Friday, February 3, 2017

Exhuming inBloom

The folks at Data & Society have decided to take a look at "The Legacy of inBloom,."  and mostly what they've revealed is that they haven't learned much of anything from the Data Overlords Disaster that was Gates' failed attempt to create a massive data mine.

Data & Society is "a research institute focused on the social and cultural issues arising from data-centric technological development." In other words, they're a thinky tank focused on data stuff. They're relatively new to the thinky tank biz, founded just two years ago by Danah Boyd, a scholar and Microsoft researcher who specializes in young folks and their social media stuff. D&S got a huge start with two MacArthur grants -- $400K from the Intelligence and Autonomy initiative, and a whopping $875K from the Enabling Connected Learning program. Big bucks have also been kicked in by a host of heavy hitters including  Microsoft Research, Ford Foundation, National Science Foundation, and, yes, the Bill and Melinda Gates Foundation.

This working paper was put together by Monica Bulger, Patrick McCormick, and Mikaela Pitcan. Bulger is an ed researcher, McCormick is a professional consultant, and Pitcan is a social scientist and mental health clinician. And whatever else we can say about the report, I will give them points for writing in mostly plain English rather than corporate edu-research gobbledeegook.

So let's go and make our visit.

Prequel to inBloom

The paper marks the beginning of inBloom in 2011 when Vicki Phillips, working at the time for the Gates Foundation, published a post entitled "Shared tools for teachers? There's an app for that!" This included an announcement about the nine-state consortium Shared Learning Collaborations, which planned to set up a platform that would gather, share, store, crunch and standardize all the education data (including setting standards for all ed software). Phillips was positively giddy.

But as the paper wryly notes, "The enthusiasm of Phillips’ blog post did not end up being an accurate predictor of the project’s future." From Day One, the folks tied to Bill Gates' big vision failed to read the room.

In November of 2011, Gates went on CNN to announce a "crisis in education" (because nobody had ever played that song before) as well as his pitch for one-to-one computing, because children are automatically fascinated by anything on a computer.

inBloom was set to launch in spring of 2013. The paper tried to connect that time with some sort of swelling storm of data issues-- Edward Snowden, the Target hacking. Sure. Perhaps we might also consider that 2013 marked the point at which Common Core and the Big Standardized Testing regimen that came stapled to it first really smacked America in the face-- and all of it backed by Bill Gates.

Maybe-- just maybe-- Americans were jumpy about a self-appointed private citizen trying to use his personal fortune to inflict his personal vision of education on the country. inBloom was the technocratic dream, the backbone and heart of a system in which we would collect everything there was to know about every student, then we could take total control of their education and shape their entire lives to be what we wanted them to be. Even Common Core, which seemed like the flagship of the reformy agenda at the time, could be arguably seen as a system of data tags meant primarily to facilitate collection and sorting of the data, so that we could standardized every student (well, every student without rich parents and access to a private school). Plus the whole thing would be hella profitable and grab big chunks of the billions of dollars of education money. Maybe-- just maybe-- Americans who saw all of this coming found it creepy and bad and just fundamentally counter to the whole spirit of American public education.

Spoiler alert: this is not a possibility that this paper is really going to explore.

inBloom's Story

The decision was that inBloom would be centralized, open source and nationally scaled. You might think that at the moment that someone said, "Let's create a computer program that monitors and collects data from all children all across the country and stores the data in one place," that somebody might have invoked 1984 and said, "Well, if that's not an actual bad idea, at the very least it''s likely to be hugely unpopular." But that smart person was not in the room.

Nor was there a smart person in the room when the run-up was put together, composed of pieces like a McKinsey report chicken littling about a plummeting education sky, or attempts to create panic by citing international tests cores, or whingig about NAEP scores, or concocting fantastic claims like a loss of trillions of dollars to the US economy, or declaring that Khan Academy's library of edu-videos was achieving great things, or Rocketship learning, and just generally as the paper puts it, many edutech evangelists felt "like their time had come."

Teachers may remember this period of the early 2010's as that time that a bunch of eduamateurs started running around telling teachers to shut up and stop sucking while the amateurs pushed a bunch of dumb ideas and nobody would listen to education professionals about any of it.  Oh, them was the days.

The paper also backtracks to talk about some earlier attempts at the same thing, and how they failed, and how nobody learned anything from any of them, because they just wanted to get funding and disrupt the hell out of something.

But finally, Race to the Top and the Consortium of States (which really sounds like a Game of Thrones thing) created "an uneasy alliance." The alliance was uneasy because even though RttT told states, "Go set up something just like what the CoS is doing,"

the nature of complex government procurement entangling federal, state, and local agencies, competing priorities, and layers of public sector oversight and accountability often chafed against inBloom’s preferred methodologies of software development.

In other words, as would repeatedly happen in reformsterland, the technocrats and software guys want to just do what they want to do, without having to answer to governments and regulations and customers (stupid, stupid end users).

The idea (the report doesn't really say whose idea it was) was that inBloom should have a fast, big, shock and awe unveiling, unburdened by all that planning and testing and slowly rolling it out as you're sure it's done right. You may recall this similar attitude of "We need this RIGHT NOW more than we need it right" was also a hallmark of the Common Core rollout.

This led to any number of problems, from how the program actually worked (or didn't) to the floating about of wildly different cost estimates to schools.

Changing Perspectives of Data 

Sigh. Just in case it isn't clear, we will be blaming the death of inBloom on everything except fatal flaws in the concept itself.

So now we're going to hear about circumstances in society at large which contributed to an atmosphere in which people had a hard time seeing how a giant data-sucking monstrosity was actually a good thing.

That crazy wikileaks, followed up by that wacky Edward Snowden gave everyone a "crash course" in both how much data the government was collecting and how bad they were at protecting it. The paper quotes Aimee Guidera of Data Quality Campaign, one of the advocacy groups pushing for Big Data to get full access to the public trough:

the issue is inBloom and education data collection became nuclear and it became synonymous with Big Brother, Edward Snowden, Target, making teachers into robots, putting teachers out of business, social engineering, lack of parent control. They were all fearful words, and once that genie was out of the bottle, you couldn’t bring the conversation back.

This is typical of the argument being pushed here, that inBloom folks just lost a war of PR in which words are just strategic weapons that have no meaning or connection to reality. In other words, their argument is that victims of the Johnstown flood would have been fine if they had not been exposed to press reports that said they were under water and drowning. inBloom became synonymous with Big Brother in the same way that being shot "becomes" synonymous with having a hole in your body. There was no "becoming." The advocates of inBloom wound up wearing that shoe because it was hteir shoe and it fit perfectly.

Meanwhile, all sorts of companies were regularly hacked and lost information. People somehow got the idea that corporations wanted to privatize many things while saying "screw the public." The Great Recession left folks with the idea that Wall Street was a bunch of out-of-touch money-sucking scumbags and that our elected officials were more concerned with taking care of Wall Street than Main Street. And the paper quotes a superintendent who was shocked-- shocked-- to discover that many people did not adore Bill Gates. I mean, here was this mega-rich guy volunteering without even being asked or elected or anything, coming to just remake a whole sector of society to suit himself, and lots of people weren't even grateful! Imagine!

The paper notes that inBloom's folks didn't always help. It cites Gates SXSW speech about inBloom in which he invoked Luke Skywalker and Yoda.

And simultaneous to all of this was the news that many citizens thought that the whole education reform scam, to which inBloom was solidly attached, was actually a big pile of baloney.

The Opposition Emerges

Into this contentious mess stepped folks like Leonie Haimson, who thought that letting a private corporation act like Big Brother was the worst of several worlds.

Once again, the paper does not address the validity of Haimson's criticisms, though it does report them. But it treats them as a political strategy, rather than a substantive critique of things that were true about inBloom. Speaking of Haimson and Diane Ravitch:

the arguments they picked, the targets they chose, and the support that they mobilized in response to inBloom honed in on the changing public data atmosphere. 

Right. Their arguments were arguments that they picked, not a reaction to Bad Things That Were True about inBloom. Those arguments only found support because of the public data atmosphere, not because they were Bad Things That Were True. They report the Haimson sent to a town hall about the issue:

Parents, do you know your child’s confidential, personal school records are going to be shared with a corporation called inBloom Inc?

This highly sensitive information will be stored on a data cloud and disclosed to for-profit corporations to help them develop and market their “learning products”

The data will include your child’s names, address, photo, email, test scores, grades, economic and racial status, and detailed disciplinary, health and special education records. 

The report presents this as if it were clever marketing that capitalized on public fears; it slides right by the point that everything Haimson said was true. The report also notes that inBloom reps were "absent" from the town hall meeting. That sounds sort of bland and neutral, but Haimson (I reached out to her for this piece) characterizes their non-attendance as "refused to attend," which was their pattern for many other gatherings about inBloom as well-- including "gatherings" to testify at NYC and NY hearings about the program.

The report suggests that inBloom, the NY State Ed Department, and other fans were slow or incomplete or 'reluctant" to communicate the awesomeness of inBloom to the public, but in fact Haimson reports e-mails obtained through the Freedom of Information Act show that the Gates Foundation forbid the NYSED to communicate anything to anybody without having the Gates folks vet it first.

In short, the report suggests that inBloom's advocates were a little slow or a little incomplete in their reading of and reaction to the growing opposition, as if they were well-meaning but too good-hearted to sense how much trouble they were in, when in fact, it seems more accurate to say that inBloom's supporters were aggressively opaque and completely in tune with the tone of those reformy years, which was "Shut up and do as your told, right now, because this will be super awesome and you guys aren't allowed to ruin it for us!"

The report does include one classy flourish-- comparing inBloom opponents to anti-vaxxers. Of course, vaccinations are a scientifically-established solution to a proven problem, while inBloom was an unproven, untested solution to an imaginary problem. No, if there were any anti-vaxxers in the room, it was the inBloom supporters.

But this report is stuck in the Battle of PR story. Take this--

Ultimately, there were never adequate assurances as to the safety of these data practices. In addition, inBloom representatives had difficulty in articulating what the concrete value would be for individual
students. With public doubts about safety, and no clear success case to show...

Quick-- What do you think comes next? Could it be "inBloom had to realize their plan was seriously flawed"? Or "inBloom needed to consider the truth and validity of the criticisms leveled against them"?


With public doubts about safety, and no clear success case to show, proponents of inBloom lost control of the narrative.

So even as we gather to consider what wrong with inBloom, we start with the premise that, really, there was nothing actually wrong with inBloom. Nope-- the whole business was just a PR battle in which inBloom supporters were outmanuevered, caught unawares because they were just too goodhearted and naively certain of their wonderful plan.

The Fall of inBloom

Nine member of the consortium were joined in 2013 by the Council of Chief State School Officers themselves, cementing inBloom as a BFF of the Common Core Complex. But those nine states included Colorado, where inBloom failed an audit, and Louisiana, where big questions came up about individualized student data, with social security numbers, stored on servers and not terribly secure. Georgia declared it would keep its own data, thank you. The coalition was shrinking down to New York, and continuing to shoot itself in the foot with pricey videos about how awesome it would be when students stared at screens and teachers stared at screens and everything there was to know about students could be shown on a data dashboard.

It took just seven months in 2013 for the consortium to completely blow apart.

But again, the report finds that none of these issues were part of inBloom. Instead, they share the insights of a senior inBloom tech who says that inBloom wanted to be a quick, agile techy start-up and it made the mistake of partnering with states, who are lumbering dinosaurs.

So New York became the last stand, in part because it had made the strongest bet on inBloom's future. But a large network of parents was ready, aided by, well, pretty much everyone who wasn't an employee or bought-in bureaucrat. The report suspects that our old enemy-- Implementation!-- was the problem. Quoting Pleasantvill superintendent Mary Fox Alter:

The Common Core in New York involved a teacher evaluation system, Common Core curriculum materials that were being built as the plane flew in the sky and a data dashboard system that was supposed to track all of this, and new tests that were data dark to us. So the implementation plan for all of this in New York contained those four elements. The data dashboard was one of the four, the other three were fatally flawed.

See? It was those other poorly implemented ideas that dragged poor, innocent inBloom to its death.


So what killed inBloom?

One theory. Gates spent lots of money but failed to build any parent or community trust. That's not wrong, though it skips the question of whether such trust could have been earned for inBloom. Does a hungry wolf trying to sneak up on sheep have a perception, communication and trust problem, or does it have a sheep are too smart to let themselves get eaten problem? But we love that communication theory:

Sharren Bates reflects that “inBloom did not have a privacy problem, inBloom did not have a parent problem. InBloom had an advocacy and perception problem.”

Another version of the communication excuse theory was that the whole thing was too rooted in dataland, championed by data people who forget to talk with the civilians. They were just such visionaries that they couldn't or wouldn't explain their advanced vision in terms that ordinary mortals could understand. Sure. That's it. We're just not bright enough here on this planet.

Could Things Have Turned out Differently? AKA, What Have We Learned?

There was a lack of consistent leadership. A visionary champion for the program might have helped.

But the "key takeaway" is that more community buy-in was needed. Which is a nicer way of saying "the public just didn't understand us well enough, and if we could get them to be less dumb, we'd be set." They kept asking what the hell inBloom actually did, and we didn't really answer them. And we were snotty. So maybe if we could have remembered to talk less, smile more.

In other words, the answer to "what have we learned," is "Nothing. Nothing at all."

The paper still wants to find the right way to impose inBloom on the public. What PR campaign can we use to get them on board? How can we talk them into seeing that we're right?

No listening. No acknowledging that misgivings about the program could be valid. We've done everything right-- we just have to find a way to make people see that. We're going to impose this program, but we have to impose it in a way that it sticks.

The problem of course is the imposing. The not listening. The not taking into account that parents didn't want it, teachers didn't want it, administrators didn't want it. Like a creepy guy who won't believe that "no" means "no," these folks remain convinced that they just need to try harder, be pushier in the right way.

And believe me-- this is not going away. last weekend my wife attended a conference where a session about community building turned out to be about finding ways to get community buy in so that parents would inflict reading software on their 0-4 year olds.

But if you're not convinced that these guys have learned nothing, check out their final chapter:

Icarus also Flew: the Legacy of inBloom

They didn't really fail:

The story of inBloom is not one of straightforward failure, but rather of shooting for the sun and being scorched during the journey. 

They were the victims of outside forces. Nobody could have foreseen that inBloom would flop:

It is unlikely that any team could have anticipated the combined historic and social factors around data use that surfaced during inBloom’s development and launch.

They dreamed too big, carried too much greatness: 

Undoubtedly, these were fatal missteps for the inBloom initiative. Yet, as David Graber argued in 2012, society needs visionaries to progress and grow. We need the risk takers who attempt to build flying cars or cure polio or land on the moon.

Yes, trying to collect all the student data and use it to enrich a private corporation in a Big Brotherly enterprise is totally like curing polio. Fun fact: the polio vaccine was developed with public resources and Jonas Salk never patented it, and so never got rich from it, letting the vaccine belong to the public. Is that the arrangement inBloom wanted to propose?

There's more, but you've really got the drift of their gist-- it turns out that inBloom was just too perfect and beautiful for the world it came into, misunderstood and unloved, ahead of its time.

Of course, there's another possibility-- folks understood inBloom and Bill Gates just fine, understood the overreaching grab for students' data, understood the danger of such data in one set of private corporate hands, understood that nobody could keep that kind of data safe, understood that there were no benefits in any of the data-driven school models except for the companies that would run them, understood that inBloom was a massive, sneaky, dishonest, expensive grab for power, one more limb of the grasping, privatizing octopus that has been ed reform.

People saw it, they understood it, and they rejected it. If after all this time Gates and his crew don't understand why, then they are among the dumbest smart people in the world.


  1. "Teachers may remember this period of the early 2010's as that time that a bunch of eduamateurs started running around telling teachers to shut up and stop sucking while the amateurs pushed a bunch of dumb ideas and nobody would listen to education professionals about any of it. Oh, them was the days."

    Amen, brother ... although I would have hyphenated "edu-amateurs."

  2. As you've said before, people who think they know everything cannot learn or grow.

    To these people, "...words are just strategic weapons that have no meaning or connection to reality". More fans of "alternate facts" for the alternate reality they wish to impose.

    At least they didn't blame the evil teachers unions. I wonder how they missed that.

  3. Thank you Peter- as usual you say it best- and most quickly. Though I will add my review soon.