I-Corps at the NIH: Evidence-based Translational Medicine

If you’ve received this post in an email the embedded videos and powerpoint are best viewed on www.steveblank.com

We have learned a remarkable process that allow us to be highly focused, and we have learned a tool of trade we can now repeat. This has been of tremendous value to us.

Andrew Norris, Principal Investigator BCN Biosciences

Over the last three years the National Science Foundation I-Corps has taught over 700 teams of scientists how to commercialize their technology and how to fail less, increasing their odds for commercial success.

To see if this same curriculum would work for therapeutics, diagnostics, medical devices and digital health, we taught 26 teams at UCSF a life science version of the NSF curriculum. 110 researchers and clinicians, and Principal Investigators got out of the lab and hospital, and talked to 2,355 customers. (Details here)

For the last 10 weeks 19 teams in therapeutics, diagnostics and medical devices from the National Institutes of Health (from four of the largest institutes; NCINHBLI, NINDS, and NCATS) have gone through the I-Corps at NIH.

87 researchers and clinicians spoke to 2,120 customers, tested 695 hypotheses and pivoted 215 times. Every team spoke to over 100 customers.

Three Big Questions
The NIH teams weren’t just teams with ideas, they were fully formed companies with CEO’s and Principal Investigators who already had received a $150,000 grant from the NIH. With that SBIR-Phase 1 funding the teams were trying to establish the technical merit, feasibility, and commercial potential of their technology. Many will apply for a Phase II grant of up to $1 million to continue their R&D efforts.

Going into the class we had three questions:

  1. Could companies who were already pursuing a business model be convinced to revisit their key commercialization hypotheses – and iterate and pivot if needed?
  2. Was getting the Principal Investigators and CEO out of the building more effective than the traditional NIH model of bringing in outside consultants to do commercialization planning?
  3. Would our style of being relentlessly direct with senior scientists, who hadn’t had their work questioned in this fashion since their PhD orals, work with the NIH teams?

Evidence-based Translational Medicine
We’ve learned that information from 100 customers is just at the edge of having sufficient data to validate/invalidate a company’s business model hypotheses. As for whether you can/should push scientists past their comfort zone, the evidence is clear – there is no other program that gets teams anywhere close to talking to 100 customers. The reason? For entrepreneurs to get out of the building at this speed and scale is an unnatural act. It’s hard, there are lots of other demands on their time, etc. But we push and cajole hard, (our phrase is we’re relentlessly direct,) knowing that while they might find it uncomfortable the first three days of the class, they come out thanking us.

The experience is demanding but time and again we have seen I-Corps teams transform their business assumptions. This direct interaction with potential users and customers is essential to commercialize science (whether to license the technology or launch a startup.) This process can’t be outsourced. These teams saved years and millions of dollars for themselves, the NIH and the U.S. taxpayer. Evidence is now in-hand that with I-Corps@NIH the NIH has the most effective program for commercializing science.

Lessons Learned Day
Every week of this 10 week class, teams present a summary of what they learned from their customers interviews. For the final presentation each team created a two minute video about their 10-week journey and a 8-minute PowerPoint presentation to tell us where they started, what they learned, how they learned it, and where they’re going. This “Lessons Learned” presentation is much different than a traditional demo day. It gives us a sense of the learning, velocity and trajectory of the teams, rather than a demo day showing us how smart they are at a single point in time.

BCN Biosciences
This video from team BCN Biosciences describes what the intensity, urgency, velocity and trajectory of an I-Corps team felt like. Like a startup it’s relentless.

BCN is developing a drug that increases anti-cancer effect of radiation in lung cancer (and/or reduces normal tissue damage by at least 40%). They were certain their customers were Radiation Oncologists, that MOA data was needed, that they needed to have Phase 1 trial data to license their product, and needed >$5 million and 6 years. After 10 weeks and 100 interviews, they learned that these hypotheses were wrong.

If you can’t see the BCN Biosciences video click here

The I-Corps experience helped the BCN Bioscience team develop an entirely new set set of business model hypotheses – this time validated by customers and partners. The “money slides” for BCN Biosciences are slides 22 and 23.

If you can’t see the BCN Biosciences presentation click here

You Can’t Outsource Customer Discovery
What we hear time and again from the Principal Investigators is “I never would have known this” or “I wouldn’t have understood it if I hadn’t heard it myself.” Up until now the NIH model of commercialization treated a Principal Investigator as someone who can’t be bothered to get out of the building (let alone insist that it’s part of their job in commercialization.) In the 21st century using proxies to get out of the building is like using barbers as surgeons.

Clinacuity
While the Clinacuity video sounds like an ad for customer discovery, listen to what they said then look at their slides. This team really learned outside the building.


If you can’t see the Clinacuity video click here

Clinacuity’s technology automatically extracts data in real-time from clinical notes, (the narrative text documents in a Electronic Health Record,) and provides a summary in real time. Their diagrams of the healthcare customer segment in slides 15-18 were outstanding.

If you can’t see the Clinacuity presentation click here

GigaGen
The GigaGen team – making recombinant gamma globulin – holds the record for customer discovery – 163 customer interviews on multiple continents.

If you can’t see the GigaGen video click here

GigaGen’s learning on customer value proposition and who were the real stakeholders was a revelation. Their next-to-last slide on Activities, Resouces and Partners put the pieces together.

If you can’t see the GigaGen presentation click here

Affinity Therapeutics
Affinity came into class with a drug coated Arterial Venous Graft – graft narrowing is a big problem.

One of things we tell all the teams is that we’re not going to critique their clinical or biological hypotheses. Yet we know that by getting out of the building their interaction with customers might do just that. That’s what happened to Affinity.

If you can’t see the Affinity video click here

Affinity was a great example of a team that pivoted their MVP. They realized they might have a completely new product – Vascular wraps that can reduce graft infection.  See slides 17-23.

If you can’t see the Affinity presentation click here

Haro
Haro is making a drug for the treatment of high risk neuroblastoma, the most common extracranial cancer in infancy and childhood. On day 1 of the class I told the team, “Your presentation is different from the others – and not in a good way.”  That’s not how I described them in the final presentation.

If you can’t see the Haro video click here

After 120 interviews the Haro found that there are oncology organizations (NCI-funded clinical development partners) that will take Haro’s compound and develop it at their own expense and take it all the way into the clinic. This will save Haro tens of millions of dollars in development cost.  See slides 12 and 13.

If you can’t see the Haro presentation click here

Cardiax
Caridax is developing a neural stimulator to treat atrial fibrillation. Their video points out some of the common pitfalls in customer discovery. Great summary from Mark Bates, the Principal Investigator: “You don’t know what you don’t know. Scientific discovery is different than innovation. You as a prospective entrepreneur need this type of systematic vetting and analysis to know the difference.”

If you can’t see the Cardiax video click here

After 80 interviews they realized they were jumping to conclusions and imparting their bias into the process. Take a look at slides 8-11 and see their course correction.

If you can’t see the Cardiax presentation click here

The other 15 presentations were equally impressive. Each and every team stood up and delivered. And in ways that surprised themselves.

The Lean Startup approach (hypotheses testing outside the building,) was the first time clinicians and researchers understood that talking to customers didn’t require sales, marketing or an MBA – that they themselves could do a pretty good first pass. I-Corps at NIH just gave us more evidence that’s true.

The team videos and slides are on SlideShare here.

A Team Effort
This blog post may make it sound like there was no one else in the room but me and the teams. But nothing could be farther from the truth. The I-Corps@NIH teaching team was led by Edmund Pendleton. Allan May/Jonathan Fay taught medical devices, John Blaho/Bob Storey taught diagnostics and Karl Handelsman/Keith McGreggor taught therapeutics. Andre Marquis, Frank Rimalovski and Dean Chang provided additional expertise. Brandy Nagel was our tireless teaching assistant. Jerry Engel is the NSF I-Corps faculty director.

Special thanks to Paul Yock of Stanford Biodesign and Alexander Osterwalder for flying across the country/world to be part of the teaching team.

I created the I-Corps/Lean LaunchPad® syllabus/curriculum, and with guidance from Allan May, Karl Handelsman Abhas Gupta and Todd Morrill adapted it for Life Sciences/Health Care/Digital Health. The team from VentureWell provided the logistical support. The I-Corps program is run by the National Science Foundation (Babu Dasgupta, Don Millard and Anita LaSalle.) And of course none of this would be possible without the tremendous and enthusiastic support and encouragement of Michael Weingarten the director of the NIH/NCI SBIR program and his team.

Lessons Learned

  • The I-Corps/Lean LaunchPad curriculum works for therapeutics, diagnostics and device teams
  • Talking to 100 customers not only affected teams’ commercial hypotheses but also their biological and clinical assumptions
  • These teams saved years and millions of dollars for themselves, the NIH and the U.S. taxpayer
  • Evidence is now in-hand that the NIH has the most effective program for commercializing science
  • In the 21st century using proxies to get out of the building is like using barbers as surgeons

Why Corporate Skunk Works Need to Die

In the 20th century corporate skunk works® were used to develop disruptive innovation separate from the rest of the company. They were the hallmark of innovative corporations.

By the middle of the 21st century the only companies with skunk works will be the ones that have failed to master continuous innovation. Skunk works will be the signposts of companies that will be left behind.

 

——–

In the 20th century companies could be leaders in a market for decades by just focusing on their core product(s). Most companies incrementally improved their products with process innovation (better materials, cheaper, product line extensions) and/or through acquisitions. Building disruptive products were thought of as “risky” and a distraction since it was not “core” to the company and did not fit existing corporate structures. Why make big bets if no one was asking for them and competitors weren’t doing so.

a-12

CIA A-12 spy plane. Developed by Lockheed Skunkworks

A few innovative companies did push the envelope. The way they did so was to set up “skunk works” to develop their most advanced, disruptive products. (IBM used the process to develop the IBM PC.) But it was Lockheed, then an aircraft manufacturer that coined the term and perfected the art. The Lockheed Skunk Works® led by Kelly Johnson was responsible for its Advanced Development Projects – everything from the P-80, the first U.S. jet fighter plane, to the U-2 and A-12 spy planes.

Skunk works differed from advanced research groups in that they were more than just product development groups. They had direct interaction with customers and controlled a sales channel which allowed them to negotiate their own deals with customers.

Decades before we were able to articulate the value of “getting out of the building” and the Lean Startup, the value in having skunk works controlling their own distribution was starkly evident. Other companies with world-class R&D groups built radical innovations only to see their company fumble the future and others reap the rewards (think of Xerox and the personal computer, Fairchild and integrated circuits, Kodak and digital photography, etc.) Common themes in these failures were, 1) without a direct connection to the customer advanced R&D groups built products without understanding user needs, and 2) the core of the company was so focused on execution of current products that it couldn’t see that the future didn’t look like the past.

Kelly Johnson’s 14 rules about how to manage a disruptive project described how to remove a small innovative team from the politics, policies, procedures and processes a large company had built to support execution of its core business (and its military customers had developed to procure large numbers of standard aircraft.)

With the vantage point of the 21st century, we can now see that a successful skunk works – separated from its corporate parent, with its own culture, in control of its own R&D and distribution channel – looked much like a startup.

But as successful as skunks works were to the companies that executed them well, innovation and execution couldn’t co-exist in the same corporate structure. Skunk works were emblematic of corporate structures that focused on execution and devalued innovation.

Until now.

Continuous Disruption Requires Continuous Innovation
In the 21st century market share is ephemeral – ask General Motors, Blackberry, Nokia, Microsoft, Blockbuster, etc. –disruption is continual.

Therefore companies need to master continuous innovation – the art of executing on core products while continually inventing new products and new businesses. That means that somehow we need to take the innovation that a skunk works removed from the core of the company and integrate the two.

Here’s how.

We need to realize that skunk works epitomize innovation by exception. But to survive companies need innovation by design.

We now know how to do just that. We can get innovation and execution to work side-by-side.

To start it requires board support and CEO and executive staff agreement. And recognition that cultural, process and procedure changes are needed to embrace learning and experimentation alongside the existing culture of execution.

I’ll provide details on how companies can organize this way in a follow-on post.

Lessons Learned

  • Skunk were advanced/disruptive product groups organizational isolated from the rest of their company
  • Skunk works had control over their sales channel and had direct customer feedback
  • World-class R&D groups that didn’t control the channel often saw their innovation die internally
  • Skunk works looked much like a startup
  • Skunk works epitomized innovation by exception
  • Companies now need innovation by design – innovation and execution that work side-by-side
  • We now know how to do this

Born Global or Die Local – Building a Regional Startup Playbook

Entrepreneurship is everywhere, but everywhere isn’t a level playing field. What’s the playbook for your region or country to make it so?

playbook

———-

Scalable startups are on a trajectory for a billion dollar market cap. They grow into companies that define an industry and create jobs.  Not all start ups want to go in that direction – some will opt instead to become a small business. There’s nothing wrong with a business that supports you and perhaps an extended family. But if you want to build a scalable startup you need to be asking how you can you get enough customers/users/payers to build a business that can grow revenues past several $100M/year.

With 317 million people the U.S. has a large enough market that most U.S. startups ignore the rest of the world until they scale in their own country. Outside the U.S. a rough rule of thumb for scale is a local population greater than 100 million (and language, cultural and/or regulatory barriers to delay or keep out U.S. entrants.) China, Russia, Brazil, India, Indonesia all meet those criteria. (Obviously this depends on industry and application.) However, most countries don’t have sufficient population to support scale with just their local market and ultimately need to be global players – from day one.

Regional Ecosystems
I’m in Australia and just spent time with some great entrepreneurs in Melbourne.

Bay of Fires Tasmania

Bay of Fires Tasmania

One of the groups I spoke to was the Australian Sports Technology Network. This group realized that Australia has a great reputation as one of the world’s best sporting nations. They realized if they could develop and promote a well-coordinated sports technologies industry, they could capture their unfair share of the $300 billon sports consumer market. So they put together a sports technology ecosystem – gathering sports startups in apparel and footwear, protective wear, equipment, nutrition, wearable devices, data and video analytics, and web and mobile solutions and brought them together with investors, retailers and distributors, universities, research centers and national sporting organizations.

Creating a vertically oriented regional ecosystem is a pretty amazing accomplishment for any country or industry.

However, in meeting some of the sports startups one of the things that struck me is that most of the founders who said they wanted to grow big hadn’t given much thought about how they would go about building size and scale.

The trap most of them fell into (common almost everywhere): they were reading the blog posts and advice of Silicon Valley-based companies and believing that it uniformly applied to them.

It doesn’t.

Born Global or Die Local
The biggest mistake for most of these startups was not understanding that optimizing their business model for the 24 million people in the Australian market would not prepare them for the size and scale they needed to get to big.

Instead of beginning with just a focus on Australia, these startups needed to use the business model canvas and articulate which of their hypotheses should be tested locally and what would require getting on an airplane to test by watching someone’s pupils dilate face-to-face.

For example, one of the critical business model hypotheses they could test locally is Product/Market fit – the connection between their Value Proposition (what product or service they were building) and the Customer Segment (who they were building it for.)

business model globals

Further refinement of Product/Market fit could be done locally by using Value Proposition Design.

bus model and value prop map

value prop map

But other critical hypotheses such as activities, resources, partners, channels needed testing offshore. For example, many of the Australian sports tech business models shared common elements. They intended to get scale for their business by growing in the U.S. while building their products in China. And their branding and demand creation activities were going to occur primarily outside of Australia. This meant they would need U.S. channel partners and Chinese manufacturers and customer acquisition and activation programs outside their home country. And as good as the Australian angel investors have been, there still is dearth of serious follow-on funding in Australia. This means that most follow-on rounds of tens of millions of dollars, if needed will likely come from outside the country.

Step 2 figure out what needs to be tested globallys

While the network was very helpful getting these startups together and introduced to investors, it wasn’t clear how and when these startups tested their “going global” hypotheses.

No one had written the playbook.

Building a Regional Startup Playbook
What’s been missing from regions outside of Silicon Valley is a “playbook.” In American football a playbook contains a sports team’s strategies and plays. It struck me that every region needs its own industry playbook on how to compete globally. For Australian sports startups, a playbook might lay out in detail the following steps:

  • Build minimal viable products and test product/market fit in Australia
  • Identify activities/resources/partners locally and then globally
  • Get seed funding in Australia
  • Trip 1 to China to understand manufacturing landscape, potential partners and rough cost of goods
  • Trip 2 to the U.S. to understand distribution channel landscape, potential partners and rough cost of customer acquisition
  • Test product/market fit in the U.S.
  • Trip 3 to China, pick manufacturing partner, start low volume production
  • Test channel and demand creation activities in the U.S.
  • Trip 4 to the U.S. Establish U.S. sales office
  • Trip 5 to the U.S. to get Series A funding in the U.S.

Each industry in a region should develop a playbook that expands and details the strategy and tactics of how to build a scalable startup. When a playbook is shared through regional collaborations (like the Australian Sports Techn Network,) entrepreneurs can jumpstart their efforts by sharing experience instead of inventing the wheel each time a new startup is launched. Now all they need is a playbook. As markets mature, and investors and the ecosystem become collectively smarter, the playbook will change over time.

Lessons Learned

  • A scalable startup typically requires a local population >100 million people
  • If your country doesn’t have that you need to be born global
  • Your country/industry needs a “go global” playbook

The Business Model Canvas Gets Even Better – Value Proposition Design

Product/Market fit now has its own book. Alexander Osterwalder wrote it. Buy it.

——

The Lean Startup process builds new ventures more efficiently. It has three parts: a business model canvas to frame hypotheses, customer development to get out of the building to test those hypotheses and agile engineering to build minimum viable products.

This week the author of the business model canvas, my friend Alexander Osterwalder, launched his new book Value Proposition Design, the sequel to his million copy best seller, Business Model Generation.

His new book does three things:
1. Introduces the Value Proposition Canvas
2. Tells you how to design new ventures with it
3. Teaches you how to use Customer Development to test it.

Value Proposition Design is a “must have” for anyone creating a new venture. It captures the core issues around understanding and finding  customer problems and designing and validating potential solutions.

Value prop design

Product/Market fit

Product/Market Fit
If you’re familiar with the Lean Startup you know that the Business Model Canvas is the tool to frame all the hypotheses of your startup. Of all the 9 boxes of the canvas, the two most important parts of the business model are the relationship between the Value Proposition (what you’re building) and the Customer Segment. These two components of the business model are so important we give them their own name, “Product/Market Fit.”

The Value Proposition Canvas functions like a plug-in to the Business Model Canvas and zooms into the value proposition and customer segment to describe the interactions between customers and product more explicitly and in more detail. This keeps things simple by giving you the big picture at the business model level and the detailed picture at the “product/market fit” level.

bus model and value prop mapvalue prop map

Integration with Customer Development and Lean Startup
Alexander and I met after he published Business Model Generation. We both realized that we had each invented one of the two parts that define the Lean Startup. In his new book he’s integrated Customer Development with the Business Model and Value Proposition Canvas and added some new tools to the mix.

Now an integral part of Value Proposition Design, several of his new tools help with testing and validation of hypotheses. These testing tools match the first two of the four steps of Customer Development. The diagram below is one of my favorites of the book and provides a simple overview of how to conduct customer discovery and customer validation in combination with the Business Model and Value Proposition Canvas. You start by extracting and prioritizing your hypotheses, then design your tests with Test Cards and finally, you conduct your tests and capture your learning. To make this all actionable Osterwalder added an Experiment Library to the book that equips you with ideas on how to test your assumptions.

3_Value_Proposition_Design_Testing_Process

Tracking Customer Development with the Value Proposition Canvas
With Customer Development you’re constantly talking to customers and partners and conducting a ton of experiments to validate and invalidate your hypotheses. All these activities, the evidence of what works and what doesn’t, and your progress towards finding a successful value proposition and business model need to be tracked. In Value Proposition Design Osterwalder shows how to do this with the Progress Board, a tool that includes a version of my investment readiness level thermometer to track progress.

5_Value_Proposition_Design_Progress_Board

Online Tools
Doing all the above together with your team is not easy when you “just” use poster-sized Canvases, sticky notes, and PowerPoint. There are simply too many Canvases you will design and trash (after rejecting and pivoting from your early tested ones), too many experiments you will conduct, and too much evidence you will produce. Keeping track of all this requires software support.

So the Value Proposition Design comes with a series of exercises that you can complete online with assessment tools that show you how you are using the Value Proposition Canvas. And last, but not least, you get access to a whole series of checklists, templates, and incredibly awesome posters that you can immediately use in your work.

Lessons Learned

  • The Value Proposition Canvas describes the details of how the value proposition and customer segments interact
  • It integrates the Customer Development process in the book
  • Product/Market fit now has its own book. Buy it

How We Used to Give Startups Very Bad Advice. 2 Minutes to See Why

If you can’t see the video click here

No Business Plan Survives First Contact With Customers. 2 Minutes to See Why

If you can’t see the video click here

Watching My Students Grow

“You cannot teach a man anything, you can only help him find it within himself.”
Galileo Galilei

One of the great things about teaching is that while some students pass by like mist in the night others remain connected forever. I get to watch them grow into their careers and cheer them on.

Its been three and a half years since I first designed and taught the Lean LaunchPad class and lots of water has gone under the bridge since then. I’ve taught hundreds of teams, the National Science Foundation Innovation Corps has taught close to 400 teams led by our nations top scientists, and the class is being taught around the world.

But I still remember a team from the first class, one which wanted to build a robotic lawnmower. It’s now been over 3 years since the team has left my classroom and I thought I’d share with you what the two founders, Jorge Heraud and Lee Redden, learned then and what they’re doing now.

The Autonomous Lawnmower
They called their company Autonomow. And they were absolutely convinced what the world needed was an auto-driving lawn mover for institutions with large green spaces.

You can see their first slide deck in class below (and here)

Like in all our Lean Launchpad classes we teach a combination of theory coupled with intense and immersive experiential learning outside the classroom. Students need to get out of the building and talk to 10-15 customers a week.

The next week they came back in class and presented this:

Each week we’d teach them about one more part of what makes up a business model. All teams struggle with finding product/market fit.

By week four their presentation looked like this:

Notice something different about the cover slide? Massive pivot. Like all great Silicon Valley companies they started with a technology and guessed who the customers will be. They’re almost always wrong. They could have never figured this out sitting inside a classroom writing a business plan.

At week five (see here) they were actually getting into farm fields wearing hip boots and overalls. Now they were figuring out how to create demand.

The Customer Development process, this relentless drive to turn hypotheses into facts is what makes this learning so rapid.

At week six they were trying to figure out their distribution channel (here) after another pivot. They got their minimal viable product (a machine vision platform) up and running in the lab.

At week seven (here) another pivot happened when farmers taught them about how to price their product. Instead of an of selling hardware they were selling a service.

BTW, notice that they were now dragging their machine vision platform through the farm fields!  If there was ever any question of whether a minimal viable product can work for hardware, see what they say in their video below.

By week eight they were learning who they needed to partner with (see here). Most importantly they found a customer who taught them while weeding carrots was nice, thinning lettuce was where the money was.

After 9 weeks their final presentation looked like this.

When I teach in universities I’m not running an incubator. What I’m trying to do is to get students to learn a way of thinking about new ventures that will stick with them for life. And I try to do by having them teach themselves, rather than us teaching at them. Whether they start a company or not, I don’t keep score.

But some teams remain connected forever. I get to watch them grow into their careers and cheer them on. This was one of those teams. After class they took this idea and formed a company – Blue River Technology.

Over the last three years they turned their vision and PowerPoint slides into real hardware that solves real customer problems. And with 3 rounds of funding, including a grant from the National Science Foundation, they’ve raised $13 million.

Take a look and see what they’ve done.

If you can’t see the video click here

“The customers had way more insights then we had. They had been thinking about their own problems for so long…If you just go out and try to sell maybe you’ll find some buyers, but you won’t be learning about what you should be doing.”

Lee Redden – Blue River Technology

——

I’m off next week on the next great adventure. We’re going to launch the I-Corps @ NIH and change how our country commercializes life sciences.

Follow

Get every new post delivered to your Inbox.

Join 166,195 other followers