2017 eLearning Predictions: Updated Hype Curve

Our 2016 hype curve was a big hit, so we’re doing it again this year. Here is the short, graphical version:

Our 2017 eLearning predictions set in terms of Gartner’s hype cycle.

And below the fold, the long version, broken into the five stages popularized by Gartner:

Innovation Trigger

As I wrote last year, “Online learning lags behind other industries in adopting new technologies. This is not because people in online learning are not interested in new technologies; it’s because we’re not a high-revenue industry compared to giant industries like consumer goods.” This is still as true as ever, and it explains why there’s a four-to-eight-year lag between technology trends as a whole and their application in eLearning.

Artificial Intelligence and Machine Learning: Our developer David Wipperfurth writes, “AI and ML don’t really fit well on the hype curve, because they will have multiple in/out fluxes with different technologies and uses. AI/ML is essentially too big and vague a category to put on such a chart.  It would be like asking where economics fits on the chart.” OK, granted. Writing specifically of intelligent apps, David adds, “This depends on what we’re talking about.  Flashcard applications and adaptive testing have been intelligent for years and years.  Lots of people probably make claims about how ‘intelligent’ their apps are.  I’d say adaptivity (outside of tests), procedural generation, and data mining are not widely utilized by mainstream pedagogy, so for those cases it might be at technology trigger.”

Following David’s lead, we define artificial intelligence and machine learning as computer-generated algorithms, or “procedural generation”: the idea that the computer derives its own rules rather than relying on a human to program its rules. This kind of machine learning is used in applications like Google’s speech recognition and translation software. The eLearning applications of this technology are very much in their infancy; we expect them to be pioneered by firms like Google, IBM (which markets its Watson technology in connection with various LMS platforms it has acquired over the years), Microsoft (which just invested in the lifelong learning space with its purchase of LinkedIn/Lynda), and even Amazon. We were surprised to find Amazon recruiting aggressively at the eLearning Guild’s DevLearn conference. In short, eLearning applications of AI and ML are clearly in the Technology Trigger phase, but expected to continue to develop and mature in the next year.

IMS Caliper: Caliper is the latest standard from IMS. A newer standard than even xAPI, Caliper is somewhat less ambitious than xAPI in the number of use cases considered and APIs specified. This is a possible reason to expect quicker adoption. Tim Martin of Rustici praised the Caliper design process at DevLearn. As Martin is one of the xAPI originators, this could foreshadow future synergies between the two standards and help xAPI get out of the Trough (see below). To understand the difference between the two standards, it’s important to note that IMS and Caliper are driven by higher ed and thus of particular interest to the Educause crowd. IMS’s best known standard, LTI, allows textbook companies to package online course content and ship it with their textbooks, regardless of the implementing university’s LMS. LTI packages vary from SCORM packages in that the typical content model follows higher education’s synchronous, weekly, semester-long format rather than the asynchronous, one-time, hour-long format of much corporate and military training. Caliper keeps these use cases and users in mind while improving on previous IMS standards in the granularity of data captured and in use of an xAPI-like model for interoperability across multiple types of tools and platforms: mobile apps feeding back in to the LMS and the like. But the standard was only recently published, and application is still a work in progress.

Internet-enabled Physical Simulators: There’s not much more to add from what we said last year, other than that we’ve  participated in a pilot that used xAPI to connect a Laerdal simulator to a learning management system, and we’ve read of another similar pilot. The use cases are clear, but the technology is not so easy to deal with right now. Standardized connectors need to be built. But because of the pilot activity, we’re moving this trend up-slope a bit.

2017 eLearning Predictions

CHEST-Web Courseworks xAPI pilot

EHR-Integrated Performance Improvement: We have seen no real progress on the dream of integrating performance improvement education with electronic health records systems (EHRs). But the hype certainly increased in 2016 thanks to the long-awaited publication of the final rule on quality payment under the Federal Medicare Access and CHIP Reauthorization Act (MACRA). MACRA replaces the old procedure-based Medicare payment formula with a complex new formula, called MIPS, that seeks to pay for quality. Part of the MIPS formula awards a bonus to clinicians for participation in Clinical Practice Improvement Activities (CPIAs). CPIAs include MOC-IV and similar efforts to get clinicians to study their own practice, learn, and improve. Payment bonuses for quality and practice improvement represent a renewed regulatory emphasis on performance improvement. At quality conferences like IHI, the MACRA/MIPS hype is strong. Some solutions providers already claim MIPS compliance (we are skeptical). When providers claim to tie reporting for reimbursement into performance improvement metrics, we expect EHR-integrated performance improvement to reach the Peak of Inflated Expectations.

Virtual Reality in eLearning: VR might evoke a bit of déjà vu in experienced eLearning observers, as we went very far down the simulation/avatar-y path a few years back: around 2006, Second Life was the talk of eLearning (it’s not now). Nonetheless, VR is something qualitatively different from 3D-animated, flat-screen computer games, and has been democratized by Google Cardboard and Samsung’s Gear goggles.  The presence of VR on the upward-sloping part of the curve was demonstrated at this year’s DevLearn conference, where a number of vendors were demonstrating VR-based experiences. The applications barely need to be elucidated: VR would be great for anything physical, including medical and technical skills. We’ve blogged about an application at Filament Games. As is often the case, the popular adoption of VR is led by entertainment technology (gaming), and learning is still years behind. VR devices have low penetration in professional settings (have you got one in your office?). While the hype is, naturally, strong, VR is still perceived as a “future thing” by most in eLearning.

2017 eLearning Predictions

Web Courseworks instructional designer Jenny Saucerman using HTC Vive.

Peak of Inflated Expectations

Subscription Learning: Subscription learning goes by many names, including spaced learning, but where we find it particularly hyped right now is in the “online continuous assessment” model that has been popularized by the American Board of Anesthesiology. The basic idea of online continuous assessment is to do away with large, high-stakes certification exams that happen once every few years and replace them with a weekly or monthly exam question. Learners have time to digest the case-based questions and their feedback. They might even apply what they learn from them in their practice. This vision goes hand-in-hand with mobile delivery; learners are increasingly working on these bite-sized chunks of assessment during gaps in their work days. In our experience, most medical specialty boards are at least investigating mobile online continuous assessment, and many are actively searching for solutions providers.

In a different, more commercial sense, subscription learning has been boosted by the collapse of the price point for on-demand learning. We’re seeing organizations with large libraries of online learning that are interested in selling their content on a subscription model: the “Netflix of self-paced learning” approach exemplified by OpenSesame. And finally we note that Coursera’s nanodegrees (see MOOCs, below) are now sold on a monthly subscription model.

In short, a lot of people are hanging their hopes on the simple idea of subscribing to learning. Expectations are a bit inflated.

Data Visualization in eLearning: Nate Silver’s The Signal and the Noise was published in 2012, and right about that time the market for self-service business intelligence software (Tableau, Qlik, Microsoft Power BI) went bonkers. Self-service BI software has democratized data visualization. Predictably, the trend has slowly disseminated outward from the higher-revenue parts of the enterprise to education, and we are now seeing constant talk about the need to offer administrative and learner-facing graphical dashboards to offer “insights” to designers and learners. This trend is clearest in corporate eLearning, where the big corporate LMSes have colorful pie charts and histograms gratuitously flung all over their interfaces. Even Moodle, which is less commercial, now has a PowerBI plug-in. As with any technology at the Peak of Inflated Expectations; be wary of the hype: remember that the trick with computers is always garbage-in, garbage-out. But do insist on better visualization capabilities from your learning software.

2017 eLearning Predictions

Tableau dashboard

Trough of Disillusionment

Gamification: It’s the proliferation of leaderboards that has landed Gamification (temporarily, we hope) in the trough of disillusionment. Many corporate LMSes offer leaderboards. They’re easy to turn on, and also easy to misuse. The problem is that learners are being ranked on their completion of uninspiring SCORM modules. If learners feel they’re being ranked arbitrarily, or that a training program is trying to spice up low-quality learning with “a little bit of competition,” then they’re going to find the leaderboard more annoying than motivating. That being said, there’s plenty of hope that gamification will pull through the trough and reach productivity relatively unscathed. The “serious professional trivia” model that we first encountered years ago with SCTE is achievable for most professional associations using available technology. The American College of Chest Physicians runs its CHEST Challenge competition in two phases: a first phase is administered online as a quiz in their LMS, and the second phase is a live “quiz show” format. The key to getting out of the Trough of Disillusionment is higher-quality, lower-cost content enabled by new tools, such as BranchTrack.

xAPI: We are big believers in xAPI. We want to be clear on that. But every technology goes through a phase of disillusionment; that’s the point of the hype curve. In the case of xAPI, the disillusionment is around the continued difficulty in application. xAPI has not reached the “plug and play” productivity of SCORM.

Speaking here as a member of the eLearning community, this is probably our own failure to learn and adapt as quickly as we might like (hey, we’ve got SCORM packages to finish). But it’s also pretty clear that the xAPI standard, in its flexibility and granularity, is ambitious, and it takes some time for ambitious standards like these to be absorbed into everyday practice. For now, it might seem as if xAPI is stuck in a “perpetual piloting” phase. But the pilots are proliferating – we’ve seen many more pilots in the last year than in the rest of the short history of xAPI put together. And that’s a good sign: as more practitioners participate in pilots, understanding spreads, and xAPI moves upward toward the Slope of Enlightenment.

Slope of Enlightenment

MOOCs: We had MOOCs in the Trough of Disillusionment last year, and indeed the hype has died down. Business models at the big MOOC platforms have shifted; both Coursera and Udacity are now pushing “nanodegrees” with a self-paced format. And their MOOCs are not entirely open anymore. Many now cost money. Meanwhile, corporate MOOCs seem to have, essentially, fizzled. The “enlightenment” we’re experiencing is best expressed by a recent quote from Udacity founder Sebastian Thrun, who told the New York Times, “If I look back at the MOOC hype, what actually happened was that people equated a cheaper delivery method with the replacement of the entire educational system. A cheaper technology is not the same as a business revolution.” We agree: MOOCs should be seen as a valuable part of an educational system that must serve an incredibly diverse range of learners and purposes. Educational institutions of all stripes shouldn’t see them as a replacement for anything, but a tool that meets certain needs.

Plateau of Productivity

Mobile Learning: Celebrate, for we’ve finally reached the point where we can look back at the train-wreck that was the early days of mobile learning and laugh. There was the wrenching transition away from Flash, the over-hyped hybrid apps, and the proliferation of unusable JQuery Mobile-based responsive websites (OK, some are still out there). We’re past that now, or a good many of us are, and we’re finally to the point where we can intelligently and productively use the various options available to us. That pleasing, placid feeling of understanding and usability…why, we must be on the Plateau of Productivity.

2017 eLearning Predictions

Courtesy of The American Board of Anesthesiology, MOCA 2.0.


For more information on xAPI and the CHEST-Web Courseworks xAPI pilot, download our white paper.

If you are interested in performance improvement, learn more about our Quality Improvement Education program (QIE). Web Courseworks’ received a Brandon Hall technology award for our QIE program in January.

Did you like this post? Share it!
Share on Facebook8Tweet about this on TwitterShare on LinkedIn150Google+2
  • Pingback: 2017 eLearning Predictions: Updated Hype Curve ...()

  • Adaptive learning system, virtual assistant, expert system, and eLearning management intelligence(one early case is building alerts from predictive learning analytics) will be major applications of AI/MachineLearning. Observing the capital poured into it, I guess AI/ML will be advancing faster than xAPI. The issue of xAPI isn’t about technology itself, it’s about people silos that leads to data silos.

  • Katie Delgado

    I would love to see a table that shows all the new learning technology with these columns:

    ~What it is and what can it do
    ~Major benefits to learners
    ~An implementation difficulty rating
    ~Cost of implementation
    ~Learning curve for end users
    ~Skill sets needed as a developer/administrator

    I think I’m going to make one. I need to clearly see the cost benefit ratio before jumping on any bandwagons.

    https://www.linkedin.com/in/katiedelgardner