Treating MOOC Platforms as Websites to be Optimised, Pure and Simple…

A month or so on from its PR launch, and with a steady trickle of press mentions since then (though no new updates on the website?), I’m guessing that the folk over at FutureLearn must be putting the hours in trying to work out what the platform offering will actually consist of, or what the sustainabilitybusiness model will actually be. (I have no inside information on the FutureLearn project…)


One of the things I have sort of picked up from online glimpses of things said and commented upon is that the USP is going to relate to the quality of teaching/pedagogy (erm, I think?!). I’m not sure if “proven” learning designs will be baked into the platform, constraining the way courses are delivered (in which case, there’s likely to be something of a bootstrap problem in getting the first courses out if they have to wait for the platform?) or whether the quality will flow “naturally” from the fact the the courses will be provided by British universities (?!), but if innovation is also to flow, it’ll be interesting to see how it’ll be supported…?

…and whether it will be done through “open” means? (I can haz API? But what would it do?!?) If it is built up from open code, I wonder to what extent it might draw on code and ideas used in other learning platforms (for example, Moodle, to which the OU is already a core contributor, I think?) or Class2Go) as well as drawing on learning from whatsoever folk managed to learn from the OU’s other open learning builds – OpenLearn/Labspace (content and community), iSpot (community and reputation), Cloudworks (community and resource sharing) or the very many expensive attempts at SocialLearn (wtf?!) that never saw the light of day? I can’t imagine a FutureLearn offering being based on the Google Coursebuilder, but it wouldn’t surprise if it ended up with something being bought in… Time to start watching the tender site, maybe, though surely that would knock any start date back too far?

One thing that would be nice to see would be project using something akin to the open, agile development process used by the @GDSteam, which is opening up the backend to View Source as well as the front-end…

I also wonder about the extent to which it might be possible to reuse ideas from commercial website design and development in the way the site is architected. This will be anathema to many, but I wonder just how far the idea could be pushed? Start with the idea of analytics, and define funnels for how folk might be expected to move through course units. Associate activities with some sort of intentional action, such as popping items into a shopping basket, or maybe the equivalent of 1-click purchases. Making it through the to end of a course can be seen as completing such sort of purchase (chuck in some open badge framework badges as a reward for good measure;-). Ad-delivery mechanisms can be rethought of as personalised content delivery (eg contextual content delivery, banner ads as signage or email-pre-emptive ads). Use search data to help refine content pages, and A/B testing to try out multiple variants of course materials and exercises (weak example). (I have never understood why the OU doesn’t engage in A/B tested delivery of course materials as a matter of course? OU courses are delivered at large enough scale, and containing more than enough content, to trial different ways of delivering content and assessment without jeopardising overall outcomes for any individual student.)

All of the above – search analysis, web analytics, contextual content/ad-serving and A/B testing – can be managed through ad servers and Google Analytics (and to a lesser extent Piwik, though they are open to additional contributions), which could provide a minimum-viable product tooling basis for a testing and analytics framework that’s ready to go now? Such an approach is far too scruffy and ad hoc, of course, for a “proper” platform project…

PS by the by, I notice that JISC Advance’s Generic eMarketplace (or GeM) for Work Based Learning (“gemforwbl”, or looking at the logo, “gee em for weeble”? (will it wobble? will it fall down?) is now open and ready for business… and as for the logo, what on earth is it supposed to represent?


Answers in the comments, etc etc, please…

PPS As ever, the opinions expressed herein are not necessarily even reflective of my own, let alone those of my employer…;-)

Author: Tony Hirst

I'm a Senior Lecturer at The Open University, with an interest in #opendata policy and practice, as well as general web tinkering...

13 thoughts on “Treating MOOC Platforms as Websites to be Optimised, Pure and Simple…”

    1. @Alan I have *no idea* what they will – or are – doing… I just hope they don’t repeat the repeatedly false started path that was followed by the SocialLearn project…

  1. Hi Tony – thanks for this. I’ve been thinking about mooc platforms too and almost wrote a post yesterday called “the VCs new plaform aka the emperor’s new clothes” – but didn’t get much past the title. Like you I’ve been wondering what “new and exciting” things FutureLearn or any other “mooc platform” will bring to the party and how different it will be to the obviously “dull and boring” systems we have just now. There is an opportunity to do something new and as you’ve outlined there a lot that could be done simply with exsiting “stuff”. I fear tho’ that the pedagogical innovation will be swept aside under the guise of analytics and assessment which will be primarily for the institutions/business side of things and not the teaching and learning.


    1. Hi Sheila – 6 years on from and I’m still laregley none the wiser about how any online courses I work are performing as content. A lot of the hype around learning analytics appears to be based on the notion of tracking performance in some way, and tuning interventions based on interaction log analysis to try to influence those metrics. Being a bear of very little brain, I could never and still can’t understand why educational institutions and educators themselves are so resistant to treating educational material as “just content”, and looking at ways of improving engagement with it in the same way that commercial sites use web analytics to evolve site design.

      It might seem irrelevant to the loftier goals of the learning analytics folk, but I still think it could be useful for a content producer/editor to know that no-one ever visits page x, that the bounce rate on page y is 95%, and that of the seven links on page z no-one has ever clicked on six of them.

      As a soft way in to A/B testing, just tweaking link text or the order in which resources are presented on a page to see if influences the click-thru rates is unlikely to jeopardise a student’s learning outcomes to any significant effect (unless the unfollowed resources are actually toxic to a student’s learning!) and may encourage students to explore otherwise unused resources, which the educator presumably merited enough to include in course materials?

      I’m not saying that treating content and flow in VLEs as “just a website” should be to the exclusion of “higher level” learning analytics, but it just seems strange to me that educators don’t even think of the stuff they publish as content that can be tweaked in ways that might influence the way it is used, or that can be edited in response to how it is actually being used.

      1. I couldn’t agree more with your comments about educational content. I think part of the problem is that the role of content in enabling learning tends to be either undervalued or over played. If you focus too much on facilitating access to improved content you run the risk of being accused of confusing content with learning. I could go on about this at length bu I’ll restrain myself :}

    1. @Tim I know the OU contributes modules, I wasn’t sure whether it contributes back to the core… (and I was too lazy to check.. so thanks:-)

      PS are you involved with FutureLearn platform at all? Any idea what’s on the cards?;-)

  2. If I had any knowledge of what was going on with FutureLearn, beyond what is already in the public domain, then should probably should not post it in a public blog comment, and I won’t.

    I will say what I thought about the technology side pretty much as soon as I heard the original announcement, and I have not had cause to change these views:

    In the OU VLE, based on Moodle, we have a single code base that, appropriately configured, runs the full module web-sites for OU students at one extreme, and the somewhat limited open-content offerings that we currently have in OpenLearn. The same code-base (configured differently again) also runs the Qualifications online sites. Additionally, all around the world lots of other people are using Moodle to facilitate all sorts of educational experiences. So, I am sure that, appropriately configured, Moodle, particularly the OU’s customised version of Moodle, could be used to deliver MOOCs. (Perhaps the most worrying area is the scalability of Moodle, beyond the level the OU already runs it at.)

    Also in Moodle’s favour is that the OU VLE development team has a lot of experience of customising Moodle and writing plug-ins. So, if there are bits of functionality that FutureLearn wants, and which are not there yet, we could almost certainly develop them. We also know a lot about how to set up Moodle, both technically and pedagogically. So, if FutureLearn wants to buy this expertise from the OU, and if they want something quick, then this seems like the most viable option.

    However, FutureLearn are an independent company. The decision is theirs, and there are many other platforms in the world that are also capable of doing the job. You list some of them.

    I think the more interesting question, as you say, is, what does a MOOC with particularly high “quality of teaching/pedagogy” look like? I am a big fan of “Martin’s 5 laws” ( as a model of pretty much any sort of learning. Certainly, FutureLearn, and the academics who contribute to it, will have to learn to learn the answer to that question. They will have to learn by doing it a few times, and feeding back amongst each other, and getting feedback from their students. They also need to get some things done quickly, to meet whatever launch date they set themselves, which does not give much time for iteration on the design of the first few courses. We need to hope they can find the right person to take some inspired guesses on the pedagogy of the first courses, and then give themselves a change to learn from experience and improve things further.

    Just to reiterate, all that is my personal opinion/speculation.

    Oh, one other point. I think that the longest-running MOOC in the world is the ‘Using Moodle ‘ course on ( which has been running since before the word MOOC was invented. That was set up as part of Martin Dougiamas’s PhD, to investigate the question “Can an open source community learn to build and use a VLE by following the 5-laws? That is by actually doing it, feeding back to each other and learning from each other, as they go along.” My personal observation over the last 5 years leads me to conclude that the answer to that question is “Yes”.

    1. Tim – thanks for that considered opinion. I was only teasing re: asking about FutureLearn details;-) though I do wonder about extent to which OU devs might be pulled into it vs 3rd party contractors, particularly in respect of extent to which FutureLearn may develop/persist/scale. The question about how FutureLearn and OU systems might/could/should feed off each other is also something worth considering, versus, for example, FutureLearn being allowed to develop completely free of OU process model and costing model encumbrance, or being allowed to try out something completely different and that is too “risky” to deploy as part of the OU’s formal, credit bearing offerings.

  3. Really interested in this, thanks Tony! Even though I’ve discovered this post 9 months late…

    I come from a web development background, so I’ve always shared your surprise that universities don’t want to do more deep optimisation work on their sites. It’s almost like it’s considered crass to take part in such ‘commercial’ practices.

    I’m running a lot of short-form courses just now and doing some deep analysis on how people are interacting with them. I definitely want to set up Google Analytics on them after your prompting here (for some reason I hadn’t thought of that before, despite using GA every day on my commercial sites) and start to look at navigation paths and goal setting. I just read Martin Hawksey’s post on event tracking in elearning too, and I think the two combined will work really well to spot places where we can improve the path through a course, increasing retention and engagement.

    Quick question, you mention APIs for getting into open course data – since this post, have you come across any MOOCs or other online courses that have opened up their analytics or other access data? Seems like it’s something institutions would do, but I haven’t spotted any yet…

    Thanks again for a great post!

    1. Hi Colin – glad to hear there’s at least one other person out there who doesn’t understand education’s unwillingness to view educational materials as just another form of content that can be optimised (I guess one reason is that the LEarning Design folk don’t want to see the stats that tell them folk aren’t using the materials as they were designed?!;-)

      Re: MOOC apis – @mhawksey is probably the best person to ask, and perhaps also @patlockley. As @futurelearn are currently sprinting on things for their course launches next month, it may be worth nudging them with the idea that their platform might benefit from some API access?!;-)

      1. Thanks true – who wants to see exactly how neglected their learning materials are :)

        Thanks for the contact pointers too – I’ve been chatting to Martin about it, but haven’t met Pat so far. I’ll drop him a line. Cheers!

Comments are closed.

%d bloggers like this: