Behold, on stage: a study in contrasts.
In one chair sits Joel Klein, who once oversaw 1.1 million students as chancellor of the New York City public schools and now runs News Corp's education technology subsidiary Amplify, holding court on education innovation. After decades in the public eye, Klein’s persona is remarkably unchanged: a rumpled blazer on his wiry frame, hair that would do Einstein proud, traces of Brooklyn in his straight-talking style.
To Klein's right sits AltSchool founder and CEO Max Ventilla, the picture of San Francisco success in a beard and sweater, his methodical manners born from years spent among engineers. He previously sold two technology companies and ran personalization across Google products, and now is talking about a "personalization first" approach to learning. His education track record to date: Less than a year overseeing a school that serves 20 students.
This may seem an unlikely TechCrunch Disrupt pairing, but the casting is no accident. Today AltSchool is a tiny experiment, but its goals are far more ambitious than its footprint would suggest. While other education startups are zeroing in on specific pain points, AltSchool is taking a comprehensive, "full-stack" approach—a Silicon Valley spin on the idea of value chains—and is experimenting in ways that could overturn our every assumption about how schools operate and evolve.
"It’s very hard to incrementally change schools. Ultimately you need a kind of full-stack approach," Ventilla says. "Everything we do is about how to make a school system get better as we increase in size, and better over time."
He now has the resources to chase that vision; in March, AltSchool raised $33 million in a Series A round led by Andreessen Horowitz and Founders Fund. The capital gives AltSchool the means to continue growing a team that today has as many engineers as it has teachers.
"Most of the products that you look at today—software products, hardware products—personalization increasingly matters, but it’s layered over an experience that at its core is not personalized," Ventilla says. A "personalization first" product has to be built differently and comprehensively, from the very start—and it’s essential to the value that Ventilla hopes AltSchool will provide. "In the 2030s, when these kids are graduating, they’re going to have to be highly independent, very dynamic, able to know themselves and get from the world what they need to be happy and successful. They need to start exercising that muscle in preschool."
Klein and his team are following a more traditional go-to-market approach, centered around district-wide sales of an Amplify tablet custom-built for the classroom. But directionally, he endorses the shift toward individualized instruction. "We’re really in the early stages with personalization, but the alternative is wholly dysfunctional," he says.
Unfortunately, the early stages of personalization look pretty dysfunctional, too. Imagine you’re a math teacher with a student who loves basketball and needs to master probability. At most schools, developing a personalized learning plan for that student that would require jumping between at least half a dozen different systems: Pinterest for inspiration; BetterLesson for planning; Mathalicious for application exercises ("In basketball, should you ever foul at the buzzer?"); Google Apps for managing assignments; Kickboard for tracking progress and attendance; and Remind101 for texting with parents. Overwhelmed yet? Multiply this process by the number of students on your roster, and suddenly "personalized learning" sounds more like a pipe dream than a promising innovation.
In the AltSchool classroom, by way of comparison, students follow personalized playlists based on their learning needs and preferences, and the tasks that typically undercut instructional time—planning, assessment, administrative processes—happen in the background. "The best technologies are invisible," Ventilla says.
If successful, AltSchool would be able "to provide parents with options for their kids in a technology-enabled way that rethinks the service model of schools and the learning experience of schools," says Michael Staton, partner at Learn Capital, which also participated in the round. It’s a daunting undertaking, encompassing everything from admissions to procurement, all in the name of personalized learning.
In practice, that means AltSchool is operating a model that is far more complex than that of a typical new school. On the "front-end" of the stack, AltSchool plans to become a network of private "micro-schools"—picture the one-room schoolhouse of the 21st century, with a couple dozen students loosely grouped by age and a modular layout that seamlessly integrates tablets and other hardware. On the "back-end," AltSchool is software and hardware designed to automate the workflows associated with instruction and operations. The back-end makes personalized learning financially and logistically viable, while the front-end serves as an laboratory for iterating on ideas and gathering data that will improve the technology over time. The goal is to make it possible for students to follow individual academic paths appropriate for their needs and preferences, but in the context of a learning community, and with a business model that can scale.
"The beauty of AltSchool is that it's much more natural to learn in this way," Ventilla says. "Personalization is important to us not just because we think it delivers a better, more satisfying outcome, but also because if an experience is personalized to you as a student, it means that you have choice. That’s going to build your ability to act with agency throughout your whole life, and to do it in a setting that’s much more mirrored to the world at large."
Ventilla’s enthusiasm makes it easy to forget the magnitude and complexity of the challenge AltSchool is undertaking. Knewton, for example—an adaptive learning "back-end" company that has also raised millions from Founders Fund—says it has no interest in the full-stack approach. "Others in the ecosystem are focused on creating content, making pedagogy choices, creating learning environments, and on and on. We are focused on one piece of the education puzzle—enabling the personalization of learning materials," Kristen Weeks, Knewton's senior manager for business development, said in a statement.
Ventilla has a strong track record as an entrepreneur—he founded Krypteian Systems and Aardvark, both acquired by Google—and by all accounts he has a rock-solid team backing him up. But $33 million for a startup school that has yet to prove it’s capable of delivering any educational impact? It’s a bold bet that AltSchool has the potential to transform the student experience, school operations, and ultimately, the structure of the education industry. Let’s look at each challenge in turn.
Transforming The Student Experience
If you grew up attending U.S. public schools, you probably have some rose-tinged nostalgia for the classrooms of your childhood—the ABCs on the wall, the scratch of pencils during spelling tests, the gerbil you should never have agreed to take home for the weekend. What you probably don’t recall is feeling like you were part of a system modeled after industrial-era factories, with their standardized assembly lines and top-down management. But in many ways that factory-mindset has shaped education for decades, requiring that students progress in lock-step along predetermined pathways: Pity the outliers, celebrate the average.
Technology is allowing educators to revisit that model. It’s now possible to individualize instruction in the same way that a tutor would for a pupil, thanks to adaptive learning solutions like Khan Academy and its competitors. But the ecosystem is in its early days—as we saw earlier, integrating solutions is often highly manual—and even factors as basic as the physical layout of school buildings can pose obstacles.
"We are at the very early innings of this new movement of finding new ways to run schools and leverage technology," says Brian Greenberg, CEO of the Silicon Schools Fund, a seed fund for new school models that "blend" traditional and technology-based instruction. "Right now a 1.0 version of blended learning is possible for a highly skilled team of educators. They accomplish that by stitching together the four or five best products on the market and manually doing some of this data exporting and importing. What they can't do is this 2.0 version that really allows students to go at different speeds."
One of those leading products on the market is Teach to One: Math, developed by New Classrooms, a nonprofit that was founded as School of One in 2009 after being incubated within the New York City Department of Education under Klein’s leadership. The catch: When it launched, School of One sounded less like a product for middle school math and more like a comprehensive vision for overhauling the factory model. That positioning landed School of One on Time magazine’s annual list of the top 50 inventions, and for a while, visitors from around the world were touring the pilot program being run out of a Chinatown elementary school in lower Manhattan. At the time I was working for a nonprofit that helped raise philanthropic funding for the pilot; I had no direct role in the project, but I still remember the feeling that we were witnessing something truly disruptive taking shape.
Even now, many of the ideas that distinguished School of One still feel new. To implement the New Classrooms model, schools reconfigure a larger space like a library into a set of stations that can accommodate students working independently, students working in small teams, and students gathered in groups of up to 20. Every day, students follow an individualized "playlist" that determines where they go and what they do when they get there. (AltSchool has adopted the "playlist" vocabulary.) Regular assessments keep tabs on students’ progress and optimize the playlists according to Teach to One’s algorithms. Five years post-launch, 15 schools are using the product.
"We made a choice to focus on just one subject and one grade span in order to really get it right," says Joel Rose, New Classrooms cofounder and CEO. "There are certainly advantages when innovating to controlling the whole school. But there are also some limitations in terms of scalability." Decision-makers in public school districts generally consider research on student outcomes to be a prerequisite for academic products and solutions; as a result, there’s a logic to proving impact one subject at a time.
Promising student outcomes data finally entered the conversation last November when a team at Teachers College at Columbia University released the first comprehensive research on Teach to One. The report found that Teach to One students gained 1.2 years worth of growth in math over the course of the school year, compared to the national average. Moreover, students who started the farthest behind made the greatest gains. "Considering the relatively disadvantaged backgrounds of [Teach to One] students, the fact that their academic gains were above the national norms is noteworthy," the authors wrote.
Even with that encouragement, expanding beyond math to other subjects is not a short-term priority. Rose knows only too well the challenge of matching in the right content to the right skill level and preferred "modality"—virtual tutor, peer-to-peer, live lecture, and so on. His team has screened over 50,000 lessons according to a custom-built rubric; only 15,000 have made the cut and now appear in students’ playlists.
"Schools are pretty complex places," Rose says. "Smart use of software can help. But in the end this movement is about redesigning the classroom in ways that are mindful about what teachers can do and what technology can do."
Ventilla espouses a similar philosophy, with a key difference: His idea of "what technology can do" is far more radical, and the reason that AltSchool sits alongside companies like SpaceX in the Founders Fund portfolio. For example, Teach to One and other learning platforms typically personalize instruction based on highly structured assessment data—in other words, multiple choice questions. That strikes AltSchool as old-fashioned and inefficient.
"A lot of our focus is on creating semi-automated or automated inline assessment so that we can understand what's happening offline in the classroom." Ventilla says. "Even in the relatively near-term that allows us to do things that are relatively sci-fi."
Just how futuristic are we talking? In the same way that a device like a Fitbit can assess your activity, AltSchool plans to install audio-visual classroom sensors that would be able to capture students’ spoken grammar, obviating the need for a formal quiz or test. From there, the possibilities keep expanding—measuring how well students persevere in the face of distractions, or making sure boys and girls participate equally in discussions—through the use of data extrapolated from sensors, wearables, and other background methods.
All this would layer on top of the loosely structured AltSchool student experience. For core subjects, students spend about half of their day following a playlist built around third-party curriculum materials, and the rest of their time doing longer-term projects that can span as many as six weeks. At the end of each week, students complete a playlist reflection that teachers review. Outside of those parameters, teachers are granted significant autonomy. "Teachers are working with a template of the AltSchool experience, and off that template almost everything can be changed," Ventilla says.
As for research, Ventilla argues that the traditional gold standard approach—a large-scale, long-term study with a control group—is meaningless in a "personalization first" context. "You’re not thinking about the global population as one unit that gets this experience or that experience," he says. "Something that’s better for 70% of the kids and worse for 30% of the kids—that’s an unacceptable outcome for us. AltSchool isn't a particular approach."
Transforming School Operations
Despite the early stage of AltSchool’s "evolution" toward a successful pedagogical framework, some parents are already clamoring for AltSchool to expand. By next fall, Ventilla plans to add at least three more Bay Area locations. Interested in becoming a "founding family" in your neighborhood? An earlier version of this sign-up form suggested that by answering a short set of questions ("Why do you want an AltSchool in your community?"; "How many families do you know who will commit to enrolling?"), you’d be on your way.
From admissions to after-school activities, "We’ve very systematically taken out the things that prevent other schools from scaling," says Ventilla, who wants to flip the "scarcity mindset" of the traditional system on its head.
With growth in mind, AltSchool is designing for efficiency: Centralized administrative systems will reduce overhead and manage vendors. Modular classroom layouts will reduce real estate expenses. Loose groupings based on age brackets, rather than specific grade levels, will change the cost structure around teachers and make it possible to more flexibly adjust capacity. Free and low-cost curriculum materials will be used at nearly every subject and skill level.
The price tag: $19,100 per student, which includes $6,000 for extracurricular enrichment at third-party organizations, to be spent at the discretion of each student’s family. While outside the reach of many families, AltSchool tuition is already lower than the $20,000-plus that cities like Washington, D.C., spend per student.
"Within the next few years we'll be lower than the California average, including extracurriculars and everything else. That’s our aim," Ventilla says. That's an unusual goal for an independent private school. (California spends between $7,500 and $10,000 per student.) How exactly will AltSchool hit that target? At this stage, Ventilla will say only that he plans to "use structural engineering and in-house hardware and software to bring down the cost year-over-year by double digits."
First, though, the AltSchool team will have to prove that its micro-school model holds up with multiple sites in operation. "To date we’ve been able to say, ‘we’ve got these amazing educators; AltSchool is what they do in this classroom,’" Ventilla says. "We need a platform that still gives them that discretion and independence but ensures a level of consistency across all the sites. That’s our main challenge."
Transforming The Education Industry
Even as he experiments with the learning experience and cost structure of the micro-school model, Ventilla is looking ahead to other sources of revenue and ways to grow the AltSchool platform. "Our focus initially is on first-party schools so that we can learn by being directly involved," he says. "Down the line you can imagine franchise instances where we work with partners who open a co-branded set of AltSchool locations, and they're involved in doing some of the things that we do centrally around hiring, facilities, and student recruitment."
That vision maps to data indicating that we’re already well on our way to a new conception of "school," even as our existing facades keep up appearances. By 2019, analysts predict that high school students will take half of their courses online, based on growth trends over the last several years. Last spring the Clayton Christensen Institute released a report that played the disaggregation of "school" and "instruction" out to its logical conclusion, predicting that: "The fundamental role of brick-and-mortar schools will pivot. Schools will focus more, for example, on providing well-kept facilities that students want to attend with great face-to-face support, high-quality meals, and a range of athletic, musical, and artistic programs and will leverage the Internet for instruction."
For a full-stack platform, that disaggregation paradoxically presents an opportunity. Someone will end up establishing the norms that bring cohesion to the education that students experience, and the first company to touch the end-to-end value chain for the industry could end up playing the lucrative role of service provider and de facto regulator.
Today, the most urgent need for industry norms is arguably around the tagging of instructional content. Looking for a 30-minute writing exercise for a student who loves history and is struggling with paragraph structure? No two systems will organize your options in the same way, and no two systems will capture progress in the same way. The Common Core standards are poised to serve as a step in the right direction, but on their own they don’t go far enough.
At Silicon Schools, Greenberg sees the problem as a consistent strain across his portfolio. "Developers need to get better about standardizing open APIs," he says. "Right now very few folks are doing that to the level that they should. Too many of the current solutions have the data in walled gardens. That is a huge limiting factor."
AltSchool is already anticipating how it could help the industry in this regard—and at the same time license aspects of its technology for a fee. "We'd love to see a future where a huge subset of the people in the education space are contributing to some kind of repository," Ventilla says. "We are creating the frameworks and underlying structure so that as people do things they add tags to the content and the data. It's our intention to contribute to the broader ecosystem and have a lot of what we do be open source."
For Ventilla, a young dad, the idea of transforming the industry hits close to home. "It's interesting that you're seeing a generation of people who grew up on the Internet have kids," he says. "I don't remember the world without the Internet; it's shaped my entire worldview."
One day, a generation of micro-school graduates might have a similar thing to say about AltSchool.