Why the Government Isn’t a Bigger Version of a Startup
There was a time when much of U.S. academia was engaged in weapon systems research for the Defense Department and intelligence community. Some of the best and brightest wanted to work for defense contractors or corporate research and development labs. And the best startups spun out of Stanford were building components for weapon systems.
Indeed, Silicon Valley was born as a center for weapon systems development and its software and silicon helped end the Cold War.
During World War II the United States did something its adversaries did not; it enlisted professors and graduate students as civilians in 105 colleges and universities to build advanced weapon systems — nuclear weapons, radar, etc. After World War II, the military-academic relationship that was so effective against Germany and Japan mobilized to face the Soviet threat and almost every research university (Massachusetts Institute of Technology, Stanford, Caltech, Harvard, Columbia, Johns Hopkins, University of Michigan, University of Wisconsin, Cornell, University of Chicago, and many others) continued to engage in weapon systems research during the Cold War.
Unique among them was Stanford, which provost Fred Terman (the father of American electronic warfare and electronic intelligence in World War I) built as a center of excellence in microwaves and electronics. Rather than focus the university inward on research, Terman took the radical step of encouraging Stanford professors and graduate students to start companies applying engineering to pressing military problems. The companies they started in the 1950s and 60s were based on Stanford’s defense contacts and contracts — microwave components, electronic warfare, and intelligence systems, and then the first wave of semiconductor companies. As there was no venture capital, these early startups were funded by early sales to weapon systems prime contractors and subcontractors.
But this quarter-century relationship between the military and universities ended with a bang in 1969. In the middle of the Vietnam War, student riots protesting military research forced the end of classified work on most college campuses. One of the unintended consequences was that many of the academics went off to found a wave of startups selling their technology to the military. For example, at Stanford after student riots in April 1969 shut down the Applied Electronics Laboratory, James de Broekert ,who was building electronic intelligence satellites, left the university and co-founded three Silicon Valley military intelligence companies: Argo Systems, Signal Science, and Advent Systems.
Within a decade, the rise of venture capital in Silicon Valley enabled startups to find commercial customers rather than military ones. And from then on, innovation in semiconductors, supercomputers, and software would be driven by startups, not the government.
After 9/11, with the memories of the fall of the twin towers, this ecosystem of military, academic, corporate, and startup actors coalesced for the decade as U.S. companies felt a patriotic duty to help their country defeat a common enemy.
But the 2013 Snowden revelations damaged that tenuous relationship yet again. In hindsight the damage wasn’t the result of what the United States was doing, but over the Pentagon’s inability and unwillingness to own up to why it was doing it: After the intelligence failure of 9/11, security agencies overcompensated by widespread, warrantless datamining as well as electronic and telephonic surveillance, including on U.S. persons.
Without a clear explanation of why this had been done, startups, which were already being funded by ever-increasing pools of venture capital, abandoned cooperation with the Defense Department and focused on high returns on social media and commercial applications. The commercial applications of big data, machine learning, artificial intelligence, drones, robotics, cyber, quantum computing, and biotechnology are the core foundations on which the Pentagon needs to build the weapons of the 21st century. Yet the development of these advanced technologies is now being driven by commercial interests, not the Defense Department.
America’s adversaries understand this. China is tightly integrating its defense establishment with startups, companies, and academia in “military-civilian fusion.” Russia, Iran, and North Korea have also fused those activities.
Reconstituting the tightly connected military-academic-commercial ecosystem that the Defense Department once had requires the Pentagon to relearn skills it once had, overcoming decades of avoiding the political and social issues of what it takes to rally the nation against a common threat. Today, every government agency, service branch, and combatant command is adopting innovation activities (hackathons, design thinking classes, innovation workshops, et al.) to tap into the creativity of a new generation of soldier — born into a digital world, comfortable with technology, and willing to improve and enhance America’s ability to fight and win.
The Government Can’t Act Like a Startup
However, those activities are not enough. The government isn’t a bigger version of a startup and can’t act like a startup does. Innovation activities in government agencies most often result in innovation theater. While these activities shape and build culture, they don’t win wars, and rarely deliver shippable or deployable products.
Startups dream in years, plan in months, evaluate in weeks, and ship in days. At times this means startups operate at speeds so fast they appear to be a blur to government agencies. It’s not that these companies are smarter than Defense Department employees, but they operate with different philosophies, different product development methodologies, and with different constraints.
The table below summarizes a few of the salient differences. Some of the most important are the least obvious. Startups can do anything. They can break the law and apologize later (as Uber, Airbnb, and Tesla did), but a government official taking the same type of risks can go to jail.
Urgency and risk-taking in a startup are integral parts of the culture, felt by 100 percent of early-stage employees. The urgency servicemembers feel on the battlefield is felt by few in government agencies, and most often there are negative incentives for risk-taking. In a startup cluster (Silicon Valley, Beijing, Tel Aviv) a failed entrepreneur is known as “experienced.” In a government agency, they’re likely known as being out of a job.
Innovation at speed is a given at a startup but the exception in a government agency. Advances in commercial technologies are occurring at no less than two, and up to ten, times the speed of comparable Pentagon-developed or acquired systems. Some of the speed is simply due to development methodologies. Waterfall development is still used by most defense contractors, resulting in updates of systems measured in years. With Agile development, used by all startups, updates can occur in weeks or sometimes days, or even hours. Some of the speed differences are because commercial companies and academics face Darwinian competitive pressures for revenue or recognition. These force rapid technical advances in fields such as machine learning, artificial intelligence, robotics, big data, and analytics.
The very definition of a contractor implies a contract. And a government contract starts with fixed requirements that only change with contract modifications. That makes sense when the problem and solution are known. But when they are unknown the traditional methods of contracting fail. Startups recognize that when new circumstances arise, they can pivot — make substantive changes to their products without any new contracts.
Startup | Government | |
Goal | For investors the goal is liquidity. For founders it’s to change the world. | Mission achievement or promotion or job security. |
Scope of activities | Can be anything. | Only what is authorized by law. |
Sense of urgency | Felt by all. | Felt by few. |
Time frame/ Pressure | Limited by cash, need for maximum speed felt by all. | Most often no pressure on individuals. Operates at or below 1/3 of startup speed. |
Innovation team | 100% volunteer. | Often ~75% assigned and/or contractors. |
Time on team/task | Until completion. | Often fixed tour, typically 2–3 years. |
Motivation | Culture, financial rewards, shared mission. | No economic alignment results in teams less robust and adaptable. |
Funding | Risk capital (“angels” and venture capital). | Internal: Enhancement to existing projects comes from existing budgets, disruptive projects come from research funding. |
Number of potential customers | Consumers: 100-plus million spending $60 trillion. Businesses: 15,000-plus with >500 employees spending ~$15 trillion. |
Hundreds of acquisition sources. |
Time/barriers to revenue | Consumers: minutes to hours.Businesses: months.Finding product/market fit. | Often years. Enormous paperwork and specialized acquisition processes. |
Innovation at speed | Rapid agile development only limited by funding. Pivots integral to the process. | Limited by contract processes and funding. Pivots counter to established culture and contracting. (Though in a crisis and on a battlefield can be as rapid as a startup.) |
Development methods | Rapid learning and discovery method called Agile/Lean. Can be 10 times government speed. | Waterfall primarily. Slow movement to Agile/Lean. |
Requirements development | Formulate hypotheses about user needs, then get out into the field and test the hypotheses. | Fixed requirements. |
Prototypes | Test a series of incremental and iterative prototypes with beneficiaries and stakeholders. | Requires new contract. Not mandatory. |
Pivots | Can change features, users, size, weight, etc. without having to painfully recycle through new requirements and acquisition processes. | Requires new contract. |
Acquisition of external resources | Whatever it takes. | Rigid requirements and acquisition processes. TechFar and Other Transaction Authority for special cases. |
Failure | Risk-taking is integral part of the culture. | Failure is career-retarding (at times, rightly so, as lives depend on pristine execution.) For major projects there is the “fishbowl management” effect. |
Who it impacts and cost of failure | Founders, investors, employees. End of company, loss of investment. | For small and mid-size programs, not felt by the program manager; impact may not be felt for months or years. Mission failure means people killed and personnel reassignment. |
Existing contractors have learned how to master the arcane defense acquisition system and live with the slow decision-making and payment processes. In some agencies, large contractors seem to “own” sections, offices, organizations, or programs. Often former government employees, at the level of GS-15 and below, will leave as staffers and return the next day working for large Beltway contractors, working or managing the same programs they previously worked. This relationship between government agency and contractor further impedes and often rejects innovation or disruption. Officers know they will likely lose their post-retirement future if they seek radical change.
This symbiotic relationship between government agencies and incumbent contractors is also a barrier to new entrants, in particular to startups with the very technologies the Department of Defense now needs. While the Pentagon has made efforts to reform the process (Other Transaction Authorities, TechFAR, mid-tier contracting, accelerators…) there is still a fundamental misunderstanding of what financial incentives would attract the best and brightest investors to guide their companies to work with the Defense Department. There are no incentives for prime contractors to invest in new ventures and none to acquire new ventures. And there is no plan for how to rapidly insert and deploy startup technologies into weapon systems.
So, the question is: What’s next? How do leaders in government think about and organize innovation in a way that makes a difference?
The answer is that, yes, government agencies need to be more agile. And yes, they need to fix the systemic internal issues that hinder their own innovators’ contributions. But, in addition, what they are missing is a comprehensive plan to build a 21st-century defense innovation ecosystem — reintegrating the military, academia, and private enterprise. To harness both their own internal innovators and this new external ecosystem the Defense Department needs what I call an innovation doctrine to organize their efforts to rapidly access and mobilize talent and technology. The Pentagon can build a mindset, culture, and process to fix this. This doctrine would let the country again capture the untapped power and passion of the best and brightest to leapfrog adversaries and win wars.
Entrepreneur-turned-educator Steve Blank is credited with launching the Lean Startup movement. He’s changed how startups are built; how entrepreneurship is taught; how science is commercialized; and how companies and the government innovate. Steve is the author of The Four Steps to the Epiphany and The Startup Owner’s Manual. His May 2013 Harvard Business Review cover story defined the Lean Startup movement. He teaches at Stanford and Columbia, where he is a Senior Fellow for Entrepreneurship; and created the National Science Foundation Innovation Corps — now the standard for science commercialization in the United States. His national Hacking for Defense class is revolutionizing how the U.S. defense and intelligence community can deploy innovation with speed and urgency.
Image: Department of Defense (Photo by HackerOne)