Diane Tavenner ( @summitps) talks about “Running Short Experiments During a Long Product Cycle” with the Summit Public Schools.

One of the most common exuses we hear: it’s too hard! we have long process cycles! people are resistant to change! so this is doing it in public education O_O

i suspect if we were playing word association, public schools do not come to mind when we say Lean Startup. Public schools are designed to hold the status quo: guaranteed revenue and customers, even when we are losing customers.
Charter schools are public schools with the express intent of innovating.
Charter schools oeprate under extreme uncertainty.
10 years ago, I launched our first school. at the time, following reall well-established facts:
only 1 quarter of them are able to attend or apply to college. #1 reason: they don’t complete the required coursework. This can be directly atributed to the quality of teacher.
Our goal was to create a school where every student enrolled completed rigourous set of courses that made them eligible for college. Every student had a high-performing teacher every period, every day.

Mission: create a public high school that gets 100% of students accepted to a four year college (not 24%, our national average).
4 years later: 98% accepted. 8 times more likely to
named by us news, 7 applicants for every seat, raised a lot of money.
4 years after that, less than half of those students actually graduated from college.
nothing bad actually happened: even with only half graduating, still doubling the national average. still among the highest.
However, it still took us 8 years that our product only delivered half the value we expected, unacceptable. also, 8 years too long — this is these kids lives. if we could teach them a little more a little better, we could ge tthere.
So, we designed a new math program. Kahn academy, technology, after 10 months it’s working: they were learning more math than their peers. The NYT wrote an article, we raised more money — but we were NOT SURE which element led to kids learning more math.
- kahn academy?
- was it that we assessed them 4 times per year?
- simply that they were doing more math?
Then we discovered lean startup! We realized that we didn’t hae a discipilined and principled approach to innovation that let us build, measure, learn effectively.

We started with a relatively small batch of students, for 1/3 of school day, for 1 subject.
Hypothesis: if students can learn to drive their own learning (determine own path and pace to goals they set themselves), they will be succseful.
We had just 3 months: pushing ship date back is not an option.
One rule: remove any feature, process or effor that does not contrubte directly to learning.
We started removing walls between classrooms. In order to test that “when kids that direct their own learning, kids learn more,” then we have to change schools fundamentally. In old way: teachers direct everything.
To test the hypothesis, we needed to have the kids do everything, teachers are there to provide resources.
We built an online math guide showing everything a student needs to know from kindergarten up, using red, green, and yellow. We created online playlists to go with concepts on guide: Kahn academy, in person, online, on-demnad testing system: show exactly what they learned and get immediate feedback. If they can show they learned a concept, it turned green.

We created a unique space with big tables, open worskpaces, looks more llike a high tech startup than a typical school. We gave every single student a computer, showed them tools, said go. We commited to empirically figuring out what works.

1. did kids learn math? could they show they learning
2. to what degree did they drive their own learning, make decisions and engage in behaviors that led to learning?

These seem obvious and seem like what every teacher does on a daily basis (measuring what kids are learning)? Sadly, everyone is still using vanity metrics. They are not measuring where each student begins and ends in a short learning cycle. They’re not looking hard enough. Is it specifically because of what the teacher did, or because of other reasons (parent, self, other)?

We create in-depth focus groups. We looked for observable behavior data. Teenagers embrace this, she gets emails with the subject “user feedback” from high school students. A cross functional team meets weekly. We have made more progress in 14 weekly cycels than we have in 10 year long cycles. Focused on learning something the kid didn’t know when the day started.

They started with 20-30 minute introduction by teacher on the subject in breakout rooms.
We found that students who attended the lectures were no more or less likely to learn.
Teachers predicted: if we knew which students would coming, if students were more prepared and brought questions. This went on for several cycles without aything changing.
but student reported “lectures aren’t timely. they’re not customized: not exactly what I need to learn. They are not peronsalized — I don’t get help with what I need to learn.” Attendance lowered, and when the groups got to be 2-3 students effectiveness increased. They noticed that teachers were throwing out their prepared lecture — teachers tutored instead.
The tutoring bar was born. like the genius bar at the Apple Store.

We continue to experiment to see what the biggest impact on learning.
We stopped giving lectures. If not for the rigourous path, we’d still be spending an extraordinary amount of teacher time on lectures and wasting our most valuable resource.
This decision cuts directly to the core of what the teacher things their job is and where they’d add value. It would have been impossible to change without clear and indisputiable evidence. This is how we got them to move to different behaviors quickly and without resistance.

Twitter and facebook or visit, we’re here in the bay area.

Read more: Lean Startup Conference 2012