While a few of the I-Corps teams are in web/mobile/cloud, most are working on advanced technology projects that don’t make TechCrunch. You’re more likely to see their papers (in material science, robotics, diagnostics, medical devices, computer hardware, etc.) in Science or Nature. The program pays scientists $50,000 to attend the program and takes no equity.
Currently there are 11 U.S. universities teaching the Lean LaunchPad curriculum organized as I-Corps “nodes” across the U.S. The nodes are now offering their own regional versions of the Lean LaunchPad class under I-Corps.
The NSF I-Corps uses everything we know about building Lean Startups and Evidence-based Entrepreneurship to connect innovation to entrepreneurship. It’s curriculum is built on a framework of business model design, customer development and agile engineering – and its emphasis on evidence, Lessons Learned versus demos, makes it the worlds most advanced accelerator. It’s success is measured not only by the technologies that leave the labs, but how many U.S. scientists and engineers we train as entrepreneurs and how many of them pass on their knowledge to students. I-Corps is our secret weapon to integrate American innovation and entrepreneurship into every U.S. university lab.
I thought it was worth sharing the progress report from the Bay Area (Berkeley, Stanford, UCSF) I-Corps node so you can see what just one of the nodes was accomplishing. Multiply this by the NSF regional nodes across the U.S. and you’ll have a feeling for the scale and breadth of the program.
If you can’t see the presentation above click here
Glad to a part of it.
The U.S. government has built an accelerator for scientists and engineers
It’s scaled across the U.S.
The program has taught ~300 teams
Balance between public/private partnerships
Listen to the podcast here
Download the podcast here
BTW, NCIIA is offering other accelerators and incubators a class to learn how to build their own versions of I-Corps here.
When I wrote Four Steps to the Epiphany and the Startup Owners Manual, I believed that Life Sciences startups didn’t need Customer Discovery. Heck how hard could it be? You invent a cure for cancer and then figure out where to put the bags of money. (In fact, for oncology, with a successful clinical trial, this is the case.)
In the real world a big pivot in life sciences far down the road of development is a very bad sign due to huge sunk costs. But pivoting early, before you raise and spend millions or tens of millions means potential disaster avoided.
Some of these pivots included changing their product/service once the team had a better of understanding of customer needs or changing their position in the value chain (became an OEM supplier to hospital suppliers rather than selling to doctors directly.) Other pivots involved moving from a platform technology to become a product supplier, moving from a therapeutic drug to a diagnostic or moving from a device that required a PMA to one that required a 510(k).
Some of these teams made even more radical changes. For example when one team found the right customer, they changed the core technology (the basis of their original idea!) used to serve those customers. Another team reordered their device’s feature set based on customer needs.
These findings convinced me that the class could transform how we thought about building life science startups. But there was one more piece of data that blew me away.
Control versus Experiment – 18% versus 60% For the last two and a half years, the teams that were part of the National Science Foundation Innovation Corps were those who wanted to learn how to commercialize their science, applied to join the program, fought to get in and went through a grueling three month program. Other scientists attempting to commercialize their science were free to pursue their startups without having to take the class.
Both of these groups, those who took the Innovation Corps class and those who didn’t, applied for government peer-reviewed funding through the SBIR program. The teams that skipped the class and pursued traditional methods of starting a company had an 18% success rate in receiving SBIR Phase I funding.
The teams that took the Lean Launchpad class – get ready for this – had a 60% success rate. And yes, while funding does not equal a successful company, it does mean these teams knew something about building a business the other teams did not.
The 3-person teams consisted of Principal Investigators (PI’s), mostly tenured professors (average age of 45,) whose NSF research the project was based on. The PI’s in turn selected one of their graduate students (average age of 30,) as the entrepreneurial lead. The PI and Entrepreneurial Lead were supported by a mentor (average age of 50,) with industry/startup experience.
This was most definitely not the hoodie and flip-flop crowd.
Obviously there’s lots of bias built into the data – those who volunteered might be the better teams, the peer reviewers might be selecting for what we taught, funding is no metric for successful science let alone successful companies, etc. – but the difference in funding success is over 300%.
The funding criteria for these new ventures wasn’t solely whether they had a innovative technology. It was whether the teams understood how to take that idea/invention/patent and transform it into a company. It was whether after meeting with partners and regulators, they had a plan to deal with the intensifying regulatory environment. It was whether after talking to manufacturing partners and clinicians, they understood how they were going to reduce technology risk. And It was after they talked to patients, providers and payers whether they understood the customer segments to reduce market risk by having found product/market fit.
Scientists and researchers have spent their careers testing hypotheses inside their labs. This class teaches them how to test the critical hypotheses that turn their idea into a business as they deal with the real world of regulation, customers and funding.
So after the team at UCSF said they’d like to prototype a class for Life Sciences, I agreed.
The goal of the Lean LaunchPad Life Sciences class at UCSF is to teach researchers how to move their technology from an academic lab into the commercial world.
We’re going to help teams:
assess regulatory risk before they design and build
gather data essential to customer purchases before doing the science
define clinical utility now, before spending millions of dollars
identify financing vehicles before you need them
We’ve segmented the class into four cohorts: therapeutics, diagnostics, devices and digital health. And we recruited a team of world class Venture Capitalists and entrepreneurs to teach and mentor the class including Alan May, Karl Handelsman, Abhas Gupta, and Todd Morrill.
The course is free to UCSF, Berkeley, and Stanford students; $100 for pre-revenue startups; and $300 for industry. – See more here
What if we could increase productivity and stave the capital flight by helping Life Sciences startups build their companies more efficiently?
We’re going to test this hypothesis by teaching a Lean LaunchPad class for Life Sciences and Healthcare (therapeutics, diagnostics, devices and digital health) this October at UCSF with a team of veteran venture capitalists.
In this three post series, Part 1 described the challenges Life Science companies face in Therapeutics and Diagnostics. This post describes the issues in Medical Devices and Digital Health. Part 3 will offer our hypothesis about how to change the dynamics of the Life Sciences industry with a different approach to commercialization of research and innovation. And why you ought to take this class.
Medical devices prevent, treat, mitigate, or cure disease by physical, mechanical, or thermal means (in contrast to drugs, which act on the body through pharmacological, metabolic or immunological means). They span they gamut from tongue depressors and bedpans to complex programmable pacemakers and laser surgical devices. They also diagnostic products, test kits, ultrasound products, x-ray machines and medical lasers.
Incremental advances are driven by the existing medical device companies, while truly innovative devices often come from doctors and academia. One would think that designing a medical device would be a simple engineering problem, and startups would be emerging right and left. The truth is that today it’s tough to get a medical device startup funded.
Class I devices are low risk and have the least regulatory controls. For example, dental floss, tongue depressors, arm slings, and hand-held surgical instruments are classified as Class I devices. Most Class I devices are exempt Premarket Notification 510(k) (see below.)
Class II devices are higher risk devices and have more regulations to prove the device’s safety and effectiveness. For example, condoms, x-ray systems, gas analyzers, pumps, and surgical drapes are classified as Class II devices.
Manufacturers introducing Class II medical devices must submit what’s called a 510(k) to the FDA. The 510(k) identifies your medical device and compares it to an existing medical device (which the FDA calls a “predicate” device) to demonstrate that your device is substantially equivalent and at least as safe and effective.
Class III devices are generally the highest risk devices and must be approved by the FDA before they are marketed. For example, implantable devices (devices made to replace/support or enhance part of your body) such as defibrillators, pacemakers, artificial hips, knees, and replacement heart valves are classified as Class III devices. Class III medical devices that are high risk or novel devices for which no “predicate device” exist require clinical trials of the medical device a PMA (Pre-Market Approval).
The FDA is tougher about approving innovative new medical devices. The number of 510(k)s being required to supply additional information has doubled in the last decade.
The number of PMA’s that have received a major deficiency letter has also doubled.
An FDA delay or clinical challenge is increasingly fatal to Life Science startups, where investors now choose to walk away rather than escalate the effort required to reach approval.
Business Model Issues
Cost pressures are unrelenting in every sector, with pressure on prices and margins continuing to increase.
Devices are a five-sided market: patient, physician, provider, payer and regulator. Startups need to understand all sides of the market long before they ever consider selling a product.
In the last decade, most device startups took their devices overseas for clinical trials and first getting EU versus FDA approval
Recently, the financing of innovation in medical devices has collapsed even further with most Class III devices simply unfundable.
Companies must pay a medical device excise tax of 2.3% on medical device revenues, regardless of profitability delays or cash-flow breakeven.
The U.S. government is the leading payer for most of health care, and under ObamaCare the government’s role in reimbursing for medical technology will increase. Yet two-thirds of all requests for reimbursement are denied today, and what gets reimbursed, for how much, and in what timeframe, are big unknowns for new device companies.
Venture Capital Issues
Early stage Venture Capital for medical device startups has dried up. The amount of capital being invested in new device companies is at an 11 year low.
Because device IPOs are rare, and M&A is much tougher, liquidity for investors is hard to find.
Exits have remained within about the same, while the cost and time to exit have doubled.
Life Sciences III – The Rise of Digital Health Over the last five years a series of applications that fall under the category of “Digital Health” has emerged. Examples of these applications include: remote patient monitoring, analytics/big data (aggregation and analysis of clinical, administrative or economic data), hospital administration (software tools to run a hospital), electronic health records (clinical data capture), and wellness (improve/monitor health of individuals). A good number of these applications are using Smartphones as their platform.
Business Model Issues
A good percentage of these startups are founded by teams with strong technical experience but without healthcare experience. Yet healthcare has its own unique regulatory and reimbursement issues and business model issues that must be understood
Most of these startups are in a multisided market, and many have the same five-sided complexity as medical devices: patient, physician, provider, payer and regulator. (Some are even more complex in an outpatient / nurse / physical therapy setting.)
Reimbursement for digital health interventions is still a work in progress
Some startups in this field are actually beginning with Customer Development while others struggle with the classic execution versus search problem
Digital Health covers a broad spectrum of products, unless the founders have domain experience startups in this area usually discover the FDA and the 510(k) process later than they should.
Seed funding is still scarce for Digital Health, but a number of startups (particularly those making physical personal heath tracking devices) are turning to crowdfunding.
Moreover, the absence of recent IPOs and public companies benchmarks creates uncertainty for VCs evaluating later investments too
Try Something New The fact that the status quo for Life Sciences is not working is not a new revelation. Lots of smart people are running experiments in search of ways to commercialize basic research more efficiently.
Universities have set up translational R&D centers; (basically university/company partnerships to commercialize research). The National Institute of Health (NIH) is also setting up translational centers through its NCATS program. Drug companies have tried to take research directly out of university labs by licensing patents, but once inside Pharma’s research labs, these projects get lost in the bureaucracy. Realizing that this is not optimal, drug companies are trying to incubate projects directly with universities and the researchers who invented the technology, such as the recent Janssen Labs program.
But while these are all great programs, they are likely to fail to deliver on their promise. The assumption that the pursuit of drugs, diagnostics, devices and digital health is all about the execution of the science is in most cases a mistake.
The gap between the development of intriguing but unproven innovations, and the investment to commercialize those innovations is characterized as “the Valley of Death.”
We believe we need a new model to attract private investment capital to fuel the commercialization of clinical solutions to todays major healthcare problems that is in many ways technology agnostic. We need a “Needs Driven/Business Model Driven” approach to solving the problems facing all the stakeholders in the vast healthcare system.
We believe we can reduce the technological, regulatory and market risks for early-stage life science and healthcare ventures, and we can do it by teaching founding teams how to build new ventures with Evidence-Based Entrepreneurship.
Part 3 in the next post will offer our hypothesis how to change the dynamics of the Life Sciences industry with a different approach to commercialization of research and innovation. And why you ought to take this class.
It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness, it was the epoch of belief, it was the epoch of incredulity, it was the season of Light, it was the season of Darkness, it was the spring of hope, it was the winter of despair, we had everything before us, we had nothing before us, we were all going direct to Heaven, we were all going direct the other way.
Life Science (therapeutics- drugs to cure or manage diseases, diagnostics- tests and devices to find diseases, devices to cure and monitor diseases; and digital health –health care hardware, software and mobile devices and applications streamline and democratize the healthcare delivery system) is in the midst of a perfect storm of decreasing productivity, increasing regulation and the flight of venture capital.
But what if we could increase productivity and stave the capital flight by helping Life Sciences startups build their companies more efficiently?
It was the best of times and the worst of times The last 60 years has seen remarkable breakthroughs in what we know about the biology underlying diseases and the science and engineering of developing commercial drug development and medical devices that improve and save lives. Turning basic science discoveries into drugs and devices seemed to be occurring at an ever increasing rate.
Yet during those same 60 years, rather than decreasing, the cost of getting a new drug approved by the FDA has increased 80 fold. Yep, it cost 80 times more to get a successful drug developed and approved today than it did 60 years ago.
75% or more of all the funds needed by a Life Science startup will be spent on clinical trials and regulatory approval. Pharma companies are staggering under the costs. And medical device innovation in the U.S. has gone offshore primarily due to the toughened regulatory environment.
At the same time, Venture Capital, which had viewed therapeutics, diagnostics and medical devices as hot places to invest, is fleeing the field. In the last six years half the VC’s in the space have disappeared, unable to raise new funds, and the number of biotech and device startups getting first round financing has dropped by half. For exits, acquisitions are the rule and IPOs the exception.
While the time, expense and difficulty to exit has soared in Life Sciences, all three critical factors have been cut by orders of magnitude in other investment sectors such as internet or social-local-mobile. And while the vast majority of Life Science exits remain below $125M, other sectors have seen exit valuations soar. It has gotten so bad that pension funds and other institutional investors in venture capital funds have told these funds to stay away from Life Science – or at the least, early stage Life Science.
WTF is going on? And how can we change those numbers and reverse those trends?
We believe we have a small part of the answer. And we are going to run an experiment to test it this fall at UCSF.
In this three post series, the first two posts are a short summary of the complex challenges Life Science companies face; in Therapeutics and Diagnostics in this post and in Medical Devices and Digital Health in Part 2. Part 3 explains our hypothesis about how to change the dynamics of the Life Sciences industry with a different approach to commercialization of research and innovation. And why you ought to take this class.
Life Sciences I—Therapeutics and Diagnostics
It was the Age of Wisdom – Drug Discovery There are two types of drugs. The first, called small molecules (also referred to as New Molecular Entities or NMEs), are the bases for classic drugs such as aspirin, statins or high blood pressure medicines. Small molecules are made by reactions between different organic and/or inorganic chemicals. In the last decade computers and synthesis methods in research laboratories enable chemists to test a series of reaction mixtures in parallel (with wet lab analyses still the gold standard.) Using high-throughput screening to search for small molecules, which can be a starting point (or lead compound) for a new drug, scientists can test thousands of candidate molecules against a database of millions in their libraries.
The second class of drugs created by biotechnology is called biologics (also referred to as New Biological Entities or NBEs.) In contrast to small molecule drugs that are chemically synthesized, most biologics are proteins, nucleic acids or cells and tissues. Biologics can be made from human, animal, or microorganisms – or produced by recombinant DNA technology. Examples of biologics include: vaccines, cell or gene therapies, therapeutic protein hormones, cytokines, tissue growth factors, and monoclonal antibodies.
It was the Season of Light The drug development pipeline for both small molecules and biologics can take 10-15 years and cost a billion dollars. The current process starts with testing thousands of compounds which will in the end, produce a single drug.
The problem is that the probability that a small molecule drug gets through clinical trials is unchanged after 50 years. In spite of the substantial scientific advances and increased investment, over the last 20 years the FDA has approved an average of 23 new drugs a year. (To be fair, this is indication-dependent. For example, in oncology, things have gotten significantly better. In most other areas, particularly drugs for the central nervous system and metabolism, they have not.)
It was the Season of Despair With the exception of targeted therapies, the science and tools haven’t made the drug discovery pipeline more efficient. Oops.
There are lots of reasons why this has happened.
Regulatory and Reimbursement Issues
Drug safety is a high priority for the FDA. To avoid problems like Vioxx, Bexxar etc., the regulatory barriers (i.e. proof of safety) are huge, expensive, and take lots of time. That means the FDA has gotten tougher, requiring more clinical trials, and the stack of regulatory paperwork has gotten higher.
Additional trials to demonstrate both clinical efficacy (if not superiority) and cost outcomes effectiveness are further driving up the cost, time and complexity of clinical trials.
In a perfect world the goal is to develop a drug that will go after a single target (a protein, enzyme, DNA/RNA, etc. that will undergo a specific interaction with chemicals or biological drugs) that is linked to a disease.
To get FDA approval new drugs have to be proven better than existing ones. Most of the low-hanging fruit of easy drugs to develop are already on the market.
Venture Capital Issues
For the last two decades, biotech venture capital and corporate R&D threw dollars into interesting science (find a new target, publish a paper in Science, Nature or Cell, get funded.) The belief was that once a new target was found, finding a drug was a technologyexecution problem. And all the new tools would accelerate the process. It often didn’t turn out that way, although there are important exceptions.
Moreover, the prospect of the FDA also evaluating drugs for their cost-effectiveness is adding another dimension of uncertainty as the market opportunity at the end of the funnel needs to be large enough to justify venture investment
In Part 2 of this series, we describe the challenges new Medical Device and Digital Health companies face. Part 3 will offer our hypothesis how to change the dynamics of the Life Sciences industry with a different approach to commercialization of research and innovation in this sector. And why you ought to take this class.
Listen to the post here
The greatest number of jobs is created when startups create a new market – one where the product or service never existed before or is radically more convenient. Yet this is where startups will run into anti-innovation opponents they may not expect. These opponents have their own name – “rent seekers” – the landlords of the status-quo.
Smart startups prepare to face off against rent seekers and map out creative strategies for doing so…. First, however, they need to understand what a rent seeker is and how they operate…
Recently, the New York and North Carolina legislatures considered a new law written by Auto Dealer lobbyists that would make it illegal for Tesla to sell cars directly to consumers. This got me thinking about the legal obstacles that face innovators with new business models.
While Tesla, Lyft, Uber, Airbnb, et al are in very different industries, they have two things in common: 1) they’re disruptive business models creating new markets and upsetting the status quo and 2) the legal obstacles confronting them weren’t from direct competitors, but from groups commonly referred to as “rent seekers.”
Rent Seekers Rent seekers are individuals or organizations that have succeeded with existing business models and look to the government and regulators as their first line of defense against innovative competition. They use government regulation and lawsuits to keep out new entrants with more innovative business models. They use every argument from public safety to lack of quality or loss of jobs to lobby against the new entrants. Rent seekers spend money to increase their share of an existing market instead of creating new products or markets. The key idea is that rent seeking behavior creates nothing of value.
These barriers to new innovative entrants are called economic rent. Examples of economic rent include state automobile franchise laws, taxi medallion laws, limits on charter schools, auto, steel or sugar tariffs, patent trolls, bribery of government officials, corruption and regulatory capture. They’re all part of the same pattern – they add no value to the economy and prevent innovation from reaching the consumer.
No regulation? Not all government regulation is rent or rent seeking. Not all economic rents are bad. Patents for example, provide protection for a limited time only, to allow businesses to recoup R&D expenses as well as make a profit that would often not be possible if completely free competition were allowed immediately upon a products’ release. But patent trolls emerged as rent seekers by using patents as legalized extortion of companies.
How do Rent Seekers win? Instead of offering better products or better service at lower prices, rent seekers hire lawyers and lobbyists to influence politicians and regulators to pass laws, write regulations and collect taxes that block competition. The process of getting the government to give out these favors is rent-seeking.
Lobbyists also work through regulatory bodies like FCC, SEC, FTC, Public Utility, Taxi, or Insurance Commissions, School Boards, etc. Although most regulatory bodies are initially set up to protect the public’s health and safety, or to provide an equal playing field, over time the very people they’re supposed to regulate capture the regulatory agencies. Rent Seekers take advantage of regulatory capture to protect their interests against the new innovators.
PayPal – Dodging Bullets PayPal consistently walked a fine line with regulators. Early on the company shutdown their commercial banking operation to avoid being labeled as a commercial bank and burdened by banks’ federal regulations. PayPal worried that complying with state-by-state laws for money transmission would also be too burdensome for a startup so they first tried to be classified as a chartered trust company to provide a benign regulatory cover, but failed. As the company grew larger, incumbent banks forced PayPal to register in each state. The banks lobbied regulators in Louisiana, New York, California, and Idaho and soon they were issuing injunctions forcing PayPal to delay their IPO. Ironically, once PayPal complied with state regulations by registering as a “money transmitter” on a state-by-state basis, it created a barrier to entry for future new entrants.
U.S. Auto Makers – Death by Rent Seeking The U.S. auto industry is a textbook case of rent seeking behavior. In 1981 unable to compete with the quality and price of Japanese cars, the domestic car companies convinced the U.S. government to restrict the import of “foreign” cars. The result? Americans paid an extra $5 billion for cars. Japan overcame these barriers by using their import quotas to ship high-end, high-margin luxury cars, establishing manufacturing plants in the U.S. for high-volume lower cost cars and by continuing to innovate. In contrast, U.S. car manufacturers raised prices, pocketed the profits, bought off the unions with unsustainable contracts, ran inefficient factories and stopped innovating. The bill came due two decades later as the American auto industry spiraled into bankruptcy and its market share plummeted from 75% in 1981 to 45% in 2012.
In these states it appears innovation be damned if it gets in the way of a rent seeker with a good lobbyist.
Much like Paypal, it’s likely that after forcing Tesla to win these state-by-state battles, the auto dealers will have found that they dealt themselves the losing hand.
Rent seeking is bad for the economy Rent seeking strangles innovation in its crib. When companies are protected from competition, they have little incentive to cut costs or to pay attention to changing customer needs. The resources invested in rent seeking are a form of economic waste and reduce the wealth of the overall economy.
Startups, investors and the public have done a poor job of calling out the politicians and regulators who use the words “innovation means jobs” while supporting rent seekers.
What does this mean for startups? In an existing market it’s clear who your competitors are. You compete for customers on performance, ease of use, or price. However, for startups creating a new market – one where either the product or service never existed before or the new option is radically more convenient for customers - the idea that rent seekers even exist may come as a shock. “Why would anyone not want a better x, y or z?” The answer is that if your startup threatens their jobs or profits, it doesn’t matter how much better life will be for consumers, students, etc. Well organized incumbents will fight if they perceive a threat to the status quo.
As a result disrupting the status quo in regulated market can be costly. On the other hand, being a private and small startup means you have less to lose when you challenge the incumbents.
If you’re a startup with a disruptive business model here’s what you need to do:
Map the order of battle
Laughing at the dinosaurs and saying, “They don’t get it” may put you out of business. Expect that existing organizations will defend their turf ferociously i.e. movie studios, telecom providers, teachers unions, etc.
Understand who has political and regulator influence and where they operate
Figure out an “under the radar” strategy which doesn’t attract incumbents lawsuits, regulations or laws when you have limited resources to fight back
Pick early markets where the rent seekers are weakest and scale
For example, pick target markets with no national or state lobbying influence. i.e. Craigslist versus newspapers, Netflix versus video rental chains, Amazon versus bookstores, etc.
Go after rapid scale of passionate consumers who value the disruption i.e. Uber and Airbnb, Tesla
AirBnb – Damn the torpedoes full speed ahead
For example, Airbnb, thrives even though almost all of its “hosts” are not paying local motel/hotel taxes nor paying tax on their income, and many hosts are violating local zoning laws. Some investors and competitors may be concerned about regulatory risk and liability. AirBNB’s attitude seems to be “build the business until someone stops me, and change or comply with regulations later.” This is the same approach that allowed Amazon to ignore local sales taxes for the last two decades.
When you get customer scale and raise a large financing round, take the battle to the incumbents. Strategies at this stage include:
Use competition among governments to your advantage, eg, if New York or North Carolina doesn’t want Tesla, put the store in New Jersey, across the river from Manhattan, increasing New Jersey’s tax revenue
Cut deals with the rent seekers. i.e. revenue/profit sharing, two-tier hiring, etc.
Buy them out i.e. guaranteed lifetime employment
Rent seekers are organizations that have lost the ability to innovate
Theylook to the government to provide their defense against innovation
“What’s gone and what’s past help
Should be past grief.”
William Shakespeare - The Winter’s Tale
We give abundant advice to founders about how to make startups succeed yet we offer few models about dealing with failure.
So here’s mine.
In my experience, living through failure has 6 stages:
Stage 1: Shock and Surprise
Stage 2: Denial
Stage 3: Anger and Blame
Stage 4: Depression
Stage 5: Acceptance
Stage 6: Insight and Change
While I had been part of a few failed startups, none of them had fallen squarely on my shoulders until Rocket Science Games where my business card said CEO. It was there that I lived through all 6 stages and came out the other side a changed man.
Stage 1: Shock and Surprise We raised $35 million and after 18 months made the cover of Wired magazine. The press called Rocket Science one of the hottest companies in Silicon Valley and predicted that our games would be great because the storyboards and trailers were spectacular. 90 days later, I found out our games are terrible, no one is buying them, our best engineers started leaving, and with 120 people and a huge burn rate, we’re running out of money and about to crash. This can’t be happening to me.
Stage 2: Deny any of it was your fault In my mind, I had done everything the investors asked me to do. I raised a ton of money and got a ton of press. We hired everyone according to our plan. It was everyone else who screwed up. I did everything right.
Stage 3: Get angry and blame everyone else This was the fault of my cofounder since he was in charge of game development, it was the engineers who bailed on me, it was the sales and marketing people who didn’t tell me how bad the games were, it was the VC’s who refused to put any more money in the company, it was Sega’s fault for making a bad gaming platform…
State 4: Get depressed When the inevitability and magnitude of the failure sunk in, I slept in a lot. There were days I’d get up late and go to bed again at 5 pm. I lost interest in anything associated with my past industry. (To this day I still can’t play a video game.)
Step 5: Gradually accept your role in the failure A few weeks after leaving, I began to think about what I should have done, could have done and pondered why I didn’t do it. (I didn’t listen, I didn’t act, I didn’t own my role as CEO, I wasn’t prepared to do what was right or leave.) This was hard and didn’t happen overnight. My wife was a great partner here. I often reverted to Stages 2 and 3, but over time I took ownership of my primary role in the debacle.
Stage 6: Gain insight and change your behavior This was the hardest part. While I stopped blaming others, understanding what I could change in my behavior took long months. It would have been much easier to just move on, but I was looking for the lessons that would make my next startup successful. I looked at the patterns of behavior, not just at my last company but also across my entire career. I learned how to dial back the hubris, get other smart people to work with me – rather than just for me, listen better, and act and do what was right – regardless of what others thought I should do.
Epilogue For my next startup I parked the behaviors that drove Rocket Science off the cliff. We established a team of founders who worked collaboratively. When my co-founders and I got the company scalable and repeatable, we hired an operating executive as the CEO and returned a billion dollars to each of our two lead investors.
Now when I listen to entrepreneurs who’ve cratered a company, I listen for their stories of failure and redemption.
Six stages of failure and redemption
Don’t get stuck in Stages 2, 3 or 4 - move forward
Don’t skip acceptance of your role
Get to insight so you can change your behavior—then commit to the challenge of doing it differently the next time
The U.S. has spent the last 70 years making massive investments in basic and applied research. Government funding of research started in World War II driven by the needs of the military for weapon systems to defeat Germany and Japan. Post WWII the responsibility for investing in research split between agencies focused on weapons development and space exploration (being completely customer-driven) and other agencies charted to fund basic and applied research in science and medicine (being driven by peer-review.)
The irony is that while the U.S. government has had a robust national science and technology policy, it lacks a national industrial policy; leaving that to private capital. This approach was successful when U.S. industry was aligned with manufacturing in the U.S., but became much less so in the last decade when the bottom-line drove industries offshore.
In lieu of the U.S. government’s role in setting investment policy, venture capital has set the direction for what new industries attract capital.
This series of blog posts is my attempt to understand how science and technology policy in the U.S. began, where the money goes and how it has affected innovation and entrepreneurship. In future posts I’ll offer some observations how we might rethink U.S. Science and National Industrial Policy as we face the realities of China and global competition.
Office of Scientific Research and Development – Scientists Against Time
As World War II approached, Vannevar Bush, the ex-dean of engineering at MIT, single-handledly reengineered the U.S. governments approach to science and warfare. Bush predicted that World War II would be the first war won or lost on the basis of advanced technology. In a major break from the past, Bush believed that scientists from academia could develop weapons faster and better if scientists were kept out of the military and instead worked in civilian-run weapons labs. There they would be tasked to develop military weapons systems and solve military problems to defeat Germany and Japan. (The weapons were then manufactured in volume by U.S. corporations.)
OSRD divided the wartime work into 19 “divisions”, 5 “committees,” and 2 “panels,” each solving a unique part of the military war effort. These efforts spanned an enormous range of tasks – the development of advanced electronics; radar, rockets, sonar, new weapons like proximity fuse, Napalm, the Bazooka and new drugs such as penicillin and cures for malaria.
The civilian scientists who headed the lab’s divisions, committees and panels were given wide autonomy to determine how to accomplish their tasks and organize their labs. Nearly 10,000 scientists and engineers received draft deferments to work in these labs.
One OSRD project – the Manhattan Project which led to the development of the atomic bomb – was so secret and important that it was spun off as a separate program. The University of California managed research and development of the bomb design lab at Los Alamos while the US Army managed the Los Alamos facilities and the overall administration of the project. The material to make the bombs – Plutonium and Uranium 235 – were made by civilian contractors at Hanford Washington and Oak Ridge Tennessee.
OSRD was essentially a wartime U.S. Department of Research and Development. Its director, Vannever Bush became in all but name the first presidential science advisor. Think of the OSRD as a combination of all of today’s U.S. national research organizations – the National Science Foundation (NSF), National Institute of Health (NIH), Centers for Disease Control (CDC), Department of Energy (DOE) and a good part of the Department of Defense (DOD) research organizations – all rolled into one uber wartime research organization.
OSRD’s impact on the war effort and the policy for technology was evident by the advanced weapons its labs developed, but its unintended consequence was the impact on American research universities and the U.S. economy that’s still being felt today.
National Funding of University Research
Universities were started with a mission to preserve and disseminate knowledge. By the late 19th century, U.S. universities added scientific and engineering research to their mission. However, prior to World War II corporations not universities did most of the research and development in the United States. Private companies spent 68% of U.S. R&D dollars while the U.S. Government spent 20% and universities and colleges accounted just for 9%, with most of this coming via endowments or foundations.
Before World War II, the U.S. government provided almost no funding for research inside universities. But with the war, almost overnight, government funding for U.S. universities skyrocketed. From 1941-1945, the OSRD spent $450 million dollars (equivalent to $5.5 billion today) on university research. MIT received $117 million ($1.4 billion in today’s dollars), Caltech $83 million (~$1 billion), Harvard and Columbia ~$30 million ($370 million.) Stanford was near the bottom of the list receiving $500,000 (~$6 million). While this was an enormous sum of money for universities, it’s worth putting in perspective that ~$2 billion was spent on the Manhattan project (equivalent to ~$25 billion today.)
World War II and OSRD funding permanently changed American research universities. By the time the war was over, almost 75% of government research and development dollars would be spent inside Universities. This tidal wave of research funds provided by the war would:
Establish a permanent role for U.S. government funding of university research, both basic and applied
Establish the U.S. government – not industry, foundations or internal funds – as the primary source of University research dollars
Establish a role for government funding for military weapons research inside of U.S. universities (See the blog posts on the Secret History of Silicon Valley here, and for a story about one of the University weapons labs here.)
Make U.S. universities a magnet for researchers from around the world
Give the U.S. the undisputed lead in a technology and innovation driven economy – until the rise of China.
The U.S. Nationalizes Research
As the war drew to a close, university scientists wanted the money to continue to flow but also wanted to end the government’s control over the content of research. That was the aim of Vannevar Bush’s 1945 report, Science: the Endless Frontier. Bush’s wartime experience convinced him that the U.S. should have a policy for science. His proposal was to create a single federal agency – the National Research Foundation – responsible for funding basic research in all areas, from medicine to weapons systems. He proposed that civilian scientists would run this agency in an equal partnership with government. The agency would have no laboratories of its own, but would instead contract research to university scientists who would be responsible for all basic and applied science research.
But it was not to be. After five years of post-war political infighting (1945-1950), the U.S. split up the functions of the OSRD. The military hated that civilians were in charge of weapons development. In 1946 responsibility for nuclear weapons went to the new Atomic Energy Commission (AEC). In 1947, responsibility for basic weapons systems research went to the Department of Defense (DOD). Medical researchers who had already had a pre-war National Institutes of Health chafed under the OSRD that lumped their medical research with radar and electronics, and lobbied to be once again associated with the NIH. In 1947 the responsibility for all U.S. biomedical and health research went back to the National Institutes of Health (NIH). Each of these independent research organizations would support a mix of basic and applied research as well as product development.
Finally in 1950, what was left of Vannevar Bush’s original vision – government support of basic science research in U.S. universities – became the charter of the National Science Foundation (NSF). (Basic research is science performed to find general physical and natural laws and to push back the frontiers of fundamental understanding. It’s done without thought of specific applications towards processes or products in mind. Applied research is systematic study to gain knowledge or understanding with specific products in mind.)
Despite the failure of Bush’s vision of a unified national research organization, government funds for university research would accelerate during the Cold War.
Coming in Part 2 – Cold War science and Cold War universities.
Large scale federal funding for U.S. science research started with the Office of Scientific Research and Development (OSRD) in 1940
Large scale federal funding for American research universities began with OSRD in 1940
In exchange for federal science funding, universities became partners in weapons systems research and development