Get to know
Jose beyond the technology and meet the person I’m dubbing the “Next Bill Gates.” Grab a cup of coffee and learn about the next frontier in education.
Listen to the Podcast or Download as MP3
DR. ROD: Talk with us about the big announcement coming from Knewton .
JOSE: We recently announced a really substantial partnership with Houghton Mifflin Harcourt to power their k12 products starting with English, Math, and English as a Second Language (ESL). Ultimately, it’ll put Knewton in every school in the country.
DR. ROD: Give me the elevator pitch on Knewton , and the role you play in education.
JOSE: Our belief is that within a few years education materials will be, largely if not entirely distributed over devices. In 10 years, it’s going to be tablets and digital devices. With that, you can data mine, you can gather data, and you can personalize education. By gather data, I mean for making assessments as to what kids know and how they learn best. That’s what Knewton does. We can figure out exactly what you’re struggling with, down to the percentile and how proficient you are with each subject no matter how granular it is. Because we’re gathering so much data, we know exactly what kids know, and we know exactly how they learn it best. We can take the entire data base of every kid who’s ever learnedthrough us and figure out who’s really similar to this kid in terms of learning style, what they know and how they learn best.
We can take the combined data power of tens of millions of people, and use it to find the optimal strategy for you. Within a few years, everybody’s going to be building personalized applications. Teachers right up to publishers and everybody in between and that’s as it should be. We’re building an infrastructure that’s so big that no one will build it for any application. With us, everyone can just plug into that matrix to make every app as powerful as possible. We don’t market any of the data. We don’t even hold your PII. We anonymize the data and we can use the combined data power to drive analytics for teachers and learning for students.
DR. ROD: Where did this start? Take me back to the beginning.
JOSE: It was obvious to me that personalized learning was going to be the future. Basic building blocks. No one was going to carry a bunch of textbooks. All of the data that students produce right now just floats away into the ether. With digital learning it can suddenly be captured. With that data, one can personalize. Therefore, it was obvious to me that we’re going to move into a world where all education content is on a device and it would all be personalized. I wanted to build a platform that anyone could tap into, that did all the hard work. We passively norm every item that you give us, right up to a major test makers standards. So, we can extract incredible amounts of data. It’s not guesswork either, it’s hard science. We know exactly what those kids know and we know exactly how they learn it best. So it occurred to me that if, like in other internet industries where data loves to be around other data, that there’s a “the whole is greater than the sum of the parts” effect. Somebody could set themselves up like Switzerland and say “we’ll work with everybody, we’ll never sell to anybody, we’ll always be independent, everybody can trust us with their data, we won’t hold PII , we’re never going to market the data, and we’re just going to use it to drive the next recommendation.” That’s what we did.
DR. ROD: Typically I find there is a great deal of fear in education about data. Who are the detractors to Knewton and what might their greatest fear be, even if it’s unfounded?
JOSE: We don’t have a lot of detractors at the moment because, the functionality we promise, and provide, is so dazzling that people give us the benefit of the doubt. We’re extremely cognizant that we hold a data set that is the most important data set in ones’ life, other than maybe your healthcare data. It’s on us to never give anybody a reason to worry about us holding it. It’s not just student’s data either. We hold publishers content, efficacy data. We know which publisher’s content works the best. We hold their IP data. So we have to be a bank for everybody. Publishers have to know that we’ll guard their IP for them. Students have to know that we don’t think of it as our data. We think of it as their data. We anonymize it and we’re custodians of it.
DR. ROD: How wide is the gap for professionals in understanding the data that you’re providing them? Also, what steps are you taking to educate the market?
JOSE: That’s an extremely good question, both parts of it becausewe spend most of our time just educating the market. Basically, nobody understands this stuff. Let me just try to explain it once, real fast.
There are basically 4 types of data in education that matter. People muddle these hopelessly and they don’t understand what’s what. Here they are.
Basic User Information: Time spent on a page, pages with highest bounce rate, time on task, click rate, etc.
Content Efficacy: How well does a piece of content work? That’s pretty difficult for most people to get because you need normed items to be able to measure how well a student’s proficiencies subsequently gained or didn’t gain.
Macro Data: Attendance data, grade book data, stuff that LMS’s can get, but you need a big macro platform in order to be able to get it. That’s what companies like InBloom or Wireless Generation (now called Amplify) focus on.
Student Proficiency Data: I don’t mean guesswork, I mean actual proficiency data. This is by far the hardest to get, because you need to build a giant taxonomy, an infrastructure that will ingest, passively normed vast quantities of content, and then tag all assessment content and all instructional content down to the common concept. Then you can measure how well the instructional content works at teaching proficiencies and so forth.
(Explanation of the types of Big Data in Education blog post by Jose here )
Those are the 4 big buckets, and they’re kind of in order of easiest-hardest to get. That was the big infrastructure that I thought somebody’s got to build this once and then everybody can share with it. That’s what Knewton’s trying to do.
DR. ROD: Is it hard to distill down because they only have access to certain components and they’ve only seen it in certain lights?
JOSE: Yes, and then people just blend them all together. They muddle them all or they think, “Hey, we’re getting number one, so we’re getting a lot of data.” But, you’re not really getting a lot of data. Not compared to what you could be getting.
Interview transcript continued on the next page…