Data In Schools: Learning From InBloom’s Failures

Education is an industry where data analytics has a lot of potential but has been slow to pervade. The amount of data a school district tracks about its students is enormous. This data, while hard to maintain, provides a huge opportunity to deliver personalized learning: a student can get the teaching and support tailored exactly to his/her needs. However, consolidating and making sense of this data is a very challenging proposition for schools and is not a part of their core competencies [1]. The data is usually messy with no standardization between various databases within a school; the attendance database could store data very differently from the grades database. Among different schools, data formats vary even more.

In 2011, the Bill and Melinda Gates Foundation and the Carnegie Corporation of New York, provided a $100 million grant to InBloom, a non-profit, to tackle this massive technological challenge [2]. InBloom promised to create value for schools by taking all the data about each student stored in fragmented data stores, massage it into a uniform format, store it in the cloud protected by world class security standards and then use it to populate easy to use dashboards [3]; teachers could then use the dashboards to track each student’s progress on various academic and non-academic dimensions and deliver personalized learning.

inbloom
A sample student dashboard InBloom created using fake data to showcase its technology [3]
While most states planned to share data only from selected districts, New York and Louisiana promised to share data from the entire state [4]. Cleaning and consolidating this data opened the doors for many other providers to use it to power other learning applications and easily deliver them to all schools on the InBloom platform. The entire platform was going to be open source for anyone to understand and contribute to. The edtech industry was buzzing with joy [2]. Unfortunately, this glee was not meant to last.

InBloom soon started to get immense backlash from parents who were uncomfortable with the idea that all their child’s data was being shared by the school with a third party. Parents and privacy advocates argued that the sensitive data could be sold by InBloom to marketers [3]. Moreover, a lot of parents believed the cloud was unsafe and data could be stolen by “hackers”. Since personalized learning was so nascent and the technology not established, the general sentiment was that the risks greatly outweighed the benefits [5]. So strong was the reaction that within two years, three states pulled out, one put its data on “indefinite hold” and three said they never planned to share any personal data in the first place. The final blow came when New York passed legislation prohibiting the state department from giving data to aggregators like InBloom, causing New York, the largest project, to back out. Within three years of launching, InBloom decided to shut down its operations [3].

In examining some of the root causes of InBloom’s failures, there are a few things it could’ve done differently to avoid the downfall.

  1. There was a significant lack of trust between InBloom and its various stakeholders. In most states, data was being shared with InBloom without any parental notification or consent, and without allowing opt out [6]. InBloom left the job of assuaging any privacy concerns to school districts. Further, InBloom maintained the right to sell the data to other providers. InBloom could have worked more closely with school districts in addressing these concerns and clearly indicated that the data would only be used for facilitating learning, and not for advertising and marketing applications.
  2. InBloom could have helped assuage the security concerns. Despite an independent third party audit that InBloom was safer than many existing systems in use [7], parents still believed the provider was insecure and InBloom made no efforts to change this. InBloom remained extremely apathetic to all security apprehensions raised and came across quite arrogant in all these interactions.
  3. Instead of going all out and getting many states on board, InBloom could have started with a pilot site and built an application that teachers in the site could have tested out to see the benefits. Involving teachers and schools as thought partners would’ve also helped them get advocates for their cause in the community when things got rough.
  4. Similarly, InBloom could have restricted the amount and types of data they required initially, and gone for ‘small and actionable’ data [7] over ‘big data’. The InBloom database included more than 400 fields about students, some of which were very intimate such as family relationships (e.g. “foster parent”) and reasons for enrollment (e.g “leaving school as a victim of a serious violent incident”) [4]. In initial applications, sensitive fields could have been omitted.

Overall, InBloom’s failure demonstrated the overarching lesson that in industries that are slow to adapt to technological change, organizations need to be quite thoughtful, deliberate and engaged when trying to roll out digital transformations.


Word Count: 798 Words

Bibliography

[1] Lorna Earl, Steven Katz, Leading Schools in a Data-Rich World, Second International Handbook of Educational Leadership and Administration, 2002.

[2]  With $100M From The Gates Foundation & Others, InBloom Wants To Transform Education By Unleashing Its Data, TechCrunch, February 2003, https://techcrunch.com/2013/02/05/with-100m-from-the-gates-foundation-others-inbloom-wants-to-transform-education-by-unleashing-its-data/

[3] InBloom Student Data Repository to Close, The New York Times, 2014, http://bits.blogs.nytimes.com/2014/04/21/inbloom-student-data-repository-to-close/?_r=1

[4] What the Failure of InBloom Means for the Student-Data Industry, The Citizen’s Guide for the Future, 2014, http://www.slate.com/blogs/future_tense/2014/04/24/what_the_failure_of_inbloom_means_for_the_student_data_industry.html

[5] InBloom’s Collapse Offers Lessons For Innovation In Education, Forbes, 2014, http://www.forbes.com/sites/michaelhorn/2014/12/04/inblooms-collapse-offers-lessons-for-innovation-in-education/

[6] InBloom Background, Parent Coalition for Student Privacy, http://www.studentprivacymatters.org/background-of-inbloom/

[7] Privacy Fears Over Student Data Tracking Lead to InBloom’s Shutdown, Bloomberg, 2014, http://www.bloomberg.com/news/articles/2014-05-01/inbloom-shuts-down-amid-privacy-fears-over-student-data-tracking

[7] The Real Lessons from InBloom, IMS Global Learning Consortium, 2014, https://www.imsglobal.org/article/real-lessons-inbloom

Previous:

Will you please Clarifai?

7 thoughts on “Data In Schools: Learning From InBloom’s Failures

  1. I’ve never heard of this company, but it doesn’t surprise me that it didn’t stick. I can’t imagine parents today wanting their kids’ data to be circulated, let alone parents 5 years ago. I think there is definitely room for digital innovation in education, but i’m not sure I see the value in digitalizing children’s data before we can figure out a school system that actually delivers a quality education.

  2. I wonder whether there is potential to leverage a supposedly un-hackable platform like blockchain to help store personal information. There seems to be so much insight to gather across industries – education, healthcare, etc. – if we would just be comfortable sharing our information. Democratization of that data would be awesome too. Why should one company have access to data, rather than having a central repository to share among everyone. The same holds true in healthcare, ranging from topics like genomic data to access to blood banks. Why can’t we all just share – particularly given that the source (people) don’t ultimately benefit from a silo’ed system?

  3. Radhika, thank you for posting this. I had very little knowledge about the company or the ed-tech industry before this.
    I personally think, privacy in this day and age is overrated. Anybody with any technology connected to the web, has most of their information whether its hard-fact data or derivable analytics, available to large corporations that may chose to sell your data or market to you based on your preference. I believe the challenge to making programs in ed-tech plausible is to acknowledge that such marketing tactics exist and actively addressing them by emphasizing that the ed-tech companies take a moral/ethical high-ground as an industry to market to children below a certain age. This needs to be policy and industry driven effort can potentially lead to a buy in from parents.

  4. This is a really interesting example of a well-intentioned platform evoking users’ fear. As we enter the digital era, it seems to me that companies are innovating faster on democratizing sharing than protection. Sharing pervades everyday interactions among civilians, while security builds up at the highest levels but is slow to evolve for regular users. This wasn’t an issue I was overly concerned about before, but with this new unpredictable administration, this concern really resonates with me. I wonder how the industry can help incentivize innovation on accelerating user privacy tools?

  5. Thanks for sharing this. I wasn’t aware of such technological developments occurring in the education sector. Personally I am torn between SubodhChawla and Christie’s comments. While so much of our information is available today and the intention behind the technology seems to be good, it doesn’t sit well that academic information is being shared with a third party. In particular the risks of sharing with such a third party need to be clearly outlined and consequences of a breach of such data made clear to all parties before implementation. I would also be interested to understand how the Bill and Melinda Gates Foundation went about selecting InBloom – were there any other similar parties competing for funding, and if so, how did InBloom differentiate itself?

  6. I loved that you posted about a company that failed. It serves as a great learning experience for us all.
    I’m perplexed why they were so committed to the cloud. Could they have had the system running on computer systems within each school, and send the data over intranets (or encrypted data sharing or even hard drives) versus relying on the cloud? Because, based on your post, the privacy concerns are what started the downward spiral, not the technology itself. The cloud seemed to be an unnecessary capability that ultimately fouled the product. As a Principal, i would be just as happy if the system was resident in my school and provided our local teaching staff a clean dashboard. My old company ran into a similar issue. We wanted to use a software suite that was only available on the cloud, and the company wouldn’t install it on our network. As a defense contractor, that wasn’t an option and we ultimately had to use a less effective, but locally implementable, product.

  7. Radika, thank you for writing about this important topic. There are few issues less pressing today than urgency to reform our education, and technology, no doubt , will play key role in its transformation in the next few decades. Despite lack of initial success, I am glad companies like InBloom are pushing the envelope towards digitization of education. Long term, the only way to increase the quality if to increase compensation for high-quality teachers and make teaching a prestigious job for the best and brightest. Technology will help us reduce number of jobs needed those allocating more funds for the few high quality teachers. I am glad you have shared this real life learning with us in the post, I am looking forward to learning more about this space while at HBS and beyond.

Leave a comment