Back to Issue

Chapter 3: What Does Blended Learning Look Like in Practice?

New blended learning programs are beginning to be spawned in many parts of the U. S. Some early-adopting charter schools have already gone through a few academic years testing different approaches to blended learning. In many cases these schools have been energized by philanthropic dollars as they’ve pursued their objectives, studied what works, and gradually expanded. Their experiences offer insights that other foundations wanting to invest in this field need to know about.

In addition to its operative contributions, the philanthropic community has also played a major role in bringing public attention to these blended-learning pathbreakers, so that other educators, policy makers, and families can know what options exist. The Jaquelin Hume Foundation, for instance, has supported the creation of several promotional videos on blended-learning schools (see sidebar on page 39). In 2011, the Michael & Susan Dell Foundation commissioned white-paper studies of five blended learning schools—Rocketship, Summit, College-ready Public Schools, KIPP Los Angeles, and the FirstLine schools in New Orleans. These in-depth profiles probe the finances and instruction and operation models of all of these schools (you can find links to them in the appendix). “We wanted to tell a very rich story about what it means to look at and understand and potentially operate models like these,” says Cheryl Niehaus, an education program officer at the Dell Foundation.

When it comes to blended learning, seeing is believing. In that spirit, here are quick profiles of some of the best-known blended-learning schools.

Rocketship Education

Before co-founding Rocketship Education in 2006 (with Preston Smith), John Danner had two careers: as a Silicon Valley CEO whose start-up, NetGravity, was acquired in a 1999 stock deal valued at $530 million, and as an educator in Nashville. He brought experiences from both these ventures to Rocketship—perhaps the best-known of the blended-learning school models. Rocketship focuses primarily on low-income and urban students, and requires fewer teachers than traditional school models, yet still achieves better results.

Rocketship’s first school opened in San Jose, California, in 2007. Mateo Sheedy Elementary School serves about 500 students, 88 percent of whom qualify for free or reduced-price lunch, and 65 percent of whom are learning English as a second language. Danner and his team subsequently opened six more elementary schools in the San Jose area, and (with funding from the Lynde and Harry Bradley Foundation) plan to expand to Milwaukee in the fall of 2013. The organization has also won charters to operate in New Orleans, Indianapolis, Memphis, Nashville, and Washington, D.C. The ultimate goal is to operate in 50 cities and serve one million children. Though a number of foundations have aided Rocketship’s growth, key aspects of the organization’s business model make day-to-day operations and replication financially sustainable in a way few other charter schools have been able to emulate.

Here’s what a Rocketship education looks like. On a balmy California May day, parent volunteers with younger kids in tow are humming around the school receptionist at Rocketship Mosaic Elementary, located on Owsley Avenue in San Jose. Most of the students and parents are Hispanic (English is a second language for 80 percent of the students at this location). Signs on the wall honor the parents who’ve put in the most volunteer hours. Some reach the “Jupiter” level—500–600 hours—by May. That’s the equivalent of a serious part-time job.

The building is bright, clean, and cheery. Some kids play outside, others sit in classrooms—very typical, colorful elementary school classrooms that are not technology-based at all. The only noticeably distinguishing feature of the school is a large Learning Lab set up in what would be the cafeteria in many schools. Since this is sunny California, children here eat lunch outside on picnic tables under a green awning, munching on Revolution Foods (a provider of healthy fare).

In the Learning Lab, dozens of children sit at Acer computers ($150 apiece) in long rows, separated by brightly painted cardboard dividers. Students cycle through this lab, spending approximately two hours per day there. The software—games and problem sets from different providers—runs about $20 per student in group licenses, and since it’s not bandwidth-gulping video, a basic Comcast broadband package suffices to power the lab. When the children shuffle in, they sit down and follow signs reminding them of the proper learning posture: headphones on, no talking, raise your hand if you need help.

By educating students for about 15 percent less than the normal state allotment, Rocketship’s business model makes replication financially sustainable—unlike many other charter schools.

The software covers basic elementary-school skills: reading and math. But each child is covering a different set of skills at any given time. Every child’s sequence of programs, sometimes called a “playlist,” is a little different. Over the years, Rocketship’s educators have developed algorithms to predict which programs different kinds of students will respond best to. If you’re an English-language learner with certain test scores, for instance, you’ll get one default playlist, to be changed as needed. Teachers get immediate data on how students are doing on different skills. During the rest of the day when the kids are not on the computers, they receive small-group instruction from teachers informed by computerized data, and do group projects.

In 2013, Rocketship announced they would be modifying their Learning Labs, placing three teachers with each pod of 90 students in the lab, and integrating more direct instruction with the online resources. Children will still cycle through work at the screen and keyboard using their individual playlist, but it will happen closer to their teacher, with the hope of gaining closer integration between what the machine offers and how the teacher follows up.

The two hours per day of computer instruction will still allow teachers to avoid burning time teaching or reviewing basic concepts that machines do better. “It’s insane that most schools do spelling in classrooms,” says Danner. There are many opportunities to teach children repetitive, drill-intensive tasks like this more effectively by computer. The teacher can focus sticking points with students individually or in small groups.

As Rocketship alters its Learning Labs, it will be crucial to see if they can maintain what has been one of their biggest comparative advantages to this point: the boost in teacher productivity that comes from reserving teachers for higher-level instruction, while using aides to oversee children’s time at the computer stations.

The Learning Lab model allowed Rocketship to operate with roughly six fewer teachers per school, meaning that “we save 25 percent of salary costs,” says Danner. “When you have that, you can grow without raising additional capital.” Even after paying its smaller number of teachers better than other schools, Rocketship is able to educate a child for about 15 percent less than California’s annual per-pupil allotment, and it plows that margin of funds into, among other things, teacher training and opening new schools. Hence, the model should be able to expand like successful businesses do, without constantly needing new financial angels or capital infusions.

“True blended schools,” says Danner, “are financially scalable.” Rocketship itself is in the midst of rapid growth. It has plans to expand from seven schools to more than 20 over the next five years.

More important than the cost savings, though, is that Rocketship’s financially attractive model gets results:

  • The network had an overall score of 855 on the 2012 California Academic Performance Index (API). The target for schools is 800.
  • Even though 90 percent of their students are low income, and 70 percent come from non-English-speaking homes, fully 80 percent of Rocketship students scored at the “proficient” or “advanced” level for math on the California Standards Test—not far from the 83 percent of students in California’s 10 most affluent districts who scored the same.

Of course, there are other things besides the blended-learning model driving results at Rocketship schools. The culture of parent involvement matters a great deal. If kids see their mothers and fathers in the school building frequently—and 500-600 hours over 30 weeks would count as frequently—they’ll get the message that learning is important.

Rocketship also has many excellent teachers. Danner exaggerates only a little when he says he has pretty much outsourced his teacher recruiting to Teach For America. He takes a number of TFA placements each year, and also hires alums of the program. While TFA members are younger and don’t have much experience, many of them come from top colleges, and its recruitment process is quite selective. More than 48,000 young people applied for about 5,800 TFA placements in 2012.

And Rocketship’s teachers are not unionized. The lack of union restrictions lets the schools adapt quickly and transform themselves as conditions demand. Rocketship also pays teachers more—at least 15 percent more, and up to 30 percent more than neighboring district teachers.

Rocketship has worked through plenty of challenges over the past few years, and will face more as it expands. For instance: student playlists often include software from multiple providers, in order to capture the best approach to each topic. But the software doesn’t integrate easily. As the Dell report on Rocketship noted:

There is no common definition of mastery across online programs. This means, for example, that when one program reports that a student has mastered fractions, this conclusion may not be shared by other online programs or by Rocketship’s own system of classroom assessments. Taken together, these issues mean that it has been difficult for Rocketship teachers to access the sort of consistent and reliable data on student progress towards the mastery of standards that they would use to directly drive classroom instruction. Instead, the data that teachers currently access is most useful for showing which students are on task, which can be helpful in motivating students and managing student behavior.

Rocketship has elected to store its software in the cloud whenever possible to save money on IT infrastructure. But, as in all blended-learning programs, IT needs are more intense than in traditional schools.

The quality of the software content isn’t always great, though it’s getting better. One of the biggest problems the school has faced is that the six different major programs they use all have different mechanisms for measuring results and giving teachers feedback. Sometimes that’s too cumbersome to help teachers know what they need to explain differently the next day.

Nonetheless, Danner is optimistic about solving blended-learning headaches as they come up. He has turned to philanthropists for help in the early stages of development and expansion before the benefits of scale are achieved. The Gates Foundation, for instance, paid for a program that helps integrate different software so children can use a single sign-on and teachers can have more useable data. This program will be useful for other schools adopting blended models as well.

With this assistance, Rocketship Education has refined its model over time so that it is more easily duplicated. “We’ve seen our peaks and valleys flatten over the last five years,” says Danner. “Now when we’re opening a new school, 50 percent of the result is almost guaranteed unless someone sets fire to the computers. Most schools don’t have that cushion.”

To capitalize on the valuable experience Rocketship is gaining, the Hume Foundation gave a $100,000 grant to the school to pay for public relations and bringing John Danner to speak at different events. This helped raise the profile of Rocketship nationally. The Eli and Edythe Broad Foundation invested $1 million, starting in 2010, to support Rocketship’s growth and expansion over a three-year period.

Alliance Technology and Math-science High School (ATAMS)

Alliance Technology and Math-Science students use the in-class rotation model. ATAMS is one of five high schools housed at the Sonia M. Sotomayor campus in Los Angeles. This brand new complex was built to ease overcrowding, and much of it is shiny and new, with an inspiring view of the mountains from the open-air common areas.

All ATAMS students have gleaming white Apple laptops, and in each large classroom students are broken into three groups doing different things. Some are progressing solo through individual problem sets on their laptops. Others are working on group projects. The rest are working with the teacher. They cycle through these three stations in each class.

The most striking thing about ATAMS classes is how many students are in each group: 16. This gives each ATAMS class an official student-to-teacher ratio of 48:1. Yet with students absorbed in their laptops, the classes feel surprisingly intimate and engaged. Moving between three stations in each class period means it “doesn’t get as boring,” says Paulina, a student. “You get to switch around. For me, it makes the time go quicker.”

Those teacher ratios create vital economic advantages for ATAMS. The combination of salary efficiencies and careful control of online content costs should ultimately allow each of the schools to save more than $1 million over five years, compared to a conventional school, even including startup costs. That encourages donors to think the model can economically spread to many places.

The material that students cover each day in an ATAMS classroom is partially differentiated for each individual. A glance around laptops in an English class, for instance, finds that everyone is reading about the USDA’s different climate zones. But students report that some of the reading passages have more advanced vocabulary than others, depending on how much a student shows he or she can handle.

Teachers get immediate data on how their students are doing—which the students pick up on. “The teachers are more on you,” says Chauncey, a young man who shows the dashboard on his laptop to a group of visitors tiptoeing through his class. “They know you more. They care about how you’re doing.” Students can see instantly how they’re doing too, he says, so they understand what they have to work on. Mickie Tubbs, the school’s principal, agrees: “That’s the magic of this model. It’s personal.”

There are kinks still being worked out in achieving this ideal of personalized education. Partial differentiation is better than no differentiation, but differentiation in ATAMS classes still occurs within a relatively narrow band. This is partly a function of the available software—something all blended-learning schools are struggling with as educators try to bridge the gap between the leading edge of adaptive software and daily classroom realities—and partially a function of sticking very closely to a three-station rotational model. The ATAMS staff is quite enthusiastic about this model, and though it does make classes go quickly, the downside is that it serves to keep the whole class on fairly similar material. After all, every subject must include group-project time, whether that group work is producing obvious benefits or not.

This portion of the ATAMS formula is the least proven. The case study on ATAMS produced by the Dell Foundation found that “staff have acknowledged that the collaborative station might lack this element of rigor . . . whereas the direct and online settings foster rigorous instruction on state standards.” Steps to rectify this include giving students specific roles in the group rotation, letting them rate themselves and each other, and offering exit slips from this station. Whether that’s enough is not clear. “Collaboration” is a perennially popular cause among educators, but making it (and related strategies like “peer tutoring”) work in practice is difficult, whether you’re in a blended school or not, and it’s possible that hewing to a three-station model will prevent breakthroughs that might otherwise result from the strengths of the online and small-group instruction.

Despite these challenges, something in the ATAMS method is powerful. In one academic year, the first cohort of mostly minority, high-poverty students achieved multiple years of academic gains, according to Frank Baxter, co-chairman of ATAMS’ umbrella organization. This brought the achievement average for the students enrolled from way below grade level up to something approaching the norm. Behavior is also much better. Students become engaged in their own progress and “there are no disciplinary problems,” says Baxter. “I walk in pretty frequently and very few heads come up. They’re so involved in what they’re doing.”

Tubbs notes that the ability of this more personalized teaching method to meet students at approximately their level means that teachers feel less burdened by their lagging students—students who the data reveal have been woefully underserved by their primary schools. Wendy Chaves, who teaches math at ATAMS, says that “a lot of kids need remediation. They have a lot of learning gaps.” She discovered that “sophomores didn’t know how to do fractions. It’s difficult to learn algebra II when you don’t have those concepts.” She’s spent much time trying to encourage students who “hated math, flunked math. Everything they tried did not work out.” But as they take charge of their own learning, she’s trying “to mold them and have them see their potential. I want to make them aware that they do have potential, they just haven’t had a chance to succeed.”

Tubbs agrees that at teacher meetings “the conversation isn’t about ‘that kid.’ It’s about ‘how can you help me solve this problem with this kid?’” This new interest in appropriate instruction—rather than mass teaching that tries to pound a uniform curriculum into varied students all at the same pace—should pay dividends in the long run. Based on experience at its blended-learning schools, the Alliance for College-Ready Public Schools (of which ATAMS is a part) is adopting blended models in the new schools it opens. The Broad Foundation is helping to fund this expansion with start-up capital so the schools can plan and purchase technology. The Hume Foundation also gave a grant of $100,000 to help ATAMS with a curious logistical problem: based on initial positive results, the school has so many visitors and queries that it needs additional staff just to handle these requests.

Summit Rainier and Tahoma

Diane Tavenner worked as a public-school teacher and administrator for a decade before founding Summit Public Schools, a network of charter schools in the San Francisco Bay area. Summit schools aim to make all students college ready. The network hires high-quality teachers and provides them with 40 days of professional training every year. Students get insights into careers and non-traditional subjects through a four-week intersession, during which they can take courses on anything from dance to professional cooking.

The personalized teaching made possible by new technology allows teachers to meet students at their level.

At Summit schools, all kids take multiple Advanced Placement classes. Almost all graduates (96 percent) are accepted at four-year colleges. Yet, according to spokeswoman Mira Browne, when Summit’s leadership team looked at the first graduates’ performance in college, they realized that many had needed remedial classes, particularly in math, and many others had not persisted to graduation. Roughly half graduated from college on schedule.

“That’s still triple the national average,” Browne says, “but for us at Summit it’s not nearly good enough. That’s not why we’re here. We want 100 percent into college, through college, and becoming contributing members of the workforce and society.”

School leaders realized that even as they successfully taught each year’s curriculum, their students had unfilled gaps from earlier in their schooling. While “our kids are taking and passing AP Calculus . . . what we weren’t able to do in the current model is fill in every single one of the holes and gaps from elementary school.” So they turned to blended learning to make sure students mastered basic skills before they took up later subjects.

In 2011 the Summit network opened two new schools called Summit Rainier and Summit Tahoma, housed in the same campus near National Hispanic University in San Jose. These schools began using blended learning in math courses, based primarily on the Khan Academy’s free library of online videos with math problem sets. A team from the Khan Academy (whose staffing was made possible, in part, by grants from the Gates Foundation) worked directly with Summit teachers to make the math sequences work for students in a classroom environment, and to give teachers useful feedback via a “coach” feature that allows a teacher or parent to log into a dashboard and check on student progress.

The schools were sufficiently happy with the results from 2011–2012 that teachers made plans to use more blended learning during the 2012–2013 school year, and to expand the use of blended learning within Summit’s charter management organization, an undertaking funded in part by a $2 million investment from the Charter School Growth Fund. Much of this curriculum development was worked out during the Summit schools’ annual 40 days of professional development. One thing teachers talked about a lot during this time was how to use blended learning to meet the needs of the top 10 percent and the bottom 10 percent of students—populations that get lost in the shuffle under mass teaching methods. In theory, blended learning should easily meet these students’ needs, yet they proved slightly more difficult to serve in Summit’s model than theory would suggest.

“We felt like we did a really good job opening up the class in a way that students could work at things at their own pace—and not get rushed if they didn’t understand, or not get slowed down—mostly with the middle 80 percent of students,” says Jesse Roe, a Summit math teacher. But at the extremes, things were more complicated.

Even students who successfully learn the year’s curriculum sometimes have unfilled gaps from earlier in their schooling. Blended learning is able to find and fill these.

“Our top 10 percent could have been challenged more. A lot of the reason for that is a bandwidth issue on our part. We’re always trying to create new content for those students, but we didn’t have time to create something really rich. So now we’re trying to improve on our top-end content. . . . We want to always have something available for students when they race ahead.”

As for the 10 percent of students who were struggling most? “We completely did our best trying to allow them to go through at their own pace, and go as far back in content as needed,” says Roe, but the problem was that Summit’s teachers were “still expecting all students to go through the algebra and geometry course.” With an eye on meeting the ninth- and tenth-grade standards on time, “we probably didn’t give them enough time to develop that foundation.” After discussion and development, though, “We’re moving toward a competency-based model, where it doesn’t matter if you’re in ninth or tenth grade.” The goal is to create an environment where a child gets instruction on the material that he has yet to master, no matter what his grade status. Better to go back and fix knowledge gaps, because “your skipping this content and not developing a foundation will just slow down your learning in the future.”

As Summit moves toward competency-based learning, it aims to focus on what students know instead of age-based grade levels. “Students work at their own pace,” notes Alex Hernandez of the Charter School Growth Fund, which is supporting Summit’s expansion. Eventually, students may be certified as “either high school-ready, college-ready, or career-ready.”

The goal is for every child to get instruction on the material he has yet to master, regardless of where that material falls in grade status.

Despite working closely with Khan Academy, there have been logistical challenges in using Khan online content as the basis for in-school classes. For instance, Khan Academy didn’t have enough material to fill an entire year of class time devoted to algebra and geometry. To stay ahead of Summit’s swifter students, the team writing Khan’s problem sets had to race to create new and deeper extensions of subject knowledge for students achieving “streaks”—getting 10 problems in a row correct, as required to move forward under Khan Academy’s mastery rubric.

Another problem was that streaming the Khan videos requires bandwidth. When 100 students watch videos simultaneously, that’s a lot of bandwidth. Summit had to massively raise its technology budget in the course of the year to upgrade its internet service.

Summit has also faced architectural challenges. In a high school, where students tend to go to different classrooms for different subjects, the computers can’t be in centralized learning labs—every classroom needs to be set up for blended learning. In some older school buildings like the one Summit Rainier and Tahoma inherited, a classroom might only feature two electrical outlets, let alone the high-capacity broadband necessary for video-based instruction.

These challenges are surmountable, but lining up dozens of power strips, or wheeling around a laptop-charging cart, just reinforces how different blended learning is from what most teachers and schools have dealt with before. Even experienced teachers become, in some ways, like first-year teachers, trying to keep an early-adopter mindset while still meeting the needs of the children in front of them. “This is messy stuff,” says Tavenner. Like any product, “you have to test and iterate.”

Carpe Diem

Carpe Diem Collegiate High School and Middle School in Yuma, Arizona, started as a traditional charter school. After losing its building lease, the school struggled to find a new space. One of the few facilities available was a former call center which featured an open floor of cubicles with offices and meeting rooms on the perimeter. School founder Rick Ogston decided to take advantage of these circumstances by shifting to a blend of heavy computer-based instruction supplemented with old-fashioned teaching.

The on-campus version of Carpe Diem (there’s also a fully online version) is small—capped at 300 students in each school for grades 6–12, according to Ogston—because “we believe it’s best for the culture.” With 300 students, the school has a grand total of four certified teachers: one math teacher, one for science, one for social sciences, and one for language arts. That creates an official student-to-teacher ratio of 75 to 1, though there are also non-certified assistant teachers/coaches who provide instruction. Because it needs only one master instructor in each field, Carpe Diem recruits the best of the best teachers—and pays them a lot more.

Students work primarily on their computers, originally using Education2020 software (a situation that was evolving as of this guidebook’s writing). There were plenty of things that weren’t perfect about this software, notes Ogston, but rather than seeking out the flashiest curriculum in each subject “we prefer one that can manage the whole package for us.” That gives the school better control over the whole education being offered. Ogston notes that “what we lose in flashiness we can make up in the workshops.”

Those workshops are a key component of teacher interaction with students at Carpe Diem, and make the 75-to-1 student-to-teacher ratio feel relatively intimate. Teachers are “handed reports several times a day electronically.” These reports inform the content of the workshops and the individualized tutoring-style instruction students receive.

This process isn’t as straightforward as it sounds. “I thought that teachers initially would understand the data more,” says Ogston, but he soon learned that blended learning requires a lot of professional development because the data can be “overwhelming.” Teachers need to learn what is important in the scoring feedback on each child, and why, “and what do I do with it?” But after much work, “they do seem to get it,” says Ogston, and it makes them more efficient. “When they are actually involved at the student-data level and know what the student needs, they don’t have to spend a lot of time just shooting to the middle. They strategically go in and one-on-one with students or small groups and teach to what they need, rather than just hope to reach some people.”

Another challenge? Acclimating students to self-directed learning. “Some students are not prepared for individualized learning,” Ogston says. “They’re used to cohort learning and not carrying their weight.” Sparking motivation becomes a key job of Carpe Diem teachers—as it is in any school.

Like Rocketship, Carpe Diem offers possibilities for sustainable expansion, because it has managed to squeeze productivity gains out of technology, primarily by employing fewer teachers. This makes the school affordable, even if per-pupil allotments stagnate. Per pupil costs run about $5,000 per year. That’s well below Arizona’s average of $7,600 per year. These savings should allow school leaders to replicate their model broadly over time without requiring large new capital infusions.

Because it needs only one master instructor in each field, Carpe Diem school recruits the best of the best teachers.

This opens up fascinating opportunities. Dan Peters, president of the Cincinnati-based Lovett and Ruth Peters Foundation, argues that while philanthropy can incubate new school models like Carpe Diem, the most promising way to expand them into a critical mass is for them to become profitable entities that generate returns they can use to expand to the next location. “One of the main reasons progress is so slow in education reform generally is because almost all reformers are focused on the nonprofit model. That means there’s very limited ability to expand, since access to capital is so difficult,” explains Peters.

If dramatic school improvements are going to be brought to a larger chunk of the 50 million kids now being instructed in America, the powerful tools of the profit motive need to be more widely employed, Peters argues. “K–12 is a $600 billion market,” he notes. “Reformist schools don’t need millions in capital, they need billions if they’re going to reach a 5 percent market share.”

As a start toward this hoped-for phase where profit margins can power continual expansion, Peters has encouraged Carpe Diem to use commercial levers to expand. His foundation provided the planning grant that allowed Carpe Diem to create its second physical school—which opened in Indianapolis in August of 2012. Then Peters helped Carpe Diem come to Cincinnati by obtaining bank loans, which will be repaid from the margin between the school’s incoming fees and its costs. “If the school is as good as we’re hoping, and it can retain its cost advantages, bank loans will make expansion much easier than waiting for philanthropic dollars,” Peters explains. In December 2012, the Cincinnati Board of Education gave the green light to the opening of a new Carpe Diem campus in the fall of 2013.

Of course, finances are only one of the hurdles that a new-schools entrepreneur must leap. Carpe Diem’s first expansion to Indianapolis was not entirely smooth. Failing to find a suitable space, Ogston had to build a new building (paid for with school start-up funds), which was completed—“dirt to fully operating school”—in 49 days. Because no one was sure if the school would open in time, some potential students enrolled elsewhere.

Local media and some bloggers also raised flags about the 75-to-1 student-to-teacher ratio. They questioned the dramatic increases of Carpe Diem students on Arizona’s annual achievement tests—92 percent proficiency on the state test in 2010, as compared to 57 percent in the rest of Yuma County, and 65 percent across Arizona generally. Critics questioned how the school could achieve such results on per-pupil costs of just over $5,000, as compared to the Arizona-wide average of $7,600, and hinted that perhaps cheating was involved.

Nonetheless, the school opened for the 2012–2013 school year. It now serves about 100 pupils from diverse backgrounds, including a number of home-schoolers used to individualized instruction but interested in coming back into the public-school fold. Ogston plans to be at the capacity of 300 students next year.

To a degree many other blended learning schools do not, Carpe Diem challenges the way people think education should look. Far from a cozy, tweedy classroom like something out of Dead Poets Society, Carpe Diem looks like a call center in Bangalore. Students spend big chunks of their days in cubicles. Because students conduct so much of their basic learning time on the computer, though, Ogston says this frees up time for individual face-to-face interactions on areas where they are stuck and looking for extra information. Students say “they actually get more one-on-one time with teachers here than they did in other schools.”

KIPP Empower Academy

KIPP (“Knowledge Is Power Program”) Public Charter Schools are already famous for their great results. The 125-member network has attracted investment from numerous philanthropies, including the Doris and Donald Fisher Fund, the Karsh Family Foundation, the Walton Family Foundation, and others. Founded in 1994 by Mike Feinberg and David Levin, KIPP schools’ basic tenets are longer school days, high expectations, and character development. Some KIPP schools have been built around small class sizes, which is what Mike Kerr planned to do when he founded KIPP Empower Academy in South Central Los Angeles, a few blocks from where the riots started in 1992.

But shortly before the school opened for kindergartners in fall 2010 (it will be K–4 by 2014), California cut state spending for schools. Some other sources of funding dried up simultaneously, and suddenly Kerr was looking at class sizes of 28–30 children. The question facing the school was how to maintain small group instruction with limited finances. Blended learning promised to help with that. Kerr implemented a blended-learning model to maintain small group time with teachers, even as class sizes rose. Grants from the Riordan Foundation, funded by former Los Angeles Mayor Richard Riordan, helped KIPP purchase the additional technology necessary to implement this model.

Now, at KIPP Empower Academy, located in a few temporary classrooms within the larger Raymond Avenue School—a turquoise and tan complex of buildings, sunny courtyards, fuchsia bougainvilleas, and a very tall fence to keep the neighborhood out—classes of kindergartners rotate through three stations. In M. J. Mathis’ room, nine children do math and reading problems on computers. A dozen children meet with an instructional aide, and half a dozen work with Mathis on specific issues.

The curriculum is challenging; in one kindergarten class, the children are writing stories about the rainforest in their workbooks. Some have just copied words (like “monkey”) from the board, but others have written little stories in full sentences. In a first-grade class, a journal question on the wall asks, “Would you rather read a book or write a book? Why?”

KIPP Empower’s classrooms are nowhere near as technology-centered as Carpe Diem’s (though there have still been challenges with wiring and bandwidth—a perennial problem with opening 21st-century schools in buildings built decades prior). Students spend just 30 minutes on the computers a few times a day. The teachers can group students for in-person instruction based on data coming out of these computer stints, though in the first years, KIPP manager Jelena Dobic reports, the data was basically useless in terms of teachers being able to quickly assess what it meant about comprehension. The school has been changing the software to close this loop, and it received a grant from the Gates Foundation to develop systems that are more user-friendly.

Technology applied in the right ways can transform, rather than reform, education.

Even without a perfect data-feedback loop, though, and with limited computer time, results from its first school year (2010–2011) put KIPP Empower Academy at the top of new KIPP schools—something that got blended learning noticed network-wide. The L.A. institution is “a very young school, and we are always careful not to overstate,” says KIPP CEO Richard Barth, but “when we look at their results both in terms of academic growth and also at other things we value—student retention, staff satisfaction, family satisfaction—they’ve had as good a couple of years as any elementary school we’ve opened.” Indeed, according to KIPP network spokeswoman Zoe Fenson, KIPP Empower Academy had 100 percent staff retention from year one to year two, and one of the highest levels of teacher satisfaction across KIPP’s nationwide network.

The L.A. school’s results cast doubt on conventional wisdom about the supposed importance of class size. If students in 28-child classes can get better results than those in 22-child classes within the same charter network, class size can’t be the most important variable. All KIPP schools have the same long school day, the same policy of allowing school leaders wide autonomy, the same strict discipline standards, and so forth. Blended learning is probably not the only variable helping KIPP Empower Academy. But it seems to be an important contribution.

School leaders in the KIPP network are empowered to design their schools as they wish, so blended learning would never be imposed top-down across the charter-school chain. Richard Barth notes that “the way things spread in KIPP is viral. The key to expanding an idea is that people try something and then it works. I can’t put it any more simply than that.”

Because blended learning seems to be working, KIPP Empower Academy’s results have sparked interest from other school leaders who are starting fresh KIPP schools or revamping old ones. KIPP Chicago had already ventured into blended learning during the 2010–2011 school year with two pilot programs, funded in part by Gates Foundation grants aimed at encouraging numerous blended-learning pilots in existing charter-school chains. The Chicago pilots took place in eighth grade at KIPP Ascend Middle School, and in first grade at KIPP Ascend Primary School. These schools found that a “power hour” modeled on Rocketship’s original Learning Lab setup was the most successful for them. So students go to a computer lab for instruction, and then are pulled out into small groups.

Based on good results at these two pilots, all KIPP Chicago students are being exposed to blended learning in some form during the 2012–2013 school year. A fully blended-learning school called KIPP Create College Prep Middle School opened in fall 2012 to serve 90 fifth graders. It will use both the power-hour model and a rotational model in classrooms throughout the day, and grow to 350 students by fall 2015.

KIPP NYC also opened a blended-learning school, KIPP Washington Heights Middle School, in fall 2012, with a focus on math. Students attend 60-minute math classes, starting with a short “Do Now” exercise, and then split into groups. One group works with the teacher while the others work on math exercises on the computers. The idea is to have multiple mini-lessons per class, tailored to students who need reinforcement on certain concepts.

Other Schools

While the best-known examples of blended-learning schools have been nurtured with philanthropic help, blended learning is a broad and grassroots movement that is surging into schools even when there is no support from donors or education-reform groups. Thurgood Marshall Middle School is located in Los Angeles, about three miles north of KIPP Empower Academy, and is housed in a space that also functions as a Pentecostal church. Principal Peter Watts heard about blended learning, and realized that he’d done a form of blended learning himself: his master’s degree is from the University of Phoenix, the for-profit institution of higher learning that enrolls hundreds of thousands of students each year, mixing computer instruction with some face-to-face components (in what might qualify as an “enriched virtual” model).

Intending to bring blended learning to his school, Watts asked Apple what it would cost to get new laptops for all his students. The price—$300,000—was far outside his budget. Watts had no deep-pocket backers, but he persevered, in the process becoming a leader in showing that blended learning can be implemented on a shoestring budget.

“My teachers and I all sat down and asked: how can we do this blended learning within our school budget?” They found dozens of old computers in the church’s basement. “We inventoried every computer on campus. We were taking keyboards and mice off teachers’ desks.” Watts had the computers refurbished as best he could. He decided to focus solely on math, using Khan Academy because it was free, and Revolution Prep because he was able to purchase it through money allocated for after-school programs and tutoring.

Needless to say, this completely bootstrapped approach to blended learning isn’t perfect. Some of the computers turned out to be lemons, and the shortage of well-functioning machines has been frustrating for Watts and his students. As a supplement, Watts managed to find money for a program to help families purchase computers for their homes, which students use for homework.

But the fascinating part of all this is that even in the absence of reliable computers, blended learning is helping the Thurgood Marshall teachers do their jobs better. They get reasonably useful data on their kids. A handful of seventh-grade students have already moved ahead to algebra after showing they’ve mastered pre-algebra. “In a traditional setting, they would have just stayed where they were with the rest of the class,” says Watts.

Other students are getting more practice on what the school now knows for sure they didn’t learn the first time. In the past, Thurgood Marshall would have put all eighth graders into algebra. After looking at data from their blended-learning program though, Watts said they realized it “would be crazy for us to put them in algebra. They’re not ready.” Instead, these eighth graders started with algebra readiness, with kids moving ahead or not moving ahead as the data indicated.

So what do the students think? “The kids do like it overall,” says Watts. “There are some kids who say ‘I want my textbook back! I don’t want to use the computer that much!’” But for the most part, “they feel like the teacher has access to them more than in the past.” They also like “knowing exactly where they are and not having to wait for a teacher to give them a grade.” Kids see what they need to work on, and “teachers have been embracing the data.” This is “the first year where they’ve not had to explain why a child got the grade they received. Parents know why their child is getting this grade, because they’ve seen the data themselves.”

Thurgood Marshall Middle School has one thing in common with other schools that enjoy more robustly funded blended programs: it is a charter school. Charter schools tend to have operational flexibility that can be crucial to establishing a blended experiment. Yet while charter schools are growing rapidly in number, and while vehicles such as the Charter School Growth Fund exist to help them expand, after 20 years their market share is still in the single digits. The lion’s share of American students continue to attend traditional public schools operated directly by large school districts. So many philanthropists wonder if it is possible to launch blended-learning programs in traditional district schools.

California’s Rogers Family Foundation is attempting to find out. Carrie Douglass, who directed much of the foundation’s attempt to bring blended learning into a conventional public school district, says that “it seemed that most of the blended-learning investment and innovation was going into charter schools. That makes sense in many ways”—after all, many foundations worry with good reason that a large gift given through the front door of a district will simply disappear into general operations—“but we feel like we will miss the boat with the power of this reform if we don’t get into districts early with really thoughtful and comprehensive pilots.”

Because there is evidence it works, blended learning is a grassroots movement surging into schools, even when there is no support from the establishment.

After years of generous, focused philanthropy, the Rogers Foundation has much political capital in the Oakland Unified School District, which has long struggled with a high dropout rate and disappointing test results. So when the foundation “reached out to schools we’ve funded in the past,” says Douglass, and said “we’re interested in making this investment. If you’re interested, let us know,” several district schools responded. In the fall of 2012, four district schools—Elmhurst Community Prep, EnCompass Academy, Korematsu Discovery Academy and Madison Middle School—launched blended-learning pilots.

Rogers and its partners (who are working directly with the schools, and are maintaining financial control of the operation) are investing $1 million to support design, implementation, and follow-up. The Stanford Research Institute is studying the Oakland pilots for results. “Something as important to us as student achievement is that teachers are happier and more effective,” says Douglass. “Teacher turnover is a huge problem in Oakland. Our goal for the first year is that these pilot teachers would go out and tell other teachers that they need to try this.” Testimony like that could make blended learning adoption a viral phenomenon.

Many philanthropists wonder if it is possible to launch blended-learning programs in traditional district schools.

Despite having to work within the constraints of the existing school model in terms of class sizes, bell schedules, and room size—all things they’d like to have the opportunity to rework—the foundation’s pilot has so far been able to deliver on its basic promises. They hope next to increase the academic rigor of the program, and there is a huge need for training so teachers can better evaluate data on their students, and help kids use programs.

The foundation will be extending the program to a second cohort of schools next year. The new group will include some charter schools as well as conventional district-run schools. The Rogers Foundation knows that this experiment is risky, but given how many students continue to receive their education in districts like Oakland, they think it’s also an experiment that needs to be run.

Rhode Island, with philanthropic help, is likewise experimenting with blended learning in district schools. Deborah Gist, Rhode Island’s Education Commissioner, uses multiple devices herself to accomplish her work and maintains a presence in social media. Expanding the use of technology in Rhode Island schools was thus on her agenda.

With support from the Hume and Peters foundations, Gist’s department decided to host a blended-learning conference in February 2012. Roughly 300 educators and other stakeholders from almost every district in the state gathered to hear from national experts on the topic. After the conference, about a dozen schools applied for a mostly taxpayer-funded grant to implement a blended-learning program.

The education department chose Pleasantview Elementary in Providence, which is one of the state’s “persistently lowest-achieving schools in one of the highest poverty neighborhoods,” according to state officials. (The state is also pursuing other grant money to make blended programs a reality at some of the other schools who applied.) The Pleasantview staff went through extensive professional development over the summer, and the school acquired enough equipment to have about a 1-to-1 ratio of students to computers. The grant will be implemented over two years. The school program will be studied closely, and if it produces good results, will be copied elsewhere.

Under the leadership of former schools chancellor Joel Klein, New York City also implemented a blended-learning model in several district schools through a program called the School of One. For now, the program covers just math, and it started as a summer-only pilot before being extended to other schools. Students learn in a large, open space with several stations. Some kids work with teachers; some work online, and some in groups. Each day, the kids take quizzes to determine what they should learn the next day.

An intriguingly data-driven experiment, School of One used an algorithm to create a personalized plan for math instruction for each student, drawing from thousands of lessons pulled from over 50 providers. The chosen software could change daily based on assessment. The Broad Foundation was an early supporter of the program and its creator Joel Rose. “We found it interesting largely because of its use of data to alter the kind of instruction a student gets based on what they need,” says Luis de la Fuente, a director at the Broad Foundation. “It’s not like it was changing the whole set-up of how a student goes to school, but it was beginning to innovate for one subject and pushing the envelope of what school could look like.”

Initial results were mixed. The algorithm didn’t completely account for the human factor in education. “We’ve gotten some criticism from teachers and parents on how that worked—that it was too in the weeds—and that they had not done a good job of creating more critical thinking opportunities for students,” says de la Fuente. Students also seemed to crave “lasting relationship time” with one teacher, even if pure data analysis suggested that was not the optimal way to be spending any unit of time.

Of the three schools that tried it, one did better than peer schools, one did the same and one did worse. Since the program is limited to math, it’s clear that other issues in the schools experimenting with it colored the overall results. Ultimately, two of the schools that tried it dropped the program, though others adopted it.

Lessons learned during the first experiment were used to improve classroom practices. And in the second year, results were clearly positive. The most recent outcomes show School of One students posting twice the gains in proficiency level of kids at middle schools citywide. While the project’s ultimate outcomes are not yet clear, it is being watched closely by donors and others anxious to bring innovation to traditional urban school districts.

“We’ve had the old model of schooling for 170 years—finding a new solution won’t happen in one fell swoop,” says Joel Rose.

Founder Joel Rose makes it clear that his team is still making discoveries and improving rapidly through trial and error. “We’ve been at this three years, and while we’re learning things every day about things like the logarithm and schedule, we’re still only 50 percent baked.” For instance, the process of evaluating the day’s work by students and creating their work schedule for the following day initially took 10 hours, but over the course of three years the process has been automated and now requires roughly five minutes. “We’ve had the old model of schooling for 170 years—finding a new solution won’t happen in one fell swoop,” he notes.

Rose’s spinoff organization, New Classrooms, is continually modifying the program, and has expanded its “Teach to One” model more broadly into eight schools in three different cities during the 2012–2013 school year. The CityBridge Foundation, for instance, helped bring the program to Washington, D.C., in fall 2012.

As the School of One experiment shows, innovation is sometimes messy. Some new ideas don’t work in quite the way you think they will. Openness to experimentation is needed. One of the big reasons education reformers have generally worked with charter schools is that trying something new in an existing bureaucracy is hard. Existing organizations, especially large ones, are seldom able to adopt changes that upend their whole business model, notes Clayton Christensen in his analysis of disruptive innovations. Settled life is just too comfortable to take new chances.

Charter schools are often new schools, and districts don’t open many new schools. But sometimes they do. In particular, districts can, these days, tap federal funds to close failing schools and re-open them as turnaround schools.

Sajan George worked in district turnaround efforts during his years at business-advisory firm Alvarez & Marsal, and after leaving, he founded Matchbook Learning, a nonprofit that implements blended learning at turnaround schools. With funding from the NewSchools Venture Fund and others, Matchbook Learning is attempting to turn around two Detroit public schools. The A. L. Holmes school started a blended-learning model during the 2011–2012 school year, and Brenda Scott Academy reopened with blended learning in the fall of 2012.

These Detroit schools face big challenges. Turnaround schools tend to be in the bottom 5 percent of performance, and many students are incredibly far behind, testing at the first- or second-grade level in seventh grade, George reports. But “they’re going from zero to something,” he says. At A. L. Holmes, eight out of ten students gained at least 10 percent on annual standardized tests. Of those, three out of ten made gains of more than 30 percent.

As of this book’s writing, Brenda Scott Academy was only a few months into its experiment, but families are voting with their feet. With bad publicity as a turnaround school, enrollment dropped from 832 in spring 2012 to 650 in fall 2012. But by December, enrollment was back up to 932, George reports. “We have kids showing up every day enrolling,” he says. Families “really see the value” because “you don’t have to be middle-income or highly educated to understand that your kids need to embrace technology in learning if they’re going to be competitive for jobs.”

The desire to help children compete in a high-tech world played a part in the decision of Mooresville, North Carolina, to adopt blended learning district wide. The district repurposed existing textbook and technology funding (with no net increase in costs) to adopt a 1-to-1 ratio of laptops to students. Students do much of their work online, with programs set at the level that students need. Constant data from the programs keep teachers and parents better informed, and help keep students from falling through the cracks. Tech facilitators at each school mentor teachers and work with them to figure out how to teach in this new way. It seems to be working. While the district has one of the lowest per-pupil funding budgets in the state, its test scores have risen over the past few years from middle of the pack to near the top—evidence that innovative districts, like charter schools, can do more with less. There may be a role for local donors in encouraging traditional school districts in their area to study the Mooresville experience.

First Lady of Digital Education

No one has done more to take blended learning from obscure concept to national movement than Gisèle Huff, the executive director (and lone employee) of the Jaquelin Hume Foundation, a small San Francisco–based philanthropy whose grants have had outsized impact.

“Education has always been a very important part of my life,” Huff says. She and her mother came to the U.S. from France after World War II with “$400 to our names, the clothes on our backs and in our suitcases. We didn’t speak any English.” The American education system made it possible for her to earn a Ph.D., have a business career, teach, and even run for Congress. She wants other children to have those same opportunities. That’s why her first years at the Hume Foundation were spent advancing school choice. But as she points out, “we were in the business of school choice for 10 years and we had bloody foreheads to show for it.” She came to the idea of digital learning after hearing Clayton Christensen speak in 2005. She loved the idea that technology could transform, not reform, education.

The Hume Foundation’s first digital-learning grant was $50,000 to the Innosight Institute, in part to support the writing and promotion of Disrupting Class. “I saw that as a seminal book and a seminal idea, and that’s what it turned out to be,” Huff says. To advance thinking about digital education, she also gave an early grant to iNACOL.

As educational entrepreneurs began starting blended learning schools, the Hume Foundation made grants for public relations tasks: $100,000 to Rocketship, $100,000 to Carpe Diem. These grants paid for videos that helped people see what blended learning looks like, and were successful to the point that Carpe Diem founder Rick Ogston was soon a major figure in a one-hour Juan Williams television special. (When Huff met Ogston and heard what he was up to, she reports that she told him “hold on to your hat—people are going to know about you.”) Rocketship has been profiled in USA Today and other publications. “It’s unbelievable what $100,000 did,” says Huff. Her foundation has also paid for journalist-education days that introduce influential writers to the concept of blended learning.

Hume Foundation money, along with grants from the Harry Singer Foundation and some individual donors, also established The Learning Accelerator. This group hopes to organize the investment of $100 million of philanthropic and commercial money to help blended-learning service providers set up operations in school districts across the country. In 2013 the group received its first substantial infusion of capital, a $750,000 grant from the Gates Foundation.

As money from other foundations has followed Hume’s giving, Huff now reports that “I’m out of charter schools. Now we’re doing districts.” Painting on this larger canvas is the next way the Hume Foundation’s risk-taking strategy could lead other funders. But even as the kinds of schools the foundation supports have changed, the organization is all-in on digital learning. “We gave up everything else,” says Huff. Arguments about teacher evaluations and merit pay and school choice remind her of reports that cities commissioned around the turn of the 20th century on what to do about the burgeoning problem of horse excrement soiling the streets. With rising populations, everyone assumed it would get worse. “A friend tells me that one proposed solution was to put diapers on the horses,” says Huff, laughing. “Meanwhile, Henry Ford was just around the corner. This is exactly what’s happening now. The reformers are putting diapers on horses. Yet digital learning is just around the corner.”

Changing an Obsolete System

Like Gisèle Huff and the Hume Foundation, Frank Baxter had scars from years of trying to improve education. “I’ve been involved in school reform since 1986,” says Baxter. His career also includes a 13-year period as CEO of investment bank Jefferies & Company, and a stint as President George W. Bush’s Ambassador to Uruguay.

In addition to acting as a donor himself, Baxter became co-chairman of the board of the Alliance College-ready Public Schools, a high-performing charter network launched in 2004 that expanded rapidly and graduated virtually all its students. Yet despite those successes, “it occurred to me that what we were doing was making the best of an obsolete system.” The campaigners who created the American public school system in the mid-1800s were “trying to prepare our pastoral, illiterate nation for the Industrial Revolution,” says Baxter. Their invention “functioned pretty well up to the middle of the 20th century.” But the world has changed since then.

The economy now values independent thought, not uniformity. Meanwhile, as the world demanded more of teachers, quality declined, since our old schools were built “on the availability of qualified women who had to work at a fraction of what they can get elsewhere today.” Reform efforts produced all kinds of responses. “Many of the ideas were really good from a management standpoint: more accountability, more training, more pay. That’s business 101. But nothing happened,” says Baxter.

People blamed just about everyone: “The teachers are no good. The kids are no good. The parents are no good.” But what, asks Baxter, if most people are doing their best, but “there’s something wrong with the model?”

That question, plus Baxter’s long-standing interest in technology (“I always think, ‘never send a person to do a machine’s job’”) led him to support blended-learning pilots in the Alliance schools, like the one described in this guidebook’s profile of the Alliance Technology and Math-science High School (ATAMS). The fact that they were running charters, not traditional public schools, allowed for the possibility of innovation. “We’ve had the luxury of being more outcome-oriented,” he says.

As his teams now study the outcomes, “I’m totally convinced that it works. There’s still not enough data to really verify that.” But students and teachers are excited about learning in a way he hasn’t seen before. “It just clearly is a better system,” he says. “It’s scalable, it’s rapidly improving, and it’s sustainable.” Recently, Baxter has been having conversations with the Los Angeles Unified School District (LAUSD) on expanding blended-learning more broadly across the city. “I was worried, because there’d be a threat of reducing the number of teachers,” he says. But he learned that due to enrollment and budget problems the LAUSD already had over 40 students in the classroom in some places, without the balancing advantages of blended-learning stations. In those stressed schools, “blended learning is going to be very attractive to them.”

First Lady of Digital Education

No one has done more to take blended learning from obscure concept to national movement than Gisèle Huff, the executive director (and lone employee) of the Jaquelin Hume Foundation, a small San Francisco–based philanthropy whose grants have had outsized impact.

“Education has always been a very important part of my life,” Huff says. She and her mother came to the U.S. from France after World War II with “$400 to our names, the clothes on our backs and in our suitcases. We didn’t speak any English.” The American education system made it possible for her to earn a Ph.D., have a business career, teach, and even run for Congress. She wants other children to have those same opportunities. That’s why her first years at the Hume Foundation were spent advancing school choice. But as she points out, “we were in the business of school choice for 10 years and we had bloody foreheads to show for it.” She came to the idea of digital learning after hearing Clayton Christensen speak in 2005. She loved the idea that technology could transform, not reform, education.

The Hume Foundation’s first digital-learning grant was $50,000 to the Innosight Institute, in part to support the writing and promotion of Disrupting Class. “I saw that as a seminal book and a seminal idea, and that’s what it turned out to be,” Huff says. To advance thinking about digital education, she also gave an early grant to iNACOL.

As educational entrepreneurs began starting blended learning schools, the Hume Foundation made grants for public relations tasks: $100,000 to Rocketship, $100,000 to Carpe Diem. These grants paid for videos that helped people see what blended learning looks like, and were successful to the point that Carpe Diem founder Rick Ogston was soon a major figure in a one-hour Juan Williams television special. (When Huff met Ogston and heard what he was up to, she reports that she told him “hold on to your hat—people are going to know about you.”) Rocketship has been profiled in USA Today and other publications. “It’s unbelievable what $100,000 did,” says Huff. Her foundation has also paid for journalist-education days that introduce influential writers to the concept of blended learning.

Hume Foundation money, along with grants from the Harry Singer Foundation and some individual donors, also established The Learning Accelerator. This group hopes to organize the investment of $100 million of philanthropic and commercial money to help blended-learning service providers set up operations in school districts across the country. In 2013 the group received its first substantial infusion of capital, a $750,000 grant from the Gates Foundation.

As money from other foundations has followed Hume’s giving, Huff now reports that “I’m out of charter schools. Now we’re doing districts.” Painting on this larger canvas is the next way the Hume Foundation’s risk-taking strategy could lead other funders. But even as the kinds of schools the foundation supports have changed, the organization is all-in on digital learning. “We gave up everything else,” says Huff. Arguments about teacher evaluations and merit pay and school choice remind her of reports that cities commissioned around the turn of the 20th century on what to do about the burgeoning problem of horse excrement soiling the streets. With rising populations, everyone assumed it would get worse. “A friend tells me that one proposed solution was to put diapers on the horses,” says Huff, laughing. “Meanwhile, Henry Ford was just around the corner. This is exactly what’s happening now. The reformers are putting diapers on horses. Yet digital learning is just around the corner.”

Changing an Obsolete System

Like Gisèle Huff and the Hume Foundation, Frank Baxter had scars from years of trying to improve education. “I’ve been involved in school reform since 1986,” says Baxter. His career also includes a 13-year period as CEO of investment bank Jefferies & Company, and a stint as President George W. Bush’s Ambassador to Uruguay.

In addition to acting as a donor himself, Baxter became co-chairman of the board of the Alliance College-ready Public Schools, a high-performing charter network launched in 2004 that expanded rapidly and graduated virtually all its students. Yet despite those successes, “it occurred to me that what we were doing was making the best of an obsolete system.” The campaigners who created the American public school system in the mid-1800s were “trying to prepare our pastoral, illiterate nation for the Industrial Revolution,” says Baxter. Their invention “functioned pretty well up to the middle of the 20th century.” But the world has changed since then.

The economy now values independent thought, not uniformity. Meanwhile, as the world demanded more of teachers, quality declined, since our old schools were built “on the availability of qualified women who had to work at a fraction of what they can get elsewhere today.” Reform efforts produced all kinds of responses. “Many of the ideas were really good from a management standpoint: more accountability, more training, more pay. That’s business 101. But nothing happened,” says Baxter.

People blamed just about everyone: “The teachers are no good. The kids are no good. The parents are no good.” But what, asks Baxter, if most people are doing their best, but “there’s something wrong with the model?”

That question, plus Baxter’s long-standing interest in technology (“I always think, ‘never send a person to do a machine’s job’”) led him to support blended-learning pilots in the Alliance schools, like the one described in this guidebook’s profile of the Alliance Technology and Math-science High School (ATAMS). The fact that they were running charters, not traditional public schools, allowed for the possibility of innovation. “We’ve had the luxury of being more outcome-oriented,” he says.

As his teams now study the outcomes, “I’m totally convinced that it works. There’s still not enough data to really verify that.” But students and teachers are excited about learning in a way he hasn’t seen before. “It just clearly is a better system,” he says. “It’s scalable, it’s rapidly improving, and it’s sustainable.”

Recently, Baxter has been having conversations with the Los Angeles Unified School District (LAUSD) on expanding blended-learning more broadly across the city. “I was worried, because there’d be a threat of reducing the number of teachers,” he says. But he learned that due to enrollment and budget problems the LAUSD already had over 40 students in the classroom in some places, without the balancing advantages of blended-learning stations. In those stressed schools, “blended learning is going to be very attractive to them.”

Sal Khan and Serendipitous Philanthropy

Sal Khan famously began filming his narrated math videos in his closet in 2004. By 2009 he and a handful of educators had launched programs using Khan Academy in a few schools. Several thousand students were watching the videos and working through his problem sets every day. “I was spending every ounce of my free time to work on it,” he writes in his 2012 book, The One World Schoolhouse. “Actually I was even spending a little of my nonfree time,” he says, and so in late summer 2009 he decided to quit his job at a hedge fund and give himself a year to make Khan Academy work.

A few months into the experiment, he was stressed out and burning through $5,000 a month to support his family. But in April, he received an unsolicited $10,000 donation from Ann Doerr, wife of the famous venture capitalist John Doerr. A friend’s child had been helped by Khan Academy videos and she was intrigued. She and Khan met for lunch; she later gave him $100,000 to stay afloat. Two months after that meeting, she sent him a text message that Bill Gates was mentioning him in a speech at the Aspen Ideas Festival. Gates turned out to be using Khan Academy videos to brush up on his own math skills and to help his children.

And so a year after quitting his hedge fund job, Sal Khan was in Gates’ offices in Kirkland, Washington, answering questions. The Gates Foundation soon invested $1.5 million in Khan Academy for operations and then later gave another $4 million for other projects. Google awarded $2 million to build out the problem sets and translate content into ten languages. Khan Academy is now being used in numerous schools, including the high-performing Los Altos School District, and in the Summit Public Schools profiled in this guidebook.

While philanthropists often devise a strategy and then look to fund people working in that area, the funding of the Khan Academy has been more about the serendipity of stumbling upon an excellent service or product and then investing in what you have grown to admire. “The Gates Foundation didn’t come at it from a traditional angle—some thesis that we want to be sponsoring open education resources and find a scalable model,” says Khan in an interview. “They kind of just bumped into me. They were using it. Bill Gates was using Khan Academy with his kids. A lot of folks at Google were using it for their kids and themselves.” These funders realized, “I’ve gotten value out of this. It would be a shame if other people wouldn’t get it because this guy can’t pay his bills.”

Khan thinks of this as more like a venture-capital mindset: investing in something you find impressive. He argues that philanthropists should do more to encourage entrepreneurship in the nonprofit sector. There are lots of people willing to launch experiments for very little money in order to have a shot at the kind of validation that comes from creating something new and effective and beloved. That kind of energy can be harnessed in the nonprofit sector just as it is in for-profit work, Khan believes.

To be sure, for-profit start-ups hope for big payoffs down the road when they go public or get acquired. Working for a nonprofit start-up is “not as much of a lottery ticket,” Khan says. So “you’d probably have to pay a little better.” Thanks to philanthropy money, at Khan Academy, people make “upper quartile Silicon Valley pay,” and that enables Khan to answer in the affirmative the question: “Are we getting the very best talent, not a subset of talent that’s just willing to sacrifice a lot?”

Yet money isn’t everything. “I’m meeting a lot of entrepreneurs who are looking to do the next cool thing,” says Khan. “They’re looking to have the next big impact. They care about money because it gives you certain abilities to do things. But they’re not in it to drive a Ferrari one day. Their pride is ‘I changed an industry.’”

So what would have happened to Khan Academy if Doerr, Gates, and Google hadn’t come along? What was his back-up business model? “Advertising,” he says. Based on page views in 2009–2010, Khan Academy could have pulled in at least $70,000–$80,000 per year in ads. “That definitely would have allowed me to survive, but it would have been unfortunate,” says Khan. “It would have become very commercial very fast.” Picture 15-second ads for Coca-Cola before each trigonometry video: “That’s what I would have had to do to keep making Khan Academy.”

dowload link source