Notes on Adaptive/Personalized Learning

Oddgeir Tveiten

January 18, 2016

Blogpost by Eilif Trondsen. Originally posted here.

We have all heard it stated many times, I am sure, that learning is a very personal and social process, where “individual context” is key to how effective someone’s learning will be—in part because that context will drive the learner’s engagement in the learning materials. We all have very different needs, backgrounds, and experiences and most of us learn in different ways and at different speeds, and these things also affect, or are affected by, how engaged we are in the learning process.

And this situation makes it very clear what challenges are faced by teachers and professors in the classroom, especially when they present the same lecture to all the students in the class, regardless of their contextual differences. It is not surprising, therefore, that many educators and edupreneurs have felt that this is an area where technology could make a big difference, by adjusting the learning process and materials presented (i.e. both content, in terms of degree of difficulty, for instance, as well as the speed in which it is delivered), to the specific learner’s context.

As a result of the market need/gap that exists, a growing number of companies have emerged in recent years to tackle this problem. In the San Francisco Bay Area, for instance, a startup school, AltSchool, was launched in 2013 to provide more customized learning for students, where teachers can deliver custom-made lesson plans that combine student interests, parental priorities and specific tasks. And the school makes the point that improvements happen in hours and days, not years. AltSchool—founded by CEO Max Ventilla (who co-founded and sold a social search company to Google for about $50 million)—recently raised $100 million to “build a network of private “micro-schools” while also developing tools for public schools.” (according to Silicon Valley Business Journal).

What AltSchool provides in form of “customized learning” is different from what Knewton and other players that offer “adaptive learning” tools and technology provide. In fact, Jess Nepom of Newton in an article on adaptive learning (see source in figure below) clarifies the terminology used around these issues by differentiating “adaptive” from “personalized” and “differentiated” learning.

Adaptive Learning

The blog post [“Adaptive Learning Market Acceleration Program”], referred to in the Reference section below by one of my favorite EdTech bloggers, Phil Hill, describes an interesting program funded by the Gates Foundation to test out adaptive learning in a number of US universities. Phil notes that “Originally planned for 10 institutions, the Gates Foundation funded 14 separate grantees at a level of ~$100,000 each. The courses must run for 3 sequential semesters with greater than 500 students total (per school), and the program will take 24 months total (starting June 2013)”. If any of you reading this post is interested in adaptive learning in Higher Education (HE), I would highly recommend monitoring the Gates program.

I have not checked out what tools and technologies the HE institutions in the Gates program are using, and some might have their own platforms and adaptive analytics engine, but I suspect that many, or perhaps most, are using tools and technologies developed by specialist firms, such as Knewton. One reason for this is that it takes very strong statistics and software–and, increasingly, AI-based– expertise to develop world-class adaptive algorithms that will be necessary to compete today and in the future in this space.

When I did a Google search on “Personalized learning startups” a few months ago, I got 174,000 hits, and while the vast bulk of these hits may not involve “true” edtech startups delivering personalized learning technology or services, a large number of companies probably think they are in this space of the learning industry, as is AltSchool. It is beyond the scope of this blog post to try to identify a large number of the leading “adaptive learning technology players” but some of the players doing things that directly or indirectly relate to adaptive, personalized or differentiated learning include these:

Knewton
Declara
ALEKS
CogBooks
Cerego
OLI [Open Learning Initiative, Carnegie Mellon University]
ASSISTments
Smart Sparrow
KnowRe
Alleyoop

Alleyoop was a college readiness network for teens that combined tailored guidance and educational content with game dynamics. It was Pearson backed, and aimed to be “Zynga for Learning” but closed down in Spring of 2013.

One of the leading players in adaptive/personalized learning is Knewton, and the company is active not only in the U.S. but also in Europe, where it has made a number of strategic deals (not only with Gyldendal (a Norwegian publisher)—described more below—but also with Sanoma (Finland), Santillana (Spain), Malmberg (Netherlands), and Elsevier. Perhaps its largest and most publicized deal, however, involves a partnership between Knewton, Pearson (the UK-based and largest learning company in the world), and Arizona State University. This three-way collaboration has generated much heated discussion in academic circles in the US, and if you want to learn more about this, two of the articles you might find interesting (both in Inside Higher Education), is this and this.

In the Nordics, the following publishing companies have made deals with Knewton to use its technology, and according to the press release when the deal with Gyldendal was announced, said that Gyldendal would use “Knewton’s infrastructure to power personalized digital products for math students that continuously adjust to each student’s unique needs [i.e. in K12].” Here are some of the aspects and implications of this deal that I find most interesting about this deal:

Gyldendal [Education] feels it has a strong technology and R&D group and thus is well positioned to work collaboratively with Knewton.

Gyldendal also has strong curriculum design expertise, and thus will be able to design curricular approaches to personalized/adaptive learning that leverages Knewton’s analytics engine.

While the current collaboration project is focused on K12, Gyldendal may be interested in leveraging its competencies built up around personalized/adaptive learning into other arenas (including corporate learning) and other services and markets.

If Gyldendal’s new products and services around personalized/adaptive learning prove successful, it not only opens new venues for the company to offer attractive products and services to customers, but may also help tie customers more closely to Gyldendal.

Gyldendal is just now in the early phases of rolling out its new sets of personalized learning products in Norway, so it will be a while before we get any details about how well these products are working. But Gyldendal, Knewton and the K12 schools involved are hopeful that they will be able to demonstrate that the new products are indeed resulting in more effective—both in terms of improving learning as well as doing so in a cost-effective manner.

But what does the Knewton deal with Gyldendal and other large publishers in Europe and collaborations with various HE institutions, especially in the US around personalized and adaptive learning, mean for Nordic edtech companies, now and in the future? Here is some of my own (“educated”) speculation, and I would be very interested in hearing your perspectives, and whether you see it the same way, or differently, both generally and for your specific country or markets you are focusing on:

Knewton’s considerable investment in adaptive learning technology will likely make it tough for Nordic startups to try to compete directly with them in the analytics engine space.

Nordic startups interested in this space may want to explore how they can develop complementary tools, technologies or services to offer Nordic publishing firms or schools/universities that decide to work with Nordic publishing houses that have collaborative deals with Knewton.

Many publishing companies may not have the technology or R&D competence that Gyldendal has, so this may open opportunities for Nordic edtech startups to provide these missing elements that publishing companies should have to fully take advantage of Knewton’s analytics engine.

The Nordic publishing companies teaming up with Knewton are still very early in their development and marketing of innovative personalized/adaptive products and services, so small, agile startups with strong technology in this space may find market opportunities and niches that these larger players won’t target, at least in the near term.

Finally, in a discussion with Karim de Coster, Senior Manager, Business Development – EMEA at Knewton, he made it clear that the company is very interested in exploring collaboration with Nordic edtech startups that have interesting technology or other capabilities that could be of interest to Knewton. We may invite Karim to the WLS2016 Conference in June, in which case we may have an opportunity to explore this issue in more depth and find out what specifically Knewton might be looking for from Nordic EdTech startups.