Privacy & Security

Analytics at AltSchool: Behind the Scenes

By Benjamin Herold — January 12, 2016 6 min read
  • Save to favorites
  • Print

AltSchool isn’t the first ed-tech company to believe there’s a market for tools that improves communication between schools and parents.

But three things make the approach of this Bay Area startup, featured in Education Week’s new special report on fresh approaches to personalized learning, unusual:

  • The company’s in-development app, dubbed Stream, is just one component of a “full stack” of K-12 software tools under development;
  • AltSchool has its own network of private schools that serve as research-and-development labs for those tools;
  • And AltSchool’s commitment to big data means that Stream (and its other tools) is evolving in almost-real time, in direct response to the wishes and experiences of the teachers and parents who are actually using it.

“Most school systems are on a one- to two-year change management timeline,” said Rajiv Bhatia, AltSchool’s vice president of product.

“Here, we’re about continuous loops,” said Bhatia, who formerly held a similar post at mobile-game company Zynga. “We want classrooms using our technology products, and we have an engineering and product team that can react to their feedback quickly.”

Last month, I had the chance to see that process in action during my visit to AltSchool headquarters in San Francisco. The classroom cameras and microphones and sensors in the company’s campuses obviously draw a lot of attention, as does AltSchool’s huge recent success in securing millions in venture capital from Mark Zuckerberg and others. The privacy worries that the company—and the rise of big data more broadly—are prompting are also a big part of the conversation.

Read more about AltSchool in the recent Ed Week story,The Future of Big Data and Analytics in K-12 Education.”

But for me, one of the most eye-opening experiences was sitting in on a product meeting for Stream, the company’s in-development parent communication app.

Having suffered through countless soul-killingly unproductive meetings during my years in academia, nonprofits, schools, and newsrooms, it was a revelation to see just how minutely and deeply everything at AltSchool is driven by data.

Stream’s premise is pretty straightforward: it’s a mobile app intended to facilitate personalized communication between schools and parents. From the schools, that means updates, alerts, and news, as well as a window into class projects, teacher reflections on each student’s progress, and individual student highlights. The app was released within the AltSchool universe earlier this fall. At the time of my visit, hundreds of AltSchool parents and teachers had been using the tool for several weeks.

During the product meeting, company leaders responsible for developing the app, as well as engineers, experts in parent outreach and user experience, and others sat down for a comprehensive review of how Stream was being used, what could be improved, and how the impact of any changes could be tracked.

That’s where big data and analytics come in. At AltSchool, there are metrics for everything. The granularity of the data was (to me at least) pretty stunning. And most importantly, there was plenty of evidence of a top-to-bottom commitment to making sure that information was used to actually drive decisions.

Among the data the product team looked at when reviewing Stream:

  • Parent usage, measured by total views per week, as well as the percentage of parents who viewed the app at least once each week;
  • Teacher adoption, measured by the frequency with which each teacher in each classroom posted updates to the app;
  • Personalization, measured by the number of student-specific posts and “highlights” per student shared over the previous two weeks;
  • Quality, measured by a review of the content of every single post that every teacher had made to Stream;
  • Parent and teacher satisfaction, measured through constant AltSchool surveys of each group.

All those disparate information flows were also looked at in relation to each other.

Lots of numbers were presented during the meeting. Some of the broader insights had to do with the company’s progress towards its core goals, including a) providing parents with greater visibility into what their children are doing and learning each day, and b) saving teachers time.

Based on the data, for example, it seemed that AltSchool parents who expressed the least satisfaction with AltSchool’s “academic transparency” had children in classrooms where teachers made the fewest high-quality Stream posts. At the same time, though, busy teachers in many classrooms were struggling to create more (and more-personalized) messages, in part because they couldn’t post to the app directly via their phones.

Interesting enough, if not entirely surprising.

But what did catch my eye as different—and what AltSchool is staking its future on—is its ability to quickly and continuously turn those types of insights into better products.

During the Stream product meeting, for example, AltSchool developers committed to creating in-app templates that would encourage teachers to create the types of posts that had been found to be most satisfying to parents. They also committed to tweaking Stream to accept mobile updates. For good measure, they also committed to adding in Instagram-type commenting and “like” features for parents. Because AltSchool has a product engineering and design team of more than 50 people, all the changes were to be made and rolled out to schools within a matter of days or weeks.

The impact of those changes would also be tracked: By the following month, AltSchool aimed to have 50 percent of its teachers making mobile posts each week, to reduce teachers’ average time spent posting individual student highlights by 10 percent, and to get 60 percent of parents to have liked or commented upon a Stream post at least once.

That kind of analytics-driven process speaks to a very different philosophy than what I’ve often observed in many other educational systems, where schools’ inability to incorporate or effectively use new technologies most often seems to lead to either hand-wringing, more trainings for teachers, or, most often, indifference.

“When one of our tools falls short in some aspect, that’s a failing of the tool, not a failing of the person using that tool,” said AltSchool CEO Max Ventilla. “If stuff is junk, then it’s not surprising that [teachers] don’t use it.”

As I explain in this week’s special report, the K-12 sector is still wrestling with whether and how to embrace analytics, especially when it comes to increased data collection inside the classroom. Privacy worries, bureaucratic hurdles within school districts, and often-stunning hubris on the part of ed-tech companies themselves have all proven to be significant hurdles time and again.

Special Report: Extending the Digital Reach: Schools Push Personalized Learning to New Heights

Perhaps AltSchool will trip over these, too. Some privacy advocates are already voicing concern over the company’s approach inside its schools. Ventilla—who believes the process laid out above will over time yield transformational technologies that teachers and parents won’t be able to live without—is certainly not lacking in ambition or confidence.

But given the way analytics is already transforming industries such as finance, consumer technology, retail sales, and professional sports, it’s hard to envision a future in which the basic principle of gathering ever-more data and analyzing that information in the hopes of spurring continuous improvement doesn’t become a big part of education, too.

Photo: Bharat Mediratta, co-founder and CTO of AltSchool, talks with colleagues during a product meeting at AltSchool headquarters in San Francisco.

See also:

A version of this news article first appeared in the Digital Education blog.