Addressing the Decline of Open Source LMS for #altc Discussion
//There is an interesting discussion on Twitter #altc that is worth addressing, regarding the decline of open source solutions in EdTech, with LMS as a key example.
Laura Czerniewicz posted in reply:
Michael Feldstein and I have dealt with this subject at e-Literate and on this blog, as Laura references, but it might be worthwhile to update the data. I do not plan on address the “why” and “what can be done” questions in this post, instead focusing on hopefully more useful information as input to the discussion.
First a note on the data for context. Our data set and analysis behind our LMS Market Analysis service for higher education is in partnership with LISTedTECH. The data are very school or institution-centric. We do not look at how many implementations are claimed by LMS providers, rather we look at institutions and check which LMSs are their primary or secondary systems. We use the public data from LMS providers as quality checks on our data, not as the source. We have data on over 11,000 institutions worldwide, with roughly 4,500 in the US and Canada. The primary regions we cover are North America, Europe, Latin America, and Oceania (Australia and New Zealand). We also have data for the Middle East, Africa, and various sub-regions of Asia, although we do not officially cover those in our market analysis. There is new data we plan to release soon on more than 3,000 K-12 institutions in the US and Canada.
In our Summer 2019 report, we showed the installed base in different global regions, based on number of institutions.
Regarding open source LMS:
- Moodle is dominant, both as an open source LMS solution and globally as the most-implemented LMS solution.
- The adoption of Moodle in the North America (25%) is different and much lower than other global regions (67 – 76%).
- As seen in our previous post, if you scale by enrollment in North America, Moodle has even lower (12%) market share in North America.
The question in the #altc discussion, however, seems to be more about the historical trends, not the installed base. And the question also seems to focus on the markets outside of North America. Laura referenced this post which shows the clear decline of Moodle and Sakai in North America, but what about other regions? The following chart shows LMS market share by institution for all of our higher education data outside of North America (including areas in the Middle East, Africa, and Asia that we do not officially cover yet).
Some notes:
- The data set is more complete in recent years, and open source implementations at smaller schools are difficult to find. Therefore there is likely a bias with more Moodle schools in recent years. In other words, there may be somewhat more of a decline that the data initially shows.
- The majority of “open source” usage is from Moodle, but this category also includes Sakai, Blackboard OpenLMS, and other open source solutions. This category does not include Canvas.
- Outside of North America there is a slight decline in open source LMS since 2015, or at least a plateau.
- I don’t have time this morning to create a separate chart, but in general open source LMS tends to drop off for more mature markets (in terms of higher LMS adoption within a region). One way to read this is that open source LMS markets are shifting over time towards developing regions.
I look forward to the continued #altc discussion, but hopefully this data will help with context.
Free and open source solutions are *not* on the decline in any way. They’re burgeoning. What many mistake for decline is the fast *natural selection* process which takes place in the FOSS world: when a better idea comes along, there’s less *social inertia (e.g. long term support contracts, sunk costs of integration, budgetary cycle, procurement processes, etc.) to fight – people can rapidly adopt better technologies. At the OERu, we think the LMS has done its dash and we’ve already moved on (but still 100% FOSS!). To be fair, LMSs like Moodle are still the best tool for a few jobs, and we incorporate them into our software mix – but as an adjunct rather than the main player.
We’ve opted for a “loosely-coupled component-based” model instead: https://tech.oeru.org/many-simple-tools-loosely-coupled We’ve termed it our NGDLE – here’s a lot more info on it: https://tech.oeru.org/2018-update-oeru-technology-stack including a case study showing how extreme the benefits of this model can be… our cost structure and flexibility will make most folk who’re stuck maintaining legacy Microsoft-based IT infrastructure weep (as well they, and their institutions, should).
Dave, thanks for comment and links. If only you didn’t use the NGDLE language . . . 🙂
I appreciate the sentiment of reduced social inertia and of movement to newer concepts such as loosely-coupled tools. I’ll take a deeper look at OERu usage.
However, these arguments do not support idea that there is no decline in any way. Certainly in North America, the usage of open source LMSs is a very clear trend, and as shown above there is a small trend, or at least plateau, outside NA in higher ed. It’s all well and good that there are different concepts benefiting from FOSS, but the broad adoptions still matter.
Well, yeah, open source LMSs appear in decline (based on the research you present). And I’m willing to agree that educational institutions are also moving away from FOSS… they’re idiotic to do so (I don’t believe in candy coating it – they’re just dumb to be locking themselves into proprietary monocultures). Overall, outside the specific market of educational technology, FOSS is going from strength to strength – which is why corporations like Microsoft are desperately trying to make over their horrible historical image… sadly, they’re having some success because people are painfully gullible.
Intrigued to know why you’re cringing about NGDLE – maybe it has different connotations in the US than in NZ… I’d be keen to understand that.
Dave – again, thanks for input.
My cringing about NGDLE is that, like many EdTech terms (e.g. see personalized learning), it has morphed from a specific usage driven by EDUCAUSE project funded by Gates Foundation into a marketing catchall phrase. The gee whiz “next generation” is overwhelming the specifics of loosely coupled ecosystem.
In your case (OERu), you clearly hit the loosely coupled aspect, but I am curious about the core platform that was key to NGDLE. The idea there was that you need to core system for non early adopter types, so that the basics are all there, yet other tools / architecture allow replacement, addition of specific functionality (specifically following the ‘do one job well’ attribute you reference). Q. What is the green center of OERu tech stack? Does that imply rosters only, or is there something else there?
To be clear, I’m cringing about the marketing overuse of term and not arguing that EDUCAUSE got “it” right and must be followed by all. Just wondering if NGDLE usage in your case has same meaning as original initialism.
Heh heh – I don’t have much nice to say about the Gates Foundation (I emigrated to NZ from Seattle) and we actively try to avoid association with them. (I have been convinced by Anand Giridharadas’ thesis on billionaire philanthropy: https://www.youtube.com/watch?v=d_zt3kGW1NM )
I think our aim is to usefully reclaim the term NGDLE, because, honestly, that’s what we’ve got. I don’t think we have a “green core”… The whole idea is to not build a structural dependency on any particular component.
If anything, our current platform has 2 main foci: our MediaWiki (https://wikieducator.org) for course development (we only allow “free cultural works” to be incorporated into the materials) with version control and templates to facilitate the development of rich, re-usable OERs, and our WordPress multisite (https://course.oeru.org) which presents courses (however, our partners can target literally any WordPress site as the presentation tool – courses are automatically “rendered” as mobile-friendly WP sites direct from WikiEducator – see https://tech.oeru.org/oeru-mediawiki-wordpress-snapshot-toolchain for how that works) to learners. And we’ve already got in mind better replacements for both MediaWiki and WordPress – but we’re taking our time.
As a matter of principle, we make *all* our materials visible even to anonymous learners. The only reason people register and log into our sites is to have persistence of their data, a record of their participation, and the ability to participate our various social media (Mastodon, Discourse, Rocket.Chat, and our own WEnotes which ties everything together – see https://tech.oeru.org/wikieducator-notes-oerus-course-feed-aggregation-and-messaging-system).
The rest of our stack is comprised of learner-facing resources that enhance pedagogy options (we use tools like SemanticScuttle, Hypothes.is, and we’ve recently introduced H5P widget integration into our OER assembly workflow), tools for collaboration (both learners and educators – Discourse, Rocket.Chat, Mastodon, Etherpad, Jitsi Meet), adminstration tools (like Mautic, NextCloud, and CollaboraOffice), and monitoring of system health and usage (all anonymous – like Matomo, YourLS, Icinga2, and others).
In future, we’ll be separating out our learner identity management (currently that’s handled by our WordPress multisite) into its own directory system. That, in turn, will facilitate our introduction of Single-Sign-On into all the services we provide, making everything even more seamless from both learner and educator perspectives. Every single component of our system is fully Free and Open Source Software.
Also, I should’ve mentioned – we also use Limesurvey to engage with learners and educators to gain insight into their preferences. We have a Moodle, too, which we use exclusively for course assessments used to award participation and completion certificates (including digital badges).
I preface this as a) my personal opinion and not connected to any past or current employment and B) having worked in edutech on both the University side implementing and running system, and aon the vendor side I can say in larger institutions the move away from open source (in APAC at least) is a strong trend. This is exactly the same trend the IT industry saw with email, website CMS and similar tools saw 20 years ago. First everyone had a bespoke niche tool that they built or cobbled together because that was all there was or commercial was just too expensive. Then reliability became an increasing concern as it became business critical, commercial options became more of a commodity and cheaper, then they realised the business risk of bespoke tools when the small number of staff leave and take institutional knowledge with them. Commercially supported tools where you have certainty for at least the contract period are increasingly attractive. Very few unis run their own email servers any more, LMS are going the same way.
Open source has its place and is still strong in the many tools out there, I am currently trying to implement an open source tool, but the business risk is high. As an example this tool has 3 code branches which are barely documented, no road map, no security reviews (and it handles some student PII), no guarantee of future updates unless the institution is willing to dedicate staff to learn and maintain it. Moodle clearly doesnt fall into this class of tool but many of its plugins do, they often are only maintained while someone has time or interest. This is both a reputational and operational risk to an institution for a business critical system, even if it’s only a small part of it.
Out of interest, Brett, how does your institution mitigate the risks of
a) an IT supplier failing to meet contracted obligations,
b) an IT supplier changing its terms of trade (e.g. jacking up its prices),
c) an IT supplier being acquired or going out of business?
How do you explain to learners the sacrifices in privacy, waived rights, and data sovereignty when forced to accept terms and conditions of 3rd party proprietary tools required by your institution for them to complete their studies?
How do your learners feel that their creative work and educational history is often contained within proprietary data formats or on proprietary cloud services, where they only way they can access it following their study is to pay that proprietary vendor to access it. Do any of them point out that this is indistinguishable from a data “hostage situation”? All food for thought.
Brett – thanks for input. yes, that matches what I am seeing in general, and not just APAC.
Dave – sorry for delay in comment being posted, as WordPress flagged as spam that had to be approved. Maybe next time you won’t describe better replacements for WP. Thanks for description – quite interesting.
No problem, Phil – happy to share 🙂 – by the way, in my experience WP Akismet spam filters trigger on external links in comments.