Steve Blank enrolled in the wrong college, joined the Air Force, dropped out of the right college and found success in Silicon Valley . Now he's back in the classroom as a sought-after teacher and, on the side, is trying to help DOD get acquisition right through Hacking for Defense, I-Corps and other efforts. His advice to the Army: Get out of the building.

In 2017, Worth magazine compiled a list of "The 25 Most Important Entrepreneurs of the Past 25 Years." Among those listed were tech giants like Amazon's Jeff Bezos, Microsoft's Bill Gates, Apple's Steve Jobs, Bloomberg LP founder Michael Bloomberg, SpaceX and Tesla's Elon Musk and entertainer-entrepreneurs such as Oprah Winfrey, rapper Jay Z and Martha Stewart.

Also on the list was a not-so-familiar name: Steve Blank.

After a successful 21-year career as a "serial entrepreneur" in California's Silicon Valley, where he started or worked on eight startups, Blank retired in 1999 at age 45. He had stepped down as CEO of E.piphany--a software company he started in his living room--just before it went public. But it was after he retired that Blank's career really took off.

An interview with Blank is a sometimes laugh-out-loud funny roller coaster of anecdotes, ideas, insights, self-deprecation and occasional profanities. Amid his cheerful and easy camaraderie, it's easy to lose sight of the fact that you're speaking with one of the smartest people you've ever talked to, someone who radically changed the trajectory and culture of Silicon Valley.

Blank grew up in a lower-middle-class neighborhood in New York City. He first enrolled at Michigan State University (he'd seen the Michigan Wolverines play football on TV but applied to the wrong college) and then, after dropping out of school, enlisted in the Air Force. After receiving combat training, he deployed to Thailand in 1974, where he repaired electronic warfare equipment for the Wild Weasels--aircraft equipped with radar-seeking missiles. "You get deployed overseas, you had to go through combat training," Blank related in a Jan. 31 interview with Army AL&T. "And I thought it was funny because I was going to be in an air base doing electronics. And the guy training me said, 'You know, the base you are going to has been attacked a couple of times. I sure hope the guy attacking you also laughed through his training.' But yes, it was mostly screwdrivers and electronics."

He thrived in the war zone atmosphere, where the assignment was to fix broken equipment, and fast, with minimum direction and few rules. "I didn't understand until decades later: In a war zone, we were incredibly innovative, at least on my level--anything it required to get the mission done."

After Thailand he was assigned to a Strategic Air Command (SAC) base in Michigan, with nuclear-armed B-52 bombers, "where there was no deviation from the technical orders and technical manuals at all, for all the right reasons. So here was an example of agile in a war zone, and an example of executing known missions with no deviation on a SAC base." When people ask why the military can't be as agile as startups, "I remind them that on the battlefield, the military historically has been more agile than startups. But when the military comes back to peacetime, it collapses into the world's largest bureaucracy. But on the battlefield, it's find, fix and finish. … Given what [the U.S. military] was trying to achieve, it was incredibly innovative. It used all kinds of people who typically never worked together and did incredibly amazing stuff. … It's not that we don't come up with this stuff, it's that the bureaucracy tends to win when we get back home."

Blank left the Air Force and went back to college, this time to the right Michigan, the University of Michigan in Ann Arbor. But the emphasis on theory over practice led Blank to drop out again . He wouldn't set foot in a college classroom again for 25 years, when he began teaching entrepreneurship at Stanford University and the University of California, Berkeley.

In 1978, while installing process-control networks in automobile assembly plants and steel mills for one of the few startups in Ann Arbor, Blank and a colleague were sent to San Jose, California, to install a system in a Ford Motor plant. He was surprised to hear radio ads from a company called Intel Corp. seeking scientists, engineers and technicians. He was stunned when he picked up a copy of the local newspaper and found a 48-page classified ad section that was almost all want ads seeking scientists, engineers and technicians. After they finished their work at Ford, his colleague flew back to Michigan and Blank stayed in San Jose. He quickly got his first job in Silicon Valley--coincidentally for a startup founded by William Perry, who would go on to help revolutionize satellite reconnaissance and later become secretary of defense.

Twenty-one years and eight startups later, Blank retired. He began teaching entrepreneurship classes at Stanford and Berkeley (he also teaches now at Columbia University, where he is a senior fellow for entrepreneurship).

He had time to think about why some startups succeeded and others failed. He "realized--heretically, at the time--that we were just missing something really big," he told Army AL&T in an October 2016 interview. Entrepreneurs were focused on turning a technology into a product, building a company and then hoping to get to an initial public offering. Instead, most failed because they forgot to discover whether anyone wanted or needed their product.

What was missing, Blank realized, was the understanding that startups had almost nothing in common with large, successful companies. Large companies know their customers, what products those customers will pay for, their pricing and their competitors. Startups know none of that.

His "Lean Startup" method boiled down to three basic steps:
• Articulate your hypotheses. What problem are you trying to solve? Who's your customer? What solution do customers want to grab out of your hands? "Hypothesis is a fancy word for 'we're just guessing,' " Blank said. "I use the word 'hypothesis' at Stanford because students pay to be there and nobody wants to learn how to guess for $50,000 a year."
• Get out of the building. Talk to at least 100 potential customers and stakeholders about your hypotheses. Have you identified the problem correctly? Can you validate your hypotheses? "Some hypotheses could be verified within 10 minutes, or it might take months," Blank said.
• Build a minimum viable product (MVP)--the smallest thing that will get you the most learning at that point in time. It can be a wireframe, a PowerPoint, hardware, etc. Get customer feedback. If the feedback is good, refine and improve your MVP until it's ready to roll out as a finished product. If it's bad, figure out where you went wrong and change direction (called a pivot). This way, if you've failed, you've failed early and inexpensively.

In 2005 Blank wrote "The Four Steps to the Epiphany," which launched the Lean Startup movement. One of his students, Eric Ries, followed up in 2011 with "The Lean Startup." Also in 2011, Blank was contacted by the National Science Foundation, which was looking for a way for its scientists to turn their research into viable commercial products. He developed the Innovation Corps (I-Corps), which is now considered the standard for science commercialization in the U.S. and has been adopted by other government agencies including the National Institutes of Health. In 2012, Blank and Bob Dorf, a fellow serial entrepreneur , released "The Startup Owner's Manual," a step-by-step guide to building a successful startup. In May 2013, Blank's article "Why the Lean Start-Up Changes Everything" was on the cover of the Harvard Business Review, one of the nation's most respected business publications. The Lean movement had gone mainstream.

Now, as the 2018 U.S. National Defense Strategy seeks to change DOD's culture and policies to drive innovation at speed, exploring "streamlined, non-traditional pathways to bring critical skills into service, expanding access to outside expertise, and devising new public-private partnerships to work with small companies, startups, and universities," Blank has brought his ideas and expertise on innovation to DOD, where they are gaining traction.

Blank's introduction to defense acquisition came in 2011, when he met Peter Newell. (See "Emergency Insurgency,".) Newell is a retired Army colonel; his last command was the Rapid Equipping Force, where he sped off-the-shelf solutions to Soldiers on the ground in Afghanistan. After retiring, he joined a former Army Special Forces colonel, Joe Felter, in starting a consulting company in Palo Alto called BMNT. (Felter is now the deputy assistant secretary of defense for South and Southeast Asia.)

"I teach a set of classes at Stanford, and one of my students was an ex-Delta Force operator who said, 'Hey, your methodology sounds a lot like the Army's Rapid Equipping Force and Pete Newell.' And I said, 'Who the heck is Pete?' " Blank said. At what the two men thought would be a quick meet-and-greet, "Pete described what he did with the Rapid Equipping Force, and I described what we did with the Lean Startup methodology and turning the federal research agencies on to I-Corps. And as he's drawing his diagram and I'm drawing mine, we discovered we basically came up with the same methodology, one that actually works from the battlefield to the boardroom. It's a big idea. Same methodology--we were just using different diagrams. His actually got deployed where lives were dependent, and mine got deployed on the cover of the Harvard Business Review."

Blank, Newell and Felter developed and taught Hacking for Defense (H4D) at Stanford, a class that unleashed teams of students on unclassified, real-world problems from DOD. As of 2018, the class has been taught at Stanford and 10 other universities nationwide. Twenty-four government agencies, including DOD, the U.S. Department of Homeland Security, the U.S. Department of State, the U.S. Department of Energy, NASA and the intelligence community, participated by providing students with real challenges to solve. A sister class, Hacking for Diplomacy, was launched at Stanford in fall 2016. A series of other classes, including Hacking for Energy and Hacking for Impact, also have been developed. Today the trio just label the program H4X, where X can mean any subject. The classes, Blank explained, create a new platform for national service, a way to expose students to parts of the U.S. government where a traditional academic path or business career would never take them.

The more he delved into DOD problems, the more Blank realized that innovation is vastly different in business versus government. "In a startup, innovation creates new products or services that people want to buy that never existed before. In an existing company, innovation can be new or more likely can improve existing products," he said. "But I'll contend innovation in government is quite different."

For DOD and intelligence agencies, innovation needs to be continuous, as DOD and intelligence agencies face what Blank calls "the Red Queen problem."

"In 'Alice in Wonderland,' they remind Alice that she has to run twice as fast just to stay in place. … So innovation in an intelligence community has two components. Think of one as replenishment. Just like we replenish ammunition, we need to replenish innovation." That, Blank said, gets DOD and the intelligence community on an even footing with potential adversaries. "The second reason is why you do innovation: To get ahead of your adversaries."

Hasn't this always been the case? Hasn't the U.S. constantly had to adapt to emerging threats? "In the 21st century, the rate of disruption is now exponential," Blank said. "In the 20th century, we had a single adversary--the Soviet Union--which was kind of innovating at our speed. Every once in a while we would do an offset strategy and they would do something else that was offset, but the clock speed was relatively simple. Yet in the 21st century we don't have one adversary; you need a scorecard just to figure it out. You just can't physically hire enough people and deploy enough weapon systems, because there are a variety of threats, everything from IEDs [improvised explosive devices] in Afghanistan to anti-access and area-denial stuff in the western Pacific for carriers. And so the question is, is this an equipment problem or is it something else? Is the third offset machine intelligence and robotics or is it something else? … It's not just an acquisition problem, it's understanding that the nature of the threat and the speed that we need to respond have changed dramatically."

Understanding the nature of the threat--knowing exactly the problem that you're trying to solve--is the foundation of Blank's Lean Startup method. "Lean methodology is not just rapid," he said. "It is not just cheap. It is not just fast. It requires deep understanding of, 'What problem are we solving? And are we actually fixing a problem or a symptom of a problem?' "

Newell, Blank said, uses the example of the difficulty of providing energy and water to remote forward operating bases in Afghanistan, which in many cases required C-17 cargo planes to airdrop fuel and water, or running convoys to the outposts amid the constant threat of ambush or IED attack. "How many men were dying to provide fuel for the equipment generators and other stuff we needed at these outposts? How much human capital and military assets were consumed trying to protect them in the first place?" Blank asked.

"I would have looked at this as a forward operating base fuel problem," Blank said. "But Pete said, 'No, no, no. It's a long-tail supply problem. Do you know how many tens of thousands of gallons of gas and other things we are using just recharging batteries and running radios?' He said if you don't understand a problem and its consequences, you end up building the wrong solution. The first solution was, 'Oh, why don't we use remote drones to drop fuel to these bases?' Yes, but once you understand the deeper problems, well, can we eliminate most of the fuel we need and just do away with half of the resupply requirements? Oh, well, gee, I was kind of excited about the drones, because that is a neat thing for a prime to build.

"Once you really understood the problem, you realized, 'Oh, what we ought to have is remote power sources run by solar cells and more efficient generators while also finding ways to recycle water.' Because everything you could save out there actually saves part of that whole supply chain."

Not everything, Blank acknowledges, needs to be lean. "I am not an expert in government requirements and acquisition, but it is even clear to me that there are some things that can be specified and contracted, just like we're doing. Not everything needs to be agile. We need to ask ourselves: 'Are we going for the 100 percent quality and perfection, and time and cost are not issues?' If so, let's use standard contracts. We know what a new pistol looks like--I need a pistol that spits out bullets. That is not an unknown thing. For God's sake, let's just spec this thing. And by the way, it shouldn't take 300 pages just to say give me more of the last stuff I just had. That's fine. And there are a ton of things that works on."

But the current defense acquisition model is broken, Blank said, "which people have been talking about, I think, since Washington got his boats on the Delaware." The problem lies at the beginning, with requirements that stifle innovation, and at the end, with prime contractors, he said.

"What is really broken is requirements. … The methodology for problem understanding is just fundamentally flawed. DOD is over-optimized for perfect performance at the expense of providing timely capabilities to the warfighter .

"And that's the part that we [Lean Startup] have gotten right: There are no facts inside the building, so get outside, and do that with speed and urgency. … And that changes our 20th-century philosophy … that we'll build things with every possible feature and we'll spend a decade doing it because our primes want to make the most money. And that runs into the prime problem."

The bulk of the current acquisition system, Blank said, is built around a waterfall requirements and acquisition process--a sequential process, with little iteration or learning. Instead, development flows steadily downward (like a waterfall)--rather than an agile system that promotes innovation.

"So how do things get built when they get built by contractors? … The word 'contract' implicitly or explicitly says we will specify all the features up front because we are going to do a great job on requirements, and you will develop it in a waterfall process and you will deliver the product," he said.

Under such a system, "it is in the contractor's interest to make the contract last as long as possible," Blank said. "… That is the antithesis of lean. It's as far from lean as you could get. Lean says no, no, no. We have a series of hypotheses on day one, but when we get out of the building … it's the notion of, we don't really know what problem we're solving. We think we do, so let's get started. But we can't spec every possible feature. So instead of waterfall engineering, we need to learn how to write contracts for agile engineering."

Until something happens to encourage defense prime contractors to focus on speed of delivery, continuous adaptation and frequent modular upgrades, Blank said, "you are not going to fix the problem."

DOD's innovation pipeline--the process from which an idea is turned into a battlefield capability--has been shrinking for decades, Blank said, while tech innovation in the private sector has grown exponentially. "Venture capital is funding AI [artificial intelligence], robotics, drones and the startup ecosystem at $70 billion a year. And very little of it is pointed to the DOD. So the question is, how can you build a wider innovation funnel that captures more than just the primes? (See "Innovation Pipeline for Success," .) And most military organizations make it incredibly difficult to work with civilians not just on the contracting side, but on the security side--almost impossible.

"In the 20th century, DOD used to own all of the innovation technology. Everything that was important was owned by DOD and the intel community. So the biggest thing that's happened to the military is that all of these technologies that used to be owned and controlled and budgeted by DOD and the primes got away. For example, NSA [the National Security Agency] used to own crypto hardware. It turns out not only did crypto go commercial, you don't need hardware anymore. You can do crypto in software. Well, we built this entire expertise about hardware. Oops.

"So our problems should have been easier, but in fact we made them harder because we still acted like, not only did we own it, but we were incapable of figuring out how to work with these people and encouraging them to build dual-use products. And again, because our primes had no interest in doing that."

Blank sees the Small Business Innovation Research (SBIR) program as a possible model for changing the way prime contractors do business. SBIR, which began in 1982, provides funding for small companies to do research and development (R&D) on U.S. government priorities. It's funded by allocating about 3 percent of the R&D budgets of 11 federal agencies.

"The problem is, for 30 years we were essentially giving out cars without requiring driver's ed, meaning most of these things would fail as commercialized technology. People didn't know how to start companies," Blank said. The I-Corps program he co-created for the National Science Foundation in 2011 has changed that. "It's taught in 81 universities. If you want to get an SBIR grant, it's kind of mandatory."

Could something similar work for defense acquisition? "Today, the DOD version of the SBIR is simply a 'set-aside' program. At DOD, awards are managed as contracts. This means that deliverables are negotiated up front before the award is made. Imagine if there were a way to make the prime the 'innovation conduit' to help translate a new capability 'through the wall.' That would be cool. 'Lockheed, your job is to fully deploy 10 new capabilities per year in this arena, but they must be externally sourced and you can own no more than 10 percent of any single firm whose solution you deploy.' "

There are no easy fixes, Blank said. "One of the biggest observations over the last couple of years--and I hope the [Army futures command] doesn't fall prey to this--is that DOD looked at startups and said, 'Oh, let's adopt a good number of the things they do. They have incubators, let's have incubators. They have accelerators, let's do that. They have hackathons, so let's do that.' … What they have created is a whole set of activities that essentially mimicked the activities of startups. But what they didn't realize is that they have created disconnected activities, none of which resulted in deployable things to the battlefield. They didn't build an innovation pipeline; they've built disconnected activities. And by the way, it's the same problem that corporate innovation is facing.

"What we lack is an innovation pipeline that is parallel to our requirements and acquisition pipeline, with data, rigor and evidence. Instead, what we mostly have are lanyards and coffee cups. We really weren't and haven't been getting much out the other end."

It's vital to remember where innovation comes from, Blank said. "Innovation tends to occur at the edges first. … All of this innovation stuff rarely happens from the center. It happens by crazy people, by outliers, and eventually gets adopted and becomes doctrine."

MR. MICHAEL BOLD provides contract support to the U.S. Army Acquisition Support Center. He is a writer/editor for Network Runners Inc., with more than 30 years of editing experience at newspapers, including the McClatchy Washington Bureau, The Sacramento Bee, the San Jose Mercury News, the Dallas Morning News and the Fort Worth Star-Telegram. He holds a B.J. in journalism from the University of Missouri.

This article will be published in the April -- June 2018 Army AL&T magazine.