22. Can We Build a Process to
Search?
Before we Execute
23. 1. Startups are a series of untested hypotheses
2. You can test these hypotheses with the same
scientific method you use in the lab, but…
3. There are no facts inside your building, so get the
hell outside
3 Big Ideas
28. And Then A Class
MBA 295: Customer Development
29. Eric Ries Extends the Model
• Took my class at U.C. Berkeley
• Co-founded IMVU, I sat on his board
– 1st implementation of Customer Development
– Paired it with an Agile Development Model
• Called it the Lean Startup
46. Then It Turned Into Another Class
A Scientific Method For Entrepreneurship
47. LaunchPad Central Software
Lean Startup
Business Model
Canvas
Customer
Development
Agile Engineering+ +
Evidence-based Curriculum (NSF I-Corps/Lean LaunchPad)
Hypotheses Experiments Data
48. LaunchPad Central Software
DataDataData
Lean Startup
Business Model
Canvas
Customer
Development
Agile Engineering+ +
Evidence-based Curriculum (NSF I-Corps/Lean LaunchPad)
Hypotheses Experiments Data
Weekly Progress
49. LaunchPad Central Software
DataDataData
Lean Startup
Business Model
Canvas
Customer
Development
Agile Engineering+ +
Evidence-based Curriculum (NSF I-Corps/Lean LaunchPad)
Hypotheses Experiments Data
Weekly Progress
Experiment
Scorecard
+
50. LaunchPad Central Software
DataDataData
Lean Startup
Business Model
Canvas
Customer
Development
Agile Engineering+ +
Evidence-based Curriculum (NSF I-Corps/Lean LaunchPad)
Hypotheses Experiments Data
Weekly Progress
Experiment
Scorecard
Data, Experiments, T
rajectory, IRL
+ =
58. Taught by Takashi Tsutsumi@ Hosei
Co-taught with Murray Low @ Columbia
Bob Dorf Co-taught with: Jon Feiber @ MDV
Taught by Jim Hornthal
And Four More Schools
Add 5-day version of the class
5-day Version
67. National Science Foundation
Commercialization Problem Prior to I-Corps
• Disparate centers
– All thought their culture was different
– All had their own commercialization program
– No best practices
– No common training/education
– No common platform
– All had technology-centric focus
• No centralized leadership
69. 69
Current Thinking about
Translational Medicine
Research Performing Institutions
Technology
(diagnostic, device, therapeutic)
Solicit &
Select
Develop Technology
Obtain Additional
Capital
70. Research Institutions
Technology
(diagnostic, device, therapeutic)
Solicit &
Select
Existing
Company
New
Company
Technology
Development Process
Regulatory
Intellectual
Property
Business
Development
Project
Management
Medical, Scientific, an
d Business Review
Licensing and Exit
Accepted
independent financing
Product
Development
Additional
Capital
Current Thinking about
Translational Medicine
Mentors
73. Answers to Hypotheses are
Outside The Lab
• You may be the smartest person in your lab
• But you are not smarter than the collective
intelligence of your potential
customers, partners, payers and regulators
• You can’t learn this by reading papers or
listening to lectures
Need a process for hypotheses testing
76. Research Institutions
Technology
(diagnostic, device, therapeutic)
Solicit &
Select
Add Evidence-based Commercialization
Medical, Scientific, an
d Business Review
Accepted
Additional
Capital
Reimburse
-men
Partners
Customers
(users, payers, etc.
)
Commercialization
Development
Process
Value
Propositio
n
Distribution
Channels
Add a parallel
Commercialization
Process
Inward-facing Outward-facing
77. Research Performing Institutions
Technology
(diagnostic, device, therapeutic)
Solicit &
Select
Add Evidence-based Commercialization
Medical, Scientific, an
d Business Review
Accepted
Additional
Capital
Customers
Value
Proposition
Channels
Partners
Reimburse-
ment
Intellectual
Property
Commercialization
78. Magnamosis
• Create a magnetic compression
anastomosis with improved outcomes
• Team:
– Michael Harrison, MD, Pediatric Surgeon
– Elisabeth Leeflang, MD, General Surgery Resident
– Michael Danty, MS, Business Development
– Dillon Kwiat, BS, Medical Device Engineer
89. • 17 out of 233 teams were identified as high performer teams
from 10 NSF I-Corps 2012-13 cohorts
Evidence-Based Entrepreneurship
90. Signal: Customer Interviews
High Performers:
• Interview ~ 2X more customers on average
• Peak customer interviews during week 4
Average Customer Interviews Per Week
Week Week
Interviews
92. High Performance Teams demonstrate
significant cadence by mid-point (week 4)
• The highest number of
hypotheses invalidation
occurs in week 4
• The highest number of
customer interviews
occur in week 4
• High performance
teams maintain a
relentless pace of
customer interviews
reaching 50% of 100
interviews goal by
week 4
Average Customer Interviews by Week
Weeks
Interviews
96. I-Corps – Insights
• It’s not just about the science
– Technology and commercialization progress require
separate processes
– PI’s can’t figure out commercialization sitting in their labs
– Technology mentorship is part of the process but it’s
insufficient
– You can’t outsource commercialization to a proxy
(consultants, market researchers, etc.)
97. Outward-Facing Commercialization &
Translational Medicine
• Getting out of the building is a big idea
• It accelerates speed of translation
• It makes our national research enormously more efficient
98. NSF Commercialization Solution
I-Corps
• One commercialization program
– Common Platform
– Centralized Leadership
• Best practices
– Disruptive idea – Lean/Evidence-based
– Common training/education
– Consistent evaluation criteria
• Nodes not centers
– Linkages between nodes
99. NSF I-Corps Results
• Early evidence of success in SBIR Phase I funding
– 18% of teams who did not take the class
– 60% of teams who did take the class
100. NSF Program Outcomes
• Scientists & Engineers trained as Entrepreneurs
– pass on their knowledge to students
• Network of Mentors/Advisors
• Increased impact of NSF-funded basic research
101. Evidence-Based Entrepreneurship
• What we now know
– Commercialization must be a parallel track
– Effort must be experiential (No Proxies)
– New metrics allow for rational discussion
– IT Platform provides tracking capacity:
Continuous improvement
102. Not About Picking the Winners
• Enable lots of low cost experiments
• Kill the Losers
• Double down on the ones that show progress
104. Early Stage Therapeutic Myths
The Idea is Key
Better ideas create value
Funding Gap
Early Stage
Investment as a
market failure
Data Quality
Findings in preclinical
research are often
not reproducible
Karl@CodonCapital.com
105. Early Stage Therapeutic Myths
The Idea is Key
Better ideas create value
Funding Gap
Early Stage
Investment as a
market failure
Data Quality
Findings in preclinical
research are often
not reproducible
Karl@CodonCapital.com
The real gap is the expertise to move early stage
research toward industrial relevance
106. Early Stage Therapeutic Myths
Karl@CodonCapital.com
Data Quality
Findings in preclinical
research are often
not reproducible
Data addressing key
development criteria
107. Early Stage Therapeutic Myths
Karl@CodonCapital.com
Data Quality
Findings in preclinical
research are often
not reproducible
Data addressing key
development criteria
The Idea is Key
Better ideas create value
Clear path to
modifying a disease
108. Early Stage Therapeutic Myths
Karl@CodonCapital.com
Data Quality
Findings in preclinical
research are often
not reproducible
Data addressing key
development criteria
The Idea is Key
Better ideas create value
Clear path to
modifying a disease
Funding Gap
Early Stage
Investment as a
market failure
Operational
planjustifying investment