The de-amplification of Amplify

Nearly five years ago, Newscorp’s Rupert Murdoch bought over Wireless Generation (90% for USD 360 mn, such a hit) with the belief that

“When it comes to K through 12 education, we see a $500 billion sector in the U.S. alone that is waiting desperately to be transformed by big breakthroughs that extend the reach of great teaching, …Wireless Generation is at the forefront of individualized, technology-based learning that is poised to revolutionize public education for a new generation of students,”

They launched Amplify in 2012. As ZDNet reports:

Amplify had a digital library, various tools and a tablet-based platform. Amplify also had analytics, curricula aligned with state standards and distribution tools Amplify and AT&T collaborated on pilots for tablets.

Newscorp is selling out Amplify because of the losses the nearly USD 1 billion investment has incurred. The report concludes that Amplify failed to win sales and acceptance because:

1. The Intel tablet based solution did not integrate well with Google and Apple and this was an issue because students preferred to bring their own devices to school
2. Integration with the information systems of districts was a pain
3. Could not compete with the developer pull from Google and Apple, even though they built a marketplace
4. Amplify was a difficult choice for decision makers because it did not have scale and expected too much in terms of implementation

Unlikely that this innovation in EdTech, even under new owners, are going to be able to compete with mass user platforms provided by Google and Apple (and perhaps Microsoft, if things go well with them next year in the mobile space).

Contrast with companies like Classdojo, Remind and Edmodo.

ClassDojo and Remind are two of the biggest names in edtech today, with tens of millions of users each, and $10.1 million and $59.5 million raised, respectively.

Also to be mentioned is Google Classroom –

Since its unveiling (Ed: In 2014), 70 million assignments have been created on Classroom and Google Apps for Education has amassed more than 40 million teacher and students users.

Interesting, all 3 are focused on the teacher. All three are looking at embedding themselves in the teacher’s workflow. All 3 have communication and storage at the core. All 3 have assignments, assessments and analytics as the focal area from the teacher’s workflow (should look at WebAssign and Fishtree as well).

They are also a far cry from what Amplify was trying to address in one crucial way, with respect to the grandness of their vision. Perhaps Amplify would have done better to Simplify their approach – perhaps go the way of Smart Sparrow or the latest Knewton beta. I am not sure how platforms like Docebo or the platform side of Edmodo are doing, but it is really important to see what users want instead of seeing what platforms or technology can do. In fact, users have been telling us a lot about their preferences by the results we see from initiatives and companies such as these – there is always a chance that it could be done better, however one could also choose to not repeat certain obvious mistakes.

In a teacher led model, offline digital interventions have been proven somewhat socially acceptable, at scale. In this approach, the teacher is really the only hope for any meaningful and scalable approach to elearning. This is an inevitable fallout of an educational model that has tightly controlled structures and rules, with the teacher being the lead implementer of those rules. Even with the xMOOCs, this model looks as it will continue to flourish.

Another inevitable fallout of the educational model is the propensity to succumb to it. By focus on existing workflows (including procedurally flipped ones), one succumbs to various imperatives – how to decrease load, how to make teachers’ lives easy, how to provide helpful analytics quickly, how to provision a bank of standardized materials or baked content, and so on. So we are then engaged with infusing technology into the status quo, rather than engaging with an environment that has changed remarkably in the past few years.

That environment is not just increased access to open educational resources, better Internet, more Apps and mobility, greater awareness and use of communication tools like Whatsapp, it is also the surge of social networks for learning, cMOOCs, gamification, learning analytics (don’t miss Caliper from IMS Global which seeks to plug the gaps of TinCan) – in general, the possibility that learning is really the process of making connections and knowledge is the network (Connectivist learning).

But more so, it is an opportunity to reinvent the wheel. The more I think about that phrase, the more attractive it sounds. If you think of the wheel as the cycle of learning, it certainly could benefit from re-invention. Of course, if you reproduce instead of reinvent, that is a failed mission from the start. If we try to reproduce a system of learning by using new technological affordances, it is likely a failed mission from the start.

Read Full Post »

#rhizo15 Learning Subjectives is #insidethebox

Another edition of the fabulous “mostly run by” Dave Cormier, Rhizomatic Learning conversation Rhizo15 begins! The question of the week, with the usual deep subversive intent is:

Build learning subjectives: How do we design our own or others learning when we don’t know where we are going? How does that free us up? What can we get done with subjectives that can’t be done with objectives?

Are we thinking inside the box? Does changing around “objectives” to “subjectives” free us from tradition – the tradition that says that learning must be designed?

Simon Worren (Worried Teacher) points out a way of looking at it – emergent outcomes.

Intended learning outcomes (ILOs) are essential in module planning to indicate the direction of teaching, however, it must be recognised that ILOs represent the lecturer’s intentions for study, not the student’s.

Carl Gombrich makes the point that design based learning (in the instructional design sense) is often at odds with traditional university based education that is more “emergent” in nature – the former more equipped to deal with “skills” and the latter with “concepts”. While most of the university education we have seen (atleast in India) is hardly emergent, it is also true that much of design based learning is hardly “aesthetic” either, at least at scale. He makes the case for a merger – “design for concepts” rather than “design for skills”. By that he suggests that either we move to a higher level of abstraction in design (say, through more loosely defined learning outcomes) or that we recognize that certain areas of study are more suited to one versus the other approach.

Sarah Honeychurch challenges the notion that rhizomatic learning is at once personal and collaborative in nature. Perhaps that learning may be greatly enhanced if it was collaborative, and perhaps we are all missing out learning from her attempts. I think the point to be made is that you may not necessarily want your learning to be public or brought about by shared experiences, but the more we learn and share collaboratively, the more we help learning as a whole.

Simon makes the point that people and ideas (and beliefs) cannot really be separated and that the learner has her own agency in deciding how, when, why, whether and where to interact with others. Collaboration cannot really be mandated to be an essential condition for learning. Also he makes the point that embracing messiness and uncertainty in learning does not necessarily mean that education systems as such should embrace messiness and uncertainty or that knowledge is only  fuzzy and uncertain.

My sense is that we are not talking about the same thing here. I think the focus is not on defining a single way in which we learn. The focus is on one possible way to learn – a way that is intensely collaborative, yet personal – which some people may find to be extremely fulfilling, so much so that they would exercise their agency and choose it to be the way they would like to learn in the world. Many people would not find this way “super-fun” and they may simply not be comfortable negotiating the messiness, but that does not mean that way of learning is unreal or useless.

Rebecca asks the million dollar question – what feeds/constrains online collaborations? This is something that we need much more work on. There is much to learn from the cMOOCs since 2008 and many other experiments across the world. Perhaps we are hitting the problem with the same old approach – trying to “design it” – trying to change the way we learn and teach by employing new ideas.

My own belief is that when we engage with new forms of online, social collaboration, the only real outcome we should be concerned with how well learners and teachers are able to negotiate this medium with each instance of such emergent learning. It is a longer term process of realigning to or establishing a new way of learning, more than a way to establish a better design paradigm that generates better traditional outputs such as grades. Maha Bali makes the case for making the subjective obvious (critical pedagogy) and makes the case that the goal is “…not to filter better performance from worse; it’s to help students learn.”

Thinking of objectives and subjectives seems very much #insidethebox (in all its variants – as the starting point for learning subjectives on a continuum, as elements that can be designed, or as inversions). We need to perhaps focus ourselves more on asking what if there was no design, no objectives and subjectives that we could identify – what then would learning really look like?

Read Full Post »

Is Free the Future of Digital Content in Education

Mary Cullinane, Houghton Mifflin Harcourt’s first Chief Content Officer, writes on Why Free Is Not the Future of Digital Content in Education. She makes the case that technology advances will not make digital content eventually free, just like in the music industry. This is because technology is adding value, not just reducing the cost of delivery or production. She writes:

As students engage with the content, the content learns more about the students and it also becomes “smarter”. A digital engine compares students’ responses to those of all other users. Equipped with that data, this adaptive learning system doesn’t just show that a student answered incorrectly. It knows why she did, and uses those insights to create a customized learning path.

In doing so, technology helps solves a big problem that has always confronted teachers: students learn at different paces. Advanced students can get bored and struggling students can give up. Now, as a teacher, I can put content in front of each learner that is personalized to his or her needs. It’s something teachers have been doing through the ages, but technology brings it to the next level of adaptivity.

However, this is not an argument that precludes “free”. Technology has added tremendous value for many services (take, for example, instant messaging) and kept the price at zero. Adaptive learning is just simply not the domain of publishers who have large repositories, author networks and organized funding. It can be done in a free, open manner too (just take a look at Khan’s work, for example). What is true, though, is that content development for adaptive learning can become expensive very quickly. We have come a long way since Ms. Lindquist : The Tutor. Now paths through content leading to mastery can be uncovered through collective intelligence, rather than having to be enumerated as before. However, adaptive content still has design requirements that are in addition to regular content development and learning design.

The other thing I would watch against is taking this as a magic wand that “solves a big problem” for teachers and personalizes learning for students. Personalized learning is a very difficult thing to crack, and it is not the same as recommendation systems such as the ones we see today that crowdsource learning patterns. Ultimately, these systems seek to harmonize existing goals (like solving an algebraic equation or learning a grammar construct) where the ontology is precise and the domain exhibits structured rules. Much of learning, however, is not that. Nor is it always goal-seeking.

Read Full Post »