[Site Feedback] Better visibility for new courses, issues with the app, accountability for absentee course creators -- with possible solutions to each

First, a simple question: Is anyone else learning a course on their desktop, but are unable to practice the same course on the app version? If anyone else is experiencing this, please reply.

Secondly, I’d like to talk about the implications of this issue.

My experience has been that relatively new courses (less than 3 weeks old) do not appear in language results on the app – and these new courses cannot be accessed on the mobile app. This seems like a major disincentive for content creators.

For example; user @MrMac (not a community member) created a new Cantonese course: http://www.memrise.com/course/1024724/another-lesson-in-cantonese-with-audio/.

It’s exceptional. Takes a novel approach (starts by learning ‘elemental’ characters), has accurate audio from a number of contributors, and it’s robust.

However, the vast majority of Memrise users interested in Cantonese language will never see this fresh content. For the time being, only people that stumble upon it on the community forums (which, I imagine, is a very small percentage of users).

Instead, they’re only going to see old courses. Ones with absentee course creators, a litany of errors, and largely irrelevant content. They’re essentially dead courses (@Hydroptere, I’ve seen you bring this up so I’m dragging you into this).

From what I can tell, the most visible courses on Memrise only have this status due to their course age, and a large number of users which have initially signed-up to learn it.

Though you could argue relevance by “seniority” and the aggregate number of learners, as well as current learners, the current model seems to make new course creation feel like a futile effort.

I’d like to suggest some possible solutions that could factor-in to “ranking” courses in the results

It goes without saying that most of this would require development work. Nonetheless:

  • Accountability - a flagging mechanism to open discussion on potential errors in content, which requires some address from course contributors in a reasonable timeframe.
  • Allow a certain amount of newer courses to show up higher in results. If they get reasonable exposure and don’t stick, they can gently slide back down.
  • Assuming it’s not a “bug” that new courses can’t be learned on the app or appear in results, having more clear parameters about what makes a course “app eligible” would be fair.
  • Simple search functionality would be ideal for the app. This would allow people to make more specialized search queries (for example, Cangjie input).

Any objections to any of this? Thoughts?

Thanks for your time.


I would also love to have more and better ways to search for courses.

Unless you’ve seen someone recommend a course in the forums, there isn’t a good way to evaluate how good courses are without spending time working on them. When we search for courses, the only thing we can really see is how many people are in the course and a rough estimate of how long it might take to complete it. What do we know about a course with a lot of users? It probably shows up high in the search rankings, and it has probably been around for a long time. That’s not a guarantee it’s a good course.

What do we know about a course with few users? Nothing—it could be a mediocre course… or it could simply be a new one.

Does the course have audio? You can’t tell by looking at the search results. Does it have many typos? Is the content presented in an order that makes sense and aids in learning? You have to spend time at it to find out…

I’d like to see users able to rate courses, perhaps after completing 1/4 of the course or 200 items. Then let us search for courses based on those ratings.

It would also be great to be able to sort by most popular, length (most items and fewest items), newest (recently published), oldest, etc. We should be able to filter out courses with no audio or filter to see only courses that specifically test audio. We should be able to filter include only or to exclude no-typing and strict-typing.


I’d like to see users able to rate courses, perhaps after completing 1/4 of the course or 200 items. Then let us search for courses based on those ratings.

I’m in full support of rating courses, in whatever permutation.

Perhaps ratings on different categories (audio quality, usefulness of content, logical course progression) – having separate rating tiers could make it easy to also incorporate the filters you suggest.

In the meantime, I’m hoping we can get a clear answer on the requirements for courses to appear in the app.


I am absolutely against rating courses. I take a number of courses which are very good, but with few users - or they are hard to find, or their level is too high for the average user etc. I stumbled over them after hours of search. There are many reasons for my opinion, among them the fact that a rating system would üplay in the hand of those having nothing to do but play the superior on others… would encourage silly power games, and would be a major disincentive for course creators. Btw, for those who know the old forums, do you remember the number of users requests concerning mistakes in forums of the in-house made courses (i.e. official Memrise courses?) - how would you rank the in-house courses - without the work of users and contributors?)

I am in favour of a system of marking courses: present/absent creator, sound/no sound. This kind of stuff.


I am all for rating courses. I stopped using the user created courses because i’m not interested in wasting time working on a language only to find out half the words were wrong. I know there are good user created courses on here but without any evaluation system there is no way of knowing if you are getting a course from someone who is fluent in that language or someone who just thinks they know what they are doing.

1 Like
  1. someone who is fluent in a language has virtually no reason to make a memrise course which is under her/his level.

  2. I, personally, I am not that risk averse… I prefer trying courses (and quit the most) to the “I am native, I know/my wife said/google translate said/I simply can ruin your nerves and work and that is why I’ll do it/etc”.

  3. every now and then the memrise user herself/himself can try and use other means than memrise, than s/he can spot mistakes her/himself - instead of expecting some other people to deliver everything for free, just like that.

  4. The “official” courses have mistakes as well - a you can see from the old and the new fora. Also, simply having a “native” to correct this and that is not automatically proof of quality.


The reason is to help others. I have multiple courses that are under my level. Almost all.


I honestly can’t interpret half of what you say here.

I’ll just speak to your first point. It is pointless to have people who do not understand a language make a course in that language. This is exactly why Memrise seems to have earned a reputation for poor quality. They are trying to sell a product here so they should consider making it a product worth buying.

Who would want to take a course buy some Joe Scmoo who simply made it up? Yes I absolutely want courses by people who understand the material they are presenting. I suspect most people do as well. Certainly from all the complaints I read from people who left Memrise that was what they were expecting.

You don’t have to do courses from Joe Scmoo. It is your choise.

1 Like

That’s the point. Without rating courses we do not always know who is qualified and who is just making garbage courses.

I suspect the people making courses are the ones who fear the rating idea. Maybe if they spent that energy improving the quality of their courses they would have nothing to fear.

Here I disagree with you. Have you made any courses?


What part are you disagreeing with?

So I guess you have not…

Let’s make a deal. You go and create a great quality course and then we evaluate you and then you can complain as much as you want if you want


Memrise is not complete without the ability to rate and review courses. You could say the product is broken at the moment. We need to be able to sort on course quality and popularity. This is such a simple and obvious idea that it does not need any further argumentation.

Memrise staff reading this should think to themselves: how can I make this happen within the next 6 months? Perhaps use an external module?

Here are some ideas from an earlier thread:


I didn’t respond to that because it is irrelevant.

You are clearly just taking this personally so maybe you need to step back and consider this about making a better product for everybody and not just a platform for you to showcase your amazing and infallible talent talent as a course creator.

Memrise staff should also remind themselves that they are selling a product. A product that only some people are paying for while they subsidize the free version for the other users. Quality matters.

As I’ve mentioned previously, I have heard from many people that they don’t use Memrise any more because of quality issues. They lost a lot of serious learners that they would have been able to turn into paying customers if it hadn’t been for this issue. And some who were paying for it.

1 Like

I want more sorting methods for course display when new users search. Courses are currently sorted by number of learners. This has an ‘age bias’ on courses (old courses have more learners) and doesn’t put the best courses at the top, plus audio courses should automatically be at the top. The only way to currently find audio courses is to open the course and try it.

Proposed sort types:
Recommended” - Senior members of that language community pick the best courses. Good definitions and high quality audio.
Audio” - only courses with 80%+ audio are shown.
Learners” - Number of learners ie the current technique.
Most words” - Biggest courses to smallest


There was a nice long discussion of this on the old forum, back when memrise first got rid of the ability to sort courses by popularity or “newest”. Fortunately archive.org still has a copy of at least one of those threads where we discussed this:


I think there were other discussions of this but I don’t know if they got saved. However, I do remember that memginee collected people’s suggestions and posted that as a suggestion on the uservoice forum, which was memrise’s official forum for feature suggestions. It got lots and lots of votes and was one of the top suggestions for a long time.

Unfortunately, as with pretty much all suggestions from users since summer 2014, memrise ignored it completely as far as we know.