Personalization: You Can't Afford It

We have seen the future, and we can't afford to live in it.

Altschool has just let out word that the tech-powered boutique of personalized education will become one more purveyor of off-the-rack computer-centered education-flavored product. There are many lessons underlined here-- I want to focus on the reminder of why, exactly, we can't have nice things.

Altschool's original vision was ambitious. Hire really good teachers. Keep class sizes small. Back up that teacher with a high-powered array of tech resources, allowing the teacher to perfectly track each student's progress in nearly-real time, then give that teacher unparalleled power to select a perfectly personalized set of materials for every single student. Keep a full IT department right on the site.

What do we dream of when we dream of True Personalized Education? Teacher-directed, with support from a powerful array of resources and facilities.

The problem is, this would be really, really expensive. Really expensive. You have to pay top dollar to lure those super-star teachers, then design your perfect educational ecosystem, then get top-of-the-line tech and hire IT people to keep it running, then buy up the resources needed to meet every possible individual student need or interest that might arrive. Ultimately you have several staff people hired for every single child. Expensive. Altschool was dropping something like $40 million a year.

You can't afford it. Hell, even the rich folks in Silicon Valley couldn't afford it.

So what happens? And how does the Personalized Education dream turn into the "personalized" education nightmare?

There are only a couple of ways to deal with the huge expense of a personalized boutique school.

One is to cut corners. 

To be prepared for any individual interest or need, really prepared, you'd need a library of tens of thousands of units, covering tens of thousands of content areas at dozens of different ability levels cross-filed by particular skill or knowledge sets involved. The library would be huge, and would need to be reviewed and updated every year. That would be expensive, and the software needed to search it for the material with just the right qualities for Pat or Chris would have to be pretty heavy duty as well.

So let's, you know, cut that library down to a couple hundred items. Let's just focus on the most common stuff, and if we find some students who aren't a perfect fit, well, if we've got materials that are Close Enough, that should do. And we can reduce some of this coursework to simple sequencing. Take the pre-test, and if you miss numbers 1 and 2, you get Drill Sheet A, and if you miss numbers 3 and 4, you get Drill Sheet B. Simple, easy to manage, fewer materials to store. Cheaper.

And getting the very best teachers to run the classroom-- well, that would be pricey, too. Let's just round up some teachers who are Good Enough. In fact, since really good teachers might start to question all the corners we've cut, let's just grab some warm bodies, train them in how to operate our system, and let it go at that. If we let the classroom be driven by the software system and not the teacher, then it's easier and cheaper to just fill in the meat widget job with a handy warm body.

But if I started this "personalized" program because I thought I could really make school awesome, why would I cut so many corners that I hurt the quality of the school.

Because I need investors.

The other way to take care of the enormous amount of money I need is to get somebody to give me that money. And investors look at my classroom a little differently.

First of all, the corner cutting appeals to them hugely. To them, every dollar I spend on that classroom is one of their dollars. Do we really need three tech guys? Couldn't one handle everything by himself? Couldn't we scale back on the library of units that we're buying every quarter?

And having a highly-qualified and experienced super-teacher in each classroom-- that's great and all, but we can't really monetize that, can we? We can't sell it as a special secret. That proprietary software, on the other hand-- we could sell that to other schools and sell them the computers to run it on. And if we could streamline that whole software program and lesson library a little more, it would be easy to package as one-size-fits-all "personalization" for any classroom in the country. Because the more All our One Size fits, the bigger the potential market for this.

By all means, keep the Original Boutique School going-- when we bring people to see this or we show them videos or we send the master teachers out to talk about it, people will pee themselves with joy and fight to buy our off-the-rack version. We will make a mint.

But investors are not showing up to pump money into a Personalized School just so every schlubb's kid can actually attend there.

And asking those investors to work around a mountain of delicious, valuable student data and leave it alone is like asking someone to come to work every day and work at a desk that sits on a mountain of $100 bills without ever touching one. Theoretically possible, but sooner or later some investor is going to say, "You know, as long as the software is already working with all this student data anyway..." In fact, that's why some of the investors are going to show up in the first place.

This is how it works

This is how "personalized learning" ends up meaning two things-- actual personalized learning in which teachers lead a classroom armed with mighty tools and resources, and faux personalized learning where the classroom is software-directed, education is algorithmically-centered, and data is mined daily and promiscuously.

We cannot afford real Personalized Learning. Okay, if we can afford trillion dollar wars without end, we could afford real Personalized Learning. But as a country, we want education cheap (particular education for children who are not our own). So real Personalized Learning remains one of those things we know how to do, but we won't do it because we don't want to. So we'll cut corners and hustle for some ROI and just generally try to look like we're doing Personalized Learning when we're really doing something else entirely.