How to Increase Chances of NIH SBIR Grants 333%? Lean Startup Principles and Life Science Startups

Rapid Prototyping at Google X

I attended a rapid prototyping workshop led by Tom Chi, a founding member of Google X, during the 2016 Mentor Capital Network Gathering.  Google X is the design team that came up with Google Glass, self-driving cars, Project Loon (a network of high altitude balloons providing wi-fi coverage to remote regions), contact lenses that measure glucose levels, and other bold innovations.

One common comment Tom hears is that Google is a huge company that can throw large amounts of money and talent against problems.  In actuality, Project Loon for example was designed by a lean team of 5 people in 4 months, rapidly cycling through 100 prototypes, using $70,000 worth of hardware.

According to Tom, the Google Glass project was a response to a one paragraph vision by Google co-founders Sergei Brin and Larry Page asking the question, what if people could know what they wanted to know as soon as they wanted to know it?  After quickly throwing out certain approaches such as brain implants in favor of something of quality achievable within a short period of time, the concept of Google Glass emerged in 10 weeks from a team of 3 people rapidly cycling through 150 prototypes made from about $20,000 worth of hardware.

Tom found the first 45 minutes of the first meeting on the Google Glass project exciting, inspiring, energizing; and the second 45 minutes depressing and a waste of time.  The discussion had devolved into smart people arguing about what color the display should be.  Nobody had direct experience, arguments were advanced in what Tom termed “guessathons”, and discussion ended when Sergei decreed it had to be red, among other things because red is the color with the lowest energy photons, which would be less taxing on the eye.

Frustrated, Tom left the meeting and made a rough prototype in 3 hours using coat hangers, a transparent report cover and a light source that could project different colored text onto the report cover. The team was quickly reconvened to try the prototype, and it was obvious within 90 seconds that red was horrible, purple only less so, and yellow, white and blue were the best.  It was true that red photons had the least energy, but for that reason, red text on the report cover was the hardest to see against the background cacophony of higher energy photons from the world seen through the report cover.

The team then recognized that Google Glass would need some user interface, and sought to turn Tom Cruise’s virtual touch screen in the air from the film, Minority Report, into reality.  It took them 45 minutes to mock up the first prototype, using hair bands, fishing line and chopsticks connected to laptop controls.  Direct experience led to very quick learning.  For example, one team member discovered the ‘Wii Tennis’ effect – the arm motions could be reduced to minimalist gestures; another went the opposite way, really getting into expansive arm gestures, until he discovered that making broad gestures with arms held above heart level for several minutes is tiring, like doing arm circles.  Such learnings suggested modifications that could be iteratively implemented and tested to end up with a product much more likely to succeed commercially.

Through experiences like this, Tom derived certain principles, such as:

  • Create the quickest path to direct experience
  • Doing is the best kind of thinking
  • Don’t guess.   Progress rapidly through cycles of Conjectures-Experiments-Actuals-Decisions

Along the way, he challenged companies to reduce cycle times for the next generation prototypes from months to hours and minutes, which could convert a product candidate with e.g. a 5% technical and commercial chance of success to a 20th generation product candidate with a 64% chance of success, or a 50th generation product candidate with a 92% chance of success.

Lean Startup

Tom’s rapid prototyping is an element of the Lean Startup methodology, pioneered and popularized by Steve Blank and Eric Ries.  Inspired, among others, by Toyota lean manufacturing practices, lean startup seeks to reduce commercial risk and shorten development cycles through accelerated, iterative, hypothesis-driven generation of ‘minimum viable products’ to perform ‘customer discovery/development’, yielding ‘validated learning’ from ‘actionable (as opposed to vanity) metrics’, ‘pivoting or persevering’ as necessary to negotiate a path charted by the ‘value and growth hypotheses’.

Lean Startup and Life Sciences

It is easy to assume that lean startup principles may be fine for a tech/consumer product like Google Glass, but not really applicable to life science products.  If a new drug cures a fatal disease, the value and growth hypotheses are self-evident.  New drug candidates cannot be rapidly and iteratively modified and tested in clinical subpopulations. When I asked Tom Chi how applicable rapid prototyping could be in life sciences, he distinguished between psychological constraints (on the part of target customers), which are very open to accelerated, iterative, validated learning through rapid prototyping, and constraints of physics/chemistry/biology (e.g., on pharmacokinetics and pharmacodynamics), for which massive parallel testing may be better, where possible.

Nevertheless, lean principles can have major utility in applicable domains of life science startups.  For example, Steve Blank started a lean startup course for life sciences to which NIH scientists subscribe (I-Corps™@NIH) on how to, e.g., rapidly define clinical utility before spending millions of dollars, understand the core and tertiary customers and the steps necessary to build sales, assess IP and regulatory risk before they design and build, and identify what data will be required by partners and acquirers before doing the science. One precursor cohort of National Science Foundation I-Corps™ graduates (a number of whom were building biotech, medical device or digital health products) had a 60% success rate in Small Business Innovation Research (SBIR) grant applications, compared to 18% by a cohort of companies that had not gone through the program (an increase of 333%)Some next-generation variants of incubators/accelerators (co-working and collaboration spaces) are being created that include simulated operating rooms, ICUs, doctor’s offices, home care settings, etc. where representative stakeholders can ‘use’ and give feedback on minimum viable products.

The lean startup methodology provides points of contrast with a more traditional approach, in which a practitioner might conceive of a novel device as a solution to a problem, raise money, spend some years and millions of dollars to develop and obtain regulatory approval for the product, then be disappointed at limited market uptake.  In a lean approach, the practitioner’s concept is treated as a hypothesis, and multiple (e.g. >100) stakeholders (including other practitioners, customers, payers and potential partners) are interrogated to validate the problem, the proposed solution, and refinements in response to customer and other stakeholder feedback, before significant time and money have been spent.  In a traditional approach, prototype optimization does occur, but early design lock is prized, typically based on the insights of the inventor and/or entrepreneur.  The assumption is that quick development is key, in contrast to quick development of a product first optimized for stakeholder acceptance and market pull.

As demonstrated by Tom Chi and Google X, the incubation and optimization phase can be quite quick and inexpensive; however, it does require some investment of time and money, as well as buy-in on the benefits of a lean approach.  Academic researchers may be able to collaborate with affiliated business school classes studying the lean start up model.  The ecosystem also needs to educate and persuade more seed investors on the benefits of the lean approach.

Could lean startup principles even apply to therapeutics development?  Approaches such as high throughput screening, in silico systems biology modeling, failing fast, critical path, target product profile, real options, repositioning/repurposing and Phase 0 and adaptive clinical trials anticipate some of the spirit of lean principles.  The late Leigh Thompson, M.D., Ph.D., former CSO of Lilly, liked to point out that we learn way more from dosing a NME under an IND in the first human than in multiple additional animal models.  However, rapid prototyping and testing of multiple successive generations of minimum viable product candidates in the clinic seem elusive (but compare with novel modular technical and regulatory approaches under interrogation e.g. for rapid vaccine development for different flu strains and common oligo backbones to address gene variants of single gene diseases).   On the other hand, testing assumptions and hypotheses with doctors, patients, providers, purchasing departments, strategic partners and other stakeholders in advance can surely reduce risk and waste.

Case Studies

Steve Blank’s blog describes half a dozen case studies of how lean startup training overturned assumptions and induced pivots in life sciences startups developing product candidates such as a radiosensitizer for cancer, a therapeutic for high risk neuroblastoma, a drug-coated arteriovenous graft, and a neural stimulator to treat atrial fibrillation.

Adoption and Broader Applications

Eric Ries started blogging on lean principles in 2008; he published The Lean Startup in 2011. His mentor, Steve Blank, published his Harvard Business Review article on the lean methodology and his The Four Steps to the Epiphany in 2013, and started blogging about the application of lean methodologies to the life science space later that year. So while these ideas are not brand new, their acceptance in the life science space still seems to be in the shallow lower left part of the “S”-shaped adoption curve.

The methodology is spreading among hundreds of scientist entrepreneurs each year, thanks to I-Corps, supported by some of the earliest stage incubator investors, like Steve Blank’s M34 Capital and a couple forward thinking angel funds like Allan May’s Life Science Angels, and is taught at many business schools. However, it does not yet seem to have the same traction with post-seed stage companies, whose investors may not relish pausing to interrogate the assumptions upon which they made their initial investments and perhaps having to face indications to pivot.

More broadly, aspects of rapid prototyping and lean principles are applicable to the introduction of novel goods or services by larger, more established organizations as well, not only at innovation leaders like Google and other learning organizations, but government agencies such as the Department of Defense and the Department of StateLean principles are even being taught to high school students. Imagine if our lawmakers performed Customer Discovery and rolled out rapidly prototyped solutions for “pain point” problems of voters! See for example Tom Chi’s anecdote at the 8:15 mark during his 2013 Kyoto TEDx talk.

Lean principles are not a panacea — Google Glass is not yet a mass success. But in the long run, the odds favor experimentation over assumptions; the wisdom of relevant crowds over intuitive genius; and products rapidly honed through dozens of evolving prototypes based on stakeholder feedback over those based on a handful of prototypes focused on technical refinement. Such principles are quickly becoming the standard in this generation.


I would love to hear feedback, particularly about:

  • specific successes and failures in applying lean startup principles to the life science space
  • extent of awareness and buy-in about lean startup principles among scientist-inventors, entrepreneurs, funders and grant-makers, partners and other stakeholders in the life science ecosystem
  • how principles of rapid prototyping could (not just could not) be applied to life science product candidate development.

Illustrative Short Slide Deck (Annotated)

1 Overview (Snapshot)

Ideally, this is the only slide the investor needs to see to ‘get’ the business opportunity, and prime the investor to check out the rest of the story. The first bullet should make crystal clear what the company does: e.g., “Next generation catheter-based devices for endovascular and neurosurgical interventions”; “Focus on novel anti-proliferative drugs that modulate a key new target in the VEGF pathway for cancer and ocular diseases.” Continue reading

Advice for Academic Start-Ups

We’ve done a fair amount of advising academic faculty about the viability and readiness of their inventions to attract start up funding.  Here are some recurring themes:

Idea vs. Business Opportunity

When should an academic with an idea about an approach or technology discuss it with potential investors?  While there can be unusual exceptions, an idea is typically not commercially fundable until it has been elaborated into a product or service in a sufficiently large market that provides a solution to a problem that customers are willing to pay for. Continue reading

Partnering Out a Program or Technology

Designing and executing a partnering campaign, like other business activities, benefit from proactive preparation informed by the full landscape of what is to be done.  Here is a brief overview.


If a company has the funds, expertise and bandwidth, it is generally preferable to develop a product candidate further to a point of higher valuation rather than partner.  In contrast, it may make sense to partner a technology platform early and often to validate the technology.  A deal is a means, not an end; a company should be clear about what the deal and its terms will do for its corporate development. Continue reading