Harvard and YC on defining learning success

“When  I started teaching, I never asked myself, ‘how am I going to teach?’  Which is strange, because that should be the first question you ask  yourself when you do something new.” — Eric Mazur, Area Dean for Applied  Physics at Harvard University

Harvard Physics

Premeds  are not very kind to physicists. Eric Mazur knew that, but took the  challenge to teach them when he started teaching Physics at Harvard. He  developed an engaging lecture style, complete with death-defying  demonstrations, and before long was one of the university’s top-rated  teachers.

“I started to believe I was the world’s best physics teacher. It was an illusion that lasted many years.”

Then  he came across a study that attempted to distinguish between  memorization and understanding. It employed a simple test, the same  real-world physics problems were asked before a physics course and  after. If learning had happened, it was expected that there would be  more correct answers afterwards than before.

The  study had taken place across various universities in the US, and the  results showed little improvement in students’ answer. No learning was  actually happening.

“I thought, ‘Not my students.’ But I’m a scientist. If you make a statement, you better show the data.”

So Eric tried the test on his students.

And when he did, a hand shot up.

“How should I answer this? According to what you taught me, or how I usually think about these things?”

Despite  Eric’s positive rating and high-performing students, his teaching  wasn’t crossing the gap between memorization and understanding. His  students weren’t progressing from knowledge to action.

Y Combinator

Startups  are one of the best places to observe new knowledge being put to use.  On top of figuring out a new product and how to sell it to new market,  startup founders face the pressure of growing quickly and the threat of  bankruptcy always around the corner.

Startup  accelerators are support programs that generally support 10–15 startups  with a goal of reaching an investment that would otherwise take a year  or two. Startup accelerators aim to get them there in 3 months instead.

Y  Combinator is the best performing startup accelerator in the world.  Known as just YC in the startup community, it’s famous alumni are  Dropbox, AirBnB and host of other household names.

YC  set the model for the proliferation of thousands of copycat startups  accelerators around the world. But it outperforms all of them. Ten years  after it started with $200k USD, it’s portfolio was valued at over $30  billion. And it’s portfolio valuation maintains a roughly tenfold  advantage over its nearest competitors.

Every  accelerator director is aware of Y Combinator, and emulates it. But if  you ask them why their program is 3 months long, or why 3 months is the  default in the industry, they probably won’t know.

Paul Graham, one of the founders of YC explains:

“We  wanted to learn how to be angel investors, and a summer program for  undergrads seemed the fastest way to do it. No one takes summer jobs  that seriously. The opportunity cost for a bunch of undergrads to spend a  summer working on startups was low enough that we wouldn’t feel guilty  encouraging them to do it. “

YC  started with ts founders’ intention to learn how to be angel investors  over a summer. The reason the program is 3 months long is that’s was the  length of the summer break at university.

Most  accelerators inadvertantly copy this and many other aspects of YC  without understanding why, or challenging whether or not this design  decision applies to their context or goals.

The proliferation of copycat accelerators starts around 2007, when the Techstars accelerator was founded. Techstars modeled itself on YC,  but instead of YC’s fortnightly mentoring conversations with the fund  partners and a weekly dinner talks, Techstars introduced a more in-depth  workshop curriculum and “mentoring days” full of 30-minute round-robin  talks with volunteer business advisors.

While  YC eventually focused on 2 programs per year in Silicon Valley,  Techstars propagated across US cities, and opened it’s playbook. So the  Techstars model propagated around the world as the default accelerator  model.

Most  accelerator directors aren’t aware of these origins, and execute the  default schedule of heavy workshops and lectures in their programs.

But at YC, the startup founders are far more in control of their time. They’re expected to attend the weekly dinner talk, but that serves less as a business school lecture and more as place for founders to share progress with each other.

“We  encourage founders to treat each dinner as a mini Demo Day and to show  each other and us what they’ve built that week. We’ve found these weekly  deadlines tend to push people to finish things in order to show them  off.”
A YC speaker dinner, 2009.

It’s  more of a competitive atmosphere, where founders feel peer pressure to  show significant progress every week. The backdrop of that is the talk  from an industry heavyweight, providing positive inspiration, some  insider insight, and a contrast to their early-stage progress.

At  YC, most of the critical learning happens in response to the  fortnightly conversation with a fund partner. They help assess progress,  and respond to the founders current challenges with introductions to  people who can address them with their own experience.

Office Hours with Paul Graham at Y Combinator. Photo: Garry Tan
“Whatever  stage a startup is at when they arrive, our goal is to help them to be  in dramatically better shape 3 months later. For most startups, better  shape translates into two things: to have a better product with more  users, and to have more options for raising money.
You  can’t make people something they’re not, but the right conditions can  bring out the best in them. And since most people have way more  potential than they realize, they’re often surprised what they’re  capable of.”

Success is defined by what the learner becomes

Harvard  Physics is a traditional, lecture-based university experience, where  the students are locked to a preset curriculum. Y Combinator is a  modern, practical educational experience, but the content is driven  primarily by the learner.

In both cases, they succeed on a much deeper level than enabling the learner to regurgitate information.

Success is defined by the evolution of the learner, and measurable by their change in behaviour.

Salim Virani

Salim Virani

Founder of Source Institute. I've developed educational programs for some of the world's top-tier accelerators and business schools, like Seedcamp, Oxford, UCL and the Royal Academy Of Engineering.