The Government Starts an Incubator: The National Science Foundation Innovation Corps

Over the last two months, the U.S. government has been running one of the most audacious experiments in entrepreneurship since World War II. They launched an incubator for the top scientists and engineers in the U.S. This week we saw the results.
This post was published on the now-closed HuffPost Contributor platform. Contributors control their own work and posted freely to our site. If you need to flag this entry as abusive, send us an email.

Over the last two months, the U.S. government has been running one of the most audacious experiments in entrepreneurship since World War II.

They launched an incubator for the top scientists and engineers in the U.S. This week we saw the results.

63 scientists and engineers in 21 teams made 2,000 customer calls in eight weeks, turning laboratory ideas into formidable startups. Nineteen of the 21 teams are moving forward in commercializing their technology.

It was an extraordinary effort.


Your Country Needs You

In July I got a call from Errol Arkilic, a program manager at the National Science Foundation (NSF) -- the $6.8-billion U.S. government agency that supports research in all the non-medical fields of science and engineering. "We've been reading your blog about your Lean Launchpad class." Wow, that's nice, I thought, a call from a fan. No, the conversation was about to get more interesting.

"Our country needs you." Say what?

"Part of the NSF charter is to commercialize the best of the science and engineering research we fund. We want to make a bet that your Lean Launchpad class can apply the scientific method to market-opportunity identification. We think your class can train scientists to start companies better than how we're doing it now."

Uh oh, where's this heading? "We want to select the best of our researchers, pay them $50,000 to take your class and see if we can change the outcome of their careers and their research."

"That's great, maybe I can set up a class for you next year," I replied. The answer shot back, "We want the class to start in 90 days."

I remember thinking, "Wow, whoever's on the other end of the phone sounds just like an entrepreneur, they were asking for the impossible." Just as I was computing whether this was possible, he added, "And we want to bring 25 new teams every quarter."

So of course, I said yes.

While they'll never admit it, the National Science Foundation was starting an incubator -- the Innovation Corps -- to take the most promising research projects in American university laboratories and turn them into startups.

The Innovation Corps -- Using the Lean LaunchPad as an Incubator for Scientists and Engineers
The Innovation Corps Startup Team
These weren't 22-year-olds who wanted to build a social shopping web site. Each of the teams selected by the NSF had a Principal Investigator -- a research scientist who was a University professor; an Entrepreneurial Lead -- a graduate student working in the Investigator's lab; and a mentor from their local area who had business and/or domain expertise. And they were hard at work at some real science.

The I-Corps Incubator Program
Unlike other incubators, our Lean LaunchPad Class had a specific curriculum. We taught them the business model/customer development/agile development solution stack. This methodology forces rapid hypothesis testing and Customer Development by getting out of the building while building the product. (The mentors in our program are there to support the methodology, but aren't there to tell stories.)

The gamble was that we could train Professors doing hard-core science, who had never been near a startup or Silicon Valley, to get out of the building and talk to customers and Pivot as easily as someone at a web startup.

The Scientists, the NSF and the teaching team were all going to go where no one had before. Given that Silicon Valley had started with scientists and engineers, not MBAs, I thought this was a bet worth making.

The Curriculum
Since the teams were in universities scattered across the U.S., we couldn't keep them in Silicon Valley for all eight weeks, so we tried an experiment in teaching remotely.

First, we brought all 21 teams to Stanford for three days of 10 hour-a-day classes in business model design and customer development. After returning to their schools, they got out of their labs while they built their products. Once a week, via Webex, they presented their Customer Development progress on line to the teaching team and the other teams. Then it was our turn, and we lectured all the teams remotely. After seven weeks they returned to Silicon Valley for their final presentations.

(The class syllabus is here. The class textbooks were The Four Steps to the Epiphany and Business Model Generation)."

Assembling the Teaching Team
We recruited two veteran Venture Capital partners to be part of the 10-week teaching team: Jon Feiber at Mohr Davidow and John Burke of True Ventures. Alexander Osterwalder joined us for the opening day, and Oren Jacob, ex-CTO of Pixar, joined us for a finale.

The First Class
As the first class settled into their seats at Stanford I wondered if we were going to be able to get them to act like startups. Most of the Principal Investigators were professors. Some had their own labs managing large groups of researchers. Their average age was in the mid-40s. Their mentors were at least that old. Only the Entrepreneurial Leads (the PI's assistants) were in their mid to late 20s.

Looking at them I wondered if: 1) hard-core science and engineering projects could rapidly pivot, 2) if the Principal Investigators would simply "assign" the work to their graduate students. I thought about the common wisdom that only 20-year-olds doing Internet startups could be agile. Some incubators would have labeled this group too old to be entrepreneurs. I smiled as I realized that I was older than most (but not all) of them.

The Stanford Lectures

Our first lecture was about 1) how to organize their thinking of what it takes to build a startup -- the business model canvas and 2) how to test their hypotheses -- the Customer Development Process.

Since the first part of the lecture was about Alexander Osterwalder's Business Model Canvas, Osterwalder flew in from Switzerland to teach slides 20 through 76. And since the rest of the slides were about Customer Development, I taught those.

View more presentations from Steve Blank.

The homework for the 21 teams in the next 24 hours? Come up with a business model canvas for their startup. And tell us how they will test each of their business model hypotheses.
As day one ended, I wondered what those canvases would look like.

Stay tuned for Part 2.

Steve Blank's blog:

Popular in the Community


What's Hot