I have been pretty relentless in mocking Knewton CEO Jose Ferriera for his “robot tutor in the sky” description of the company’s adaptive learning product. And I pledge to continue this proud tradition. It’s the gift that keeps on giving.
But the truth is that most of the vendors in this space, both startups and traditional textbook vendors, are not doing a whole lot better than Knewton. The reason the “robot tutors” quote remains so useful is that it is a beautifully compact and self-explanatory illustration of the same symptom that shows up throughout the industry. Vendors struggle find ways to pitch their products that educators won’t find both offensive and ridiculous. In today’s episode of the evergreen marketing sitcom, we’ll see examples of Pearson making analogies or prominent references to Netflix, Google Ads, and one other surprise tech hype guest. (You don’t expect me to give you all the spoilers up front, do you?)
For the most part, it’s not that these companies are conspiring against educators (although we do get periodic outbreaks of “disrupting education” talk driven by Silicon Valley bros who are want to make a virtue out of the fact that they don’t know how to sell to schools). Nor is it that the people in these companies are stupid, or that they don’t understand their own products. Rather, it’s a manifestation of a larger problem of our culture. Most people who are not teachers, whether they work for education vendors or not, don’t understand what teaching is. They have a very rough Gradgrindian sense of it. Learning, in this view, is not an activity undertaken by learners but something that is done to them. You screw off the tops of the students heads, pour the knowledge in, and voila! They have been learned! Learning can be done to students equally well by a person or a machine, since it is not “personal” in any way that relates to normal understanding of the word, and it certainly is not interpersonal.
The people inside education companies that work with customers closely in designing the products often (though not always) have a more sophisticated understanding of teaching than this. But the further away the employees are from product development, the less likely they are to understand how teaching really works. The inner workings of the classroom are just not part of their daily lives. As I wrote in my review of InstructureCon, it’s very hard and rare for any company of over, say, 100 employees to be strongly aligned and consistent across its culture. This is doubly (or even trebly) true when the alignment you’re trying to get cuts against the grain of strongly held cultural notions like the ones we have of teaching. So there are a lot of employees inside most of these companies that don’t get it. Some of them are decision-makers. Some of them are allowed to speak in public. Some of them are both. The result is generally…not good.
It’s probably worth providing a more detailed version of the full disclosure thing than usual for this particular post. Pearson is a client of MindWires, the consulting company that Phil and I run together. When we do consult for vendors (as opposed to universities, which is the other half of our main business), our focus is on helping them understand the needs of students, educators, and schools. We don’t take money to advise them on marketing except in the narrow sense that we try to help them understand the kinds of educational problems that their customers want to hear (and talk) about how to solve. One of the reasons we try to be meticulous about drawing this line is that it leaves our hands clean to write the kind of public critique that I am about to offer here. We always tell this to our corporate clients and let them know that we could write a critical piece about them at any time without warning or editorial review, as long as we are not writing about proprietary information that is part of our consulting work with them.
So here goes….
A good education is not like a good mix tape
The first examples of the problem come from a pair of interviews of Pearson CIO Albert Hitchcock in a UK tech publication called V3. The first is entitled “Pearson aims to become the ‘Netflix of education’“.
A vendor is already in deep trouble when that’s the headline. The piece contains gems like
“The analogy I use is to be the Netflix for education. We want to create a single platform to deliver all educational content and services, irrespective of the age and stage of the pupil,” said Hitchcock.
“Much as you would consume movies through Netflix, or buy services through Amazon, we want education to be delivered through this single, quality user experience, but available to all ages and stages of learners.”
Given his profile on LinkedIn, Mr. Hitchcock appears to be new to education except for whatever memories he has of his own days as a student. Let me offer a couple of suggestions on how to get along in education for him and the many vendor employees who are in a similar situation:
- Never, ever, say that you want your company to be the Uber of education, the Airbnb of education, the Pokemon GO of education, or the [insert name of tech darling] of education unless you really enjoy being a recluse (or you are secretly a double agent for your employer’s direct competitor).
- If you absolutely must say something like the above, then do not say you are the Netflix of education. Honestly, Netflix isn’t even great at being the Netflix of movies. The last time they recommended a movie that I actually wanted to watch was…uh…never.
There is a recurring cultural fantasy that “solving” the education “problem” consists of creating a customized playlist of little content bits. So really, more like the Spotify of education, if you want to play that game. This idea enrages educators because it trivializes what they do. Nobody who has taught believes that proper sequencing of content chunks is the hard part. (For a fully fleshed out prior example—or a “worked example” in teaching parlance—of how the sorts of comments Mr. Hitchcock have made typically play out in educational corporate branding over time, see my post-mortem on Udacity’s pivot away from higher education.)
The second noteworthy quote from Mr. Hitchcock came in the article, “Struggling to find a data scientist? Use machine learning instead“:
“We’re starting to meta-tag all the content so it can be chopped up into component parts and reassembled on the fly [for each user],” explained Hitchcock.
He likened it to the techniques used by firms such as Google in selecting which ad to display online based on its understanding of each user.
“But our use case is much richer. We’re changing the nature of the content, so we need deeper analytics, and we use Pearson’s unique intellectual property to modify that learning content on the fly,” he said.
We’ve gone from Netflix to Google Ads. Out of the frying pan, into the fire. Can we get a spamming algorithm analogy just for good measure?
PR disaster aside, the day that Pearson fires all its data scientists is the day they should go back to selling paper trade books. First, educational contexts for machine learning are not just “much richer” than serving ads. They are orders of magnitude more complex. If Mr. Hitchcock wants to get a sense of what that means, he need look no further than Pearson’s own breadth and depth of published research. Like this paper on assessing case analyses in bioengineering ethics, to pick the one that happened to be first on a list of publications on Pearson’s research web site.
Second, many of the most interesting and valuable patterns that need to be spotted in any class that has a human teacher are the ones that happen at least partly outside of the digital content. No amount of rich metatagging will solve that problem. You will always need human scientists to make sense of what’s happening in the digital part of the education by looking at the analog part. Most publishers (though not Pearson, ironically enough) have been very slow to understand that this kind of anthropological data science work is a core competency for their business now. Mr. Hitchcock’s comments strongly suggest a basic lack of understanding of teaching, which is deeply worrying given his senior role in the company and apparent ability to get press access.
Serving our robot overlords
The next gem comes in the form of a sponsored article at Education Dive written by Christa Ehmann, Senior Vice President and Chief Education Officer at Smarthinking, Inc. – a Pearson Company (quoted from her LinkedIn profile). It’s not Gradgrindian in the classic sense but there is definitely a grinder in here somewhere. The article is called “A vision for PERSONalized learning.” OK, so we know from the start that the piece isn’t going to be subtle, but at least the emphasis is promising. There’s a lot of the standard quoting of definitions and citing of trends in the beginning that you can frankly skim or skip. The nub of the piece starts here, in the seventh paragraph:
The discussions are vibrant, and the technical developments are rapid. Unfortunately, the heart of any successful personalized learning system is often neglected: the human teacher, tutor, coach, or mentor. [Emphasis in original.] While these indispensable people are indeed recognized as working in conjunction with and alongside personalized learning elements, they are typically positioned as distinctly separate.
Personalized learning vision, research, and development will be far more transformative if it focuses on opportunities to strategically interject human-delivered instruction of all types throughout its technology and content. By positioning human teachers at the center of these systems, we can create truly transformative and seamless learning environments that are both scalable and economically viable.
Now we’re getting somewhere. There’s some language in there that troubles me. But I’ll give you a pass for now since at least we’re talking about humans in general and teachers in particular. Tell me more about this vision…
To understand what this might mean in practice, imagine a scenario like this:
- A content author melds text with media drawn from news reports, factually represented fictional scenes, lab settings, and verified social media content.
- Using Virtual or Augmented Reality devices such as Microsoft’s Hololens®, the learner is then placed within the content as a participant.
- A human teacher or tutor is then seamlessly interjected into the learner’s experience, walking through key processes (e.g., demonstrating a chemical engineering process or helping a nursing student calculate dosages for a life-or-death injection).
- While text and other content remain accessible in the background…
Stop. Just stop. Don’t go any further. Please.
So step one for bringing “the human teacher, tutor, coach, or mentor” into “the heart” of “PERSONalized learning” starts with a content author. Who, in fairness, is a person, at least until Mr. Hitchock figures out how to replace content authors with algorithms. Step 2 is Microsoft’s Hololens®.
Let me say that again. Step 2 is Microsoft’s Hololens®.
Finally, in Step 3, a human teacher or tutor is “seamlessly interjected into the learner’s experience.”
The use of passive voice alone should have been enough to set off alarm bells. Microsoft’s Hololens® coming before the teacher who is the PERSON headlining the piece should have been a five-alarm fire for whichever PR PERSON approved (and paid for!) this article. Most disturbing of all, though, is that unlike Mr. Hitchcock, Ms. Ehmann has the title “Chief Education Officer” and has a PhD in Education from Oxford. She should know better.
Update: In fairness to Ms. Ehmann, please see my comment in the Google+ thread below, which provides a more charitable and probably fairer way to frame the problems with her article.
I know from my public work as an analyst, our consulting work representing schools when we have dealt with Pearson on their behalf, and our consulting work directly with them that these examples do not represent the monolithic views of the company. Pearson has historically been the antithesis of Instructure in terms of internal alignment. Asking what Pearson thinks about education has only been slightly less complex and semantically indeterminate than asking what America thinks about the 2016 Presidential election.
Update: Case in point: Here’s an article in EdSurge by a Pearson employee entitled “Why We Don’t Need a ‘Netflix for Education‘.”
Despite recent efforts to reduce their alignment problem (including their massive effort around “efficacy”), these articles suggest the problem is stubbornly persistent. The fact that two senior executives were allowed to make these kinds of egregious statements in public does not inspire confidence that folks inside the company who really do understand teaching and learning are the ones who are driving the bus (assuming the bus has a driver).
But the larger point is that Pearson is no more unique than Knewton in being unable to stop itself from putting its foot in its mouth on topics related to digital education. The shift to digital requires these companies to understand much more about teaching than they did when they were just shipping books off and letting teachers do what they like with them (which sometimes was almost nothing). And by “companies,” I mean all the employees in the company, and certainly all employees who are allowed to make strategic decisions or speak in public. Because these examples of Pearson’s public statements are really, really bad. If you look around at the other textbook and adaptive learning vendors, both large and small, most of them are not doing much better, either.
Going back to the point I made early on about where we draw the line between our consultant and analyst roles, Phil sometimes jokingly calls posts like this one “free consulting.” In that spirit, we offer these gifts to our colleagues in Pearson and the rest of the ed tech industry:
We made these for you. Use them directly or just use them as models. Use them in your external communications and, more importantly, in your internal communications.
Or you can just keep letting Netflix, Google Ads, and Microsoft’s Hololens® be your dominant themes when you talk about the future of education. That could work."PEARSONalized Learning",