offer.) This could be very easy, and more effective than anything Duolingo or Memrise are promoting.
  • If you're targeting intermediate or advanced learners, then you have a problem: They don't really need courses, because they're already reading books, watching TV, writing essays, etc. But they still run into dialog that they don't understand, and words that they can't define. In other words, they want to use their own content. And this makes the software harder to use, because you need to access the content somehow.
Actually, there is at least one popular tool for intermediate and advanced students who are reading their own content: Amazon's Word Wise for the Kindle, which includes definitions for hard words, plus a flashcard system. Basically, if you're reading English, Amazon has a built-in readlang-like tool!

Your offer to Assimil looks great, I'd be among the customers (unless the terms and conditions are just as unlikable as those for their e-courses).
You are partially right about the beginners. I think the huge underestimated option is cooperation between the platform and the content creators. Imagine a publisher of a popular coursebook offering it with an official Anki SRS deck. Anki gets content and publicity on the back cover, the publisher gets a bonus over other publishers' coursebooks and the list can serve as an advertisement too ("like this Anki deck? wait till you see the whole thing!")I honestly don't know why is this such a stupid idea. But it probably is, since no publisher is doing this.

How nice of Amazon to have such a tool. Another thing I cannot have despite wanting to pay.

But the easiest way to simplify things is basically to package something like Assimil into a high-quality app, and aim solely at beginning students. Serious intermediate students might want to choose their own books and films or something unreasonable like that. ;-)

Actually, I've seen a czech app trying to teach languages with movies. Of course it looks like it doesn't have all your functions, but it is going in that direction. They are trying to connect it with a store, where you buy access to individual movies. They have only few now (and I am not gonna pay to try it out on English), but they are adding new ones. Hard to tell how to make such a service like subs2srs, connect it easily with a lot of content, while not being the owner of Netflix.
4 x

User avatar
emk
Black Belt - 1st Dan
Posts: 1620
Joined: Sat Jul 18, 2015 12:07 pm
Location: Vermont, USA
Languages: English (N), French (B2+)
Badly neglected "just for fun" languages: Middle Egyptian, Spanish.
Language Log: viewtopic.php?f=15&t=723
x 6330
Contact:

Re: ReadLang bought by Duolingo

Postby emk » Wed Nov 22, 2017 3:14 am

We keep beating each other to posting! So please forgive me if you've already responded to some of this in your previous post. :-)

Cavesa wrote:Emk, I highly value your experience, but I think there is a huge flaw in your thought here. And it is one you probably share with many of the creators and marketers of language learning tools. You are looking at the issue only through the eyes of an English native. I don't mean to offend you. But in your priviledged part of the world, extremely few people NEED to learn a foreign language. They need to pass an obligatory class, or they are curious, or they just vaguely thing "it could bring me a better salary, if I learnt Spanish." Extremely few people need it, because the whole world is learning English and because those positions for English+Spanish (or something) speakers have lots of applicants native of the other language and with excellent command of English as their second language.

I absolute agree that ESL is big business! I mean, the French subways have been covered with these posters for as long as I've been traveling there:

Image

And I know that there's an enormous hunger for better English. I once watched a French entrepreneur with maybe A2 English try to convince a panel of American investors to invest money in her company. She did this on stage, in public. It was agonizing to watch. English skills are incredibly valuable, and people will work for them. The money and desire are there.

Cavesa wrote:ESL is a huge monster with huge PR, it is extremely profitable business.

Yup. But as I mentioned above, nearly all of that money goes into textbooks and language schools. Not even Rosetta Stone's ferocious marketing engine has ever managed to grab even 1% of that money. Once you look at market for self-study courses and apps, you're looking a tiny slice of the pie.

But there's a subtler problem with the European market. According to various French software entrepreneurs that I've heard, it's comparatively easy to build a company in a single country. But once you try to expand into a second country, you wind up fighting the language barrier every step of the way. You need to open local offices, hire people, rewrite all your advertising copy, and restart your marketing campaign from zero. Even worse, you lose nearly all your word-of-mouth advertising when you move into a new country. And so European companies often get lazy, and they settle for being a big fish in the small pond of a single national market. Again, an excellent example here is Assimil: They've got pocket-sized courses that work better than any of the apps on the market, and they're been trapped in the French market for 3 generations now.

Cavesa wrote:So, both conditions are actually fulfilled. 1:enough serious learners-check. 2:learners willing to pay-check. The only problem: the sellers somehow cannot see this, as they are too focused on catching the attention of a usual English native vaguely interested in learning a language like "it would be so nice, if I could order my food in Spanish, when I visit Mexico."

I mean, I don't want to be too discouraging here. Somebody's going to figure this out eventually, maybe even in the next 10 years for all I know.

But on this forum, we already know of several low-tech courses which work very reliably, including Pimsleur, Assimil, and Michel Thomas. Any of the Czech learners you describe could use those to reach a basic conversational level in less than a year for a tiny fraction of what they spend on schools. So why doesn't Assimil, for example, already dominate the European market? It's cheap, effective, and available in most Western European base languages at least. If you can figure out how to succeed where Assimil failed, you'll earn literally billions. :-)

And just to be clear: I've been cheering for Steve Rideout for years now. I really wanted him to get filthy stinking rich. I was rooting for lang-8 and Khatzumoto and LingQ and that Verbling group chat feature, too. But they were all selling to intermediate learners, which is a much smaller market than the dabblers.

Cavesa wrote:I think the huge underestimated option is cooperation between the platform and the content creators. Imagine a publisher of a popular coursebook offering it with an official Anki SRS deck. Anki gets content and publicity on the back cover, the publisher gets a bonus over other publishers' coursebooks and the list can serve as an advertisement too ("like this Anki deck? wait till you see the whole thing!")I honestly don't know why is this such a stupid idea. But it probably is, since no publisher is doing this.

Actually, book publishers in general tend to be hopelessly clueless about technology. So it's perfectly possible they're just being stupid and they're passing up free money. A book publisher's usual approach is to outsource technology stuff to an incompetent, overpriced contracting firm, and have the project end in disaster after burning a giant bonfire of money. Someday, some internet startup will eat their lunch.

But also, the publishers probably don't care. A publisher's major customers are schools, and schools are conservative, bureaucratic and broke. Educational technology is a notoriously difficult market—it has all the paperwork and headaches of selling a huge multinational company, but without the giant checks. Actually, you might have more luck trying to convince a private language school to advertise their use of spaced repetition software. Maybe somebody like Wall Street English would know how to brag convincingly about how their advanced technology makes them better than the other schools? Surely some of those schools would be willing to throw aside the bureaucracy, and they clearly know how to market. Maybe partner with one big private school chain per country or region?

I'm pretty sure there's some way to make money off of top-quality language learning tools. I'm only spending my time writing down all the obstacles I know about in hopes that some smart entrepreneur figures out how to reliably sell tools like readlang!
4 x

User avatar
Adrianslont
Blue Belt
Posts: 827
Joined: Sun Aug 16, 2015 10:39 am
Location: Australia
Languages: English (N), Learning Indonesian and French
x 1936

Re: ReadLang bought by Duolingo

Postby Adrianslont » Wed Nov 22, 2017 4:19 am

Awesome post Cavesa. Awesome post emk. I really enjoyed reading your insights.

One thing I would really like to clarify is that I can totally understand Steve Rideout taking a job at DL and selling an unprofitable Readlang to them. If I were him ie talented, hardworking and young, I would probably do that, too. Steve obviously thought it was worth trying to make his fortune with Readlang and I am sure he developed a great skill set working on it. I wish him all the best - I just hope for my own selfish reasons that you are wrong when you suspect that DL will euthanise Readlang. And I think DL with their experience and market penetration would have a better chance of making a success out of Readlang than a one person outfit. If it disappears I guess Steve Kauffman will get my money because I am intimated just reading about LWT!

One of the biggest issues - as discussed in this thread and others - the large slice of the market is beginners. They are the low hanging fruit, easier to develop for and the segment where most potential profit lies. I’m just hoping someone comes up with a concept for intermediate and advanced learners that a business can see a future in. Not every business has to be Google or even Assimil. I’m just saying that the market for non-beginners IS also huge - if not as huge - and Cavesa has detailed where much of the market lies.

I find the whole subject of software development (philogenesis not ontogenesis) fascinating. Spreadsheets for Apple 2 are widely acknowledged as being the point when personal computers took off, then there was word processing and then presentation software. And Microsoft spent the 90s maturing the whole Office suite and Adobe captured graphics and publishing. The big guys picked areas with lots of users first to maximise profit and fought for market share and why not - but lots of smaller developers have filled lots of gaps for niche products for dentists or whoever and made a great living.
4 x

User avatar
Adrianslont
Blue Belt
Posts: 827
Joined: Sun Aug 16, 2015 10:39 am
Location: Australia
Languages: English (N), Learning Indonesian and French
x 1936

Re: ReadLang bought by Duolingo

Postby Adrianslont » Wed Nov 22, 2017 4:53 am

I am so slow. You two posted two more awesome posts before I got mine up.

I have just a few more things to add:

1. I don’t think Assimil have “failed” emk. Did you say they employ fifty people? I imagine the owners living comfortable lives in comfortable French homes, eating fine food and drinking fine wine. This is all my imagination of course but they have been around for three generations so I don’t think “failed” is an appropriate word. They’ve been doing well enough to keep going. I think you make good points for untapped potential, though, Cavesa.

2. Emk, I think your idea of developing an app for Assimil is awesome. And yes maybe selling decks or apps through a college is a better way to go.

3. Emk, I loved your description of outsourcing development and burning money - I’ve seen it happen.

3. Interestingly, for Indonesian, I have seen movement in the area of “bundling” anki decks with classes (face to face and online classes/courses). Neither example has set the world on fire - they weren’t really trying to - but it has been happening for a few years now at least. The small institute I have studied with in Indonesia has a couple of small anki decks they give you - or they give you paper cards if you prefer. Just single words and only a few hundred cards but they have audio and actually nice audio quality. Also, The University of Hawaii has an online course called The Indonesian Way and that has about 8 anki decks if I remember correctly as part of the course.

Cheers, I’m going to do some actual study!
3 x


Return to “Language Programs and Resources”

Who is online

Users browsing this forum: No registered users and 2 guests