Over the next week or so, The Hechinger Report will be publishing excerpts from Richard Whitmire’s book On the Rocketship: How Top Charter Schools are Pushing the Envelope. In this second installment, software engineer Ben Slivka finally ships his educational software but finds his real customer base isn’t what he expected.
CHAPTER 20: Slivka’s Sci-Fi Software Gets Shipped
February 2009: Bellevue, Washington
When Ben Slivka decided to create his own interactive learning software in 2006, the online options for students, parents, and teachers were pretty bleak. For decades, the big textbook publishers had offered up catalog after catalog of new and updated textbooks that were plush and expensive, something akin to Microsoft’s computer-based software. Nobody wants to walk away from cash cows like that. And so, for the same reason Microsoft delayed shifting to the web, the textbook publishers delayed going into online education. What made more sense to them was duplicating their textbooks on CDs or placing them online, thus making them digital cash cows.
But digital learning doesn’t work that way. If it’s not interactive and personalized for each student it might as well be a frayed old textbook. The few independently developed digital math programs were linear, asking students to proceed from one step to the next. And if you run into trouble? Raise your hand and ask the teacher for help. What was the point of that? Might as well return everyone to classroom lectures.
Slivka knew he had an opportunity to create something different, mostly because he and his team weren’t educators. They were software writers, and good ones. Slivka says software engineers are not like journalists. A really good journalist might be twice as good, or twice as productive, as a lesser journalist but a software writer can be exponentially better. A talented software engineer on a roll can do in one night what a team of 150 engineers fail to do over the previous week. The reason DreamBox did what other online programs couldn’t do, says Slivka, is simple: “We had better software engineers.”
Slivka’s team created a platform, essentially an engine with all the nuts and bolts hidden, with a programming language resting on top, a language that actual educators could use to write interactive lessons. The first teacher Slivka hired was Mickelle Weary, who had taught his son at a Seattle private school seven years earlier. “I must have made an impression on him [Slivka],” said Weary, who was a first-year teacher. “You know, first-year teachers often aren’t great, but I tried really hard. I was trying to do individual learning, letting [Slivka's son] and another really smart kid to do things together.” Succeeding at personalized learning made an impression on Slivka.
The best way to experience DreamBox is to sign on as a prospective parent and take a quick trip through the program. To entice the kids, the math programs might be constructed as treasure hunts displayed as elaborate maps. Students who successfully solve logistical problems along that pathway get a little closer to the treasure. The real treasure, of course, is the hidden mapping that sucks up every move made by your child and then uses that information to build a complete package of math skills. Personalized learning.
“Kids are not books, where one page follows the next,” said Dan Kerns, DreamBox’s chief architect. “They are all over the map. You might have one kid at the ‘normal’ and everyone else at different parts of the bell curve, with the hard-to-reach at one end and the bored at the other. Part of engagement and teaching is finding out what kids need to learn. We have deeply integrated assessment into learning. We’re constantly asking, ‘Does the kid really know the stuff we’re teaching?’ If the machine doubts that, it loops back and gives a refresher. … At the end of the day in a DreamBox classroom all the kids are doing something different. That’s because it has measured and adapted.”
And so, in February 2009 DreamBox was launched. On its own merits, DreamBox may have been as innovative as Dropbox, but the reception was vastly different. Whereas Dropbox soared, DreamBox sputtered. DreamBox was aimed at parents of young children, but how many of those parents were even aware their children needed a program such as DreamBox? That kind of awareness doesn’t even develop until middle school when kids start bringing home lousy math grades.
As it turned out, Slivka had created a software model that was light years ahead of its business model. Few parents were interested in buying a product they didn’t know they needed. Kids weren’t finding it on their own. One interesting development surfaced early, however. The DreamBox team noticed that “parents” listing their family size at between twenty and thirty children were buying DreamBox. As it turned out, these were teachers buying it for their classrooms. “They raved about it,” said Kerns. “They told us it solved real problems they had.” Soon, DreamBox reoriented itself to classroom sales but that wasn’t an easy switch. How many tech-naive school superintendents are going to take a chance on a bunch of startup guys unknown in the education world? It was great that teachers liked DreamBox, but that didn’t necessarily translate into a workable business model. Companies like DreamBox need lots of cash to keep innovating, but the company wasn’t generating that cash and investors were getting hard to find. It wasn’t at all clear that DreamBox, despite being one of the first truly adaptive learning programs to emerge, would make it.