How good intentions can be misinterpreted – a Big Data flameout

In this case, the good intentions have to do with improving school efficiencies, making teachers’ jobs easier, giving students more flexibility and facilitating parents’ choices. I’m talking about Inbloom, the software platform that was to revolutionise the way schools manage their data. The idea was that students’ data would be standardised and aggregated, so that teaching could be more tailored to the student’s abilities, trends and problems could be spotted sooner, and transferring schools would no longer be a big hassle.

Inbloom educational software

Sounds good, right? Inbloom would give the schools significant cost savings, free up management resources, and make it easier to spot trends and identify problems… Several states adopted Inbloom as the educational standard, and The Bill and Melinda Gates Foundation and the Carnegie Corporation of New York contributed significant funding. Things were going well, schools were signing up, and the administrators and officials were getting excited about the digitalised future of education. But at the end of last month Inbloom announced that it was closing its virtual doors. What happened?

There is so much to talk about here (the role of Big Data in education, the value of standardised curriculums, the security of cloud storage…) that I will leave for future posts. But it is worth looking at what Inbloom wanted to do, why it is a very good idea, the mistakes that they made, and what we can learn from all this.

Inbloom wanted to standardise school data, and to make it easy for schools to input, analyse and share this information. Traditionally US schools have spent untold resources on data input in different, often incompatible platforms. If a student changed schools, transferring his or her data from one school to another could take up to several weeks, which didn’t exactly facilitate integration. And with data spread out so thinly, it was impossible to analyse, so teachers were often working with sparse information and applying general assumptions to disperse groups. Patterns were almost impossible to see, schools could not tell which subjects were working or not, and teachers did not in general have the necessary data to be able to effectively personalize the classes.

Having all the data in one place makes it easier for teachers and school administrators to analyse teaching methods, track student performance, identify outliers… And the data is more easily transferrable if students change schools. Up until now the data has been in different databases, with different formats, and was not easily transferrable, which made following a child’s progress as he or she passed through several schools very difficult indeed.

But, Inbloom made several big mistakes. When it comes to their kids’ data, parents are understandably sensitive. Inbloom did not focus on speaking directly to the parents, addressing their concerns and explaining the security controls. It chose to let the schools do that, and even dropped out of its social media channels. But it turned out that the schools and state departments in general were not equipped to talk about Big Data fears, and took the easier route of backing down. The media backlash prompted states to drop out of the programme, and the loss of the New York state schools was the final blow.

Frankly, some of the parents’ fears are quite understandable. For instance, data permanence: for how long is the information stored? Let’s say your son or daughter one day decides to run for President of the United States. Do you really want the number of times that he or she was late to class to be a matter of public record? Even if your child’s ambitions are nowhere near that high, parents do not feel comfortable handing over sensitive data like academic record, health information, family relationships and disciplinary actions to a private company, even a non-profit that cannot in any way sell the data.

And as for public record, who owns the data, the school or the state? The data is put in by the school, but the state is Inbloom’s client. If a school wanted certain data deleted, for example, it would have to get permission from the state. And what if a parent simply doesn’t want the data shared, is there an opt-out option? Do you have to change schools, move states, leave the country?

The thing is, this data is all already in the system anyway. Health information, attendance records, etc., are stored on computers somewhere. What Inbloom wanted to do was to aggregate the data for efficiency and for insight. What was not well handled was the data control issue. It seems that we are more comfortable if our data is fragmented, because is it less likely that a bigger picture about us as individuals can be drawn. We understandably fear the labelling and the misunderstanding.

As the CEO said in a message on Inbloom’s webpage, “this concept is still new, and building public acceptance for the solution will require more time and resources than anyone could have anticipated”. Sensitivity should be added to the list, as well as perseverance. Inbloom or its successors need to work with all of the participants – the states, the schools, the parents and the students – to figure out what everyone wants, to calm fears, to add features and even to help draft a Students’ Privacy Bill of Rights.

We’re not yet comfortable with Big Data, especially when it concerns our kids. But at least we’re having the discussion, we’re starting to get a glimpse of the possibilities, and better information and communication will eventually help us to accept that the efficiencies are worth it. May Inbloom’s demise not be in vain.

— x —

For more on online education, check out my Flipboard “Internet and Education”:

flipboard education

Leave a Reply

Your email address will not be published. Required fields are marked *