THE LOONSHOTS LETTER

How to win the next war

In this letter:

Why good teams kill great ideas, and so many innovation labs fail
The beautiful baby problem
Inventor vs. champion
Changing structure rather than culture

How to win the next war

The quote at the top is from Kill Chain, by Christian Brose, which anyone who cares about national security should read. Brose was staff director of the Senate Armed Services Committee from 2015 to 2018 and before that the principal national security advisor to John McCain.

Brose and others have made a strong case that over the past two decades, while the US focused on overmatched battles (Libya, Iraq, non-state actors) and grew complacent with its single-superpower status, our near-peer competitors, Russia and China, focused on the US. They studied our methods and legacy systems, identified weaknesses, and developed a new type of warfare – combining speed, cyber, and new technologies – designed to exploit those vulnerabilities (the recent FireEye and SolarWind breaches are just one example). We’re racing to catch up.

The rise of peer-level competitors, by itself, is not only common and natural, it can be an advantage. Single superpowers are historical anomalies. Productive trade strengthens rather than weakens lead powers. (China, India, and the Islamic Empire dominated world trade and economy for over a thousand years.) But new competition creates new threats, which require new ideas. Those ideas may initially seem … crazy.

Which brings us to Loonshots. When senior defense leaders have reached out, they’ve been clear about their goal: their agencies and services need to innovate faster and better than ever before.

As most strategists know, the US began many of its 20^th century conflicts behind its adversaries in crucial technologies or strategies, yet managed to catch up just in time. That worked when battlefield advantage was driven primarily by hardware. But the source of battlefield advantage is now changing rapidly from hardware to software. What worked for the past hundred years, when innovation meant better planes, better ships, better guns, won’t work anymore, when innovation means better algorithms, better data, faster decisions. The pace is too fast for current organizational models to withstand.

An accelerating pace of competition has also overwhelmed many private sector companies. The rise of cloud services, social, and unprecedented access to capital have shrunk barriers to new entrants. Executive teams that I’ve met with have been asking the same questions as defense leaders: How can we design organizations to innovate at a pace and scale that exceeds anything that came before?

A first step, as I wrote for War on the Rocks, is to look to history, to the last time a very large organization, the US military, suddenly found itself far behind a dangerously underestimated adversary. That was eighty years ago, on the brink of world war. The Allies lagged Nazi Germany in crucial science and technologies including U-boats (submarines), missiles, jet aircraft, and nuclear fission.

Vannevar Bush, the dean of engineering at MIT, quit his job, talked his way into a meeting with President Franklin Roosevelt, and convinced him of a new idea. He wouldn’t try to change the culture of the military. Instead, with the president’s support, he created a new structure. As I and others have described, Bush’s efforts turned the course of the war.

It’s a new century, with new technologies, but the same problem. And it’s time, once again, to create a new structure.

I recently spoke about the need and a path forward to Stanford’s class on Technology, Innovation, and Modern Warfare. The class is taught by Steve Blank, the founder of the Lean Startup movement and the National Science Foundation, National Institutes of Health, and Hacking for Defense innovation programs; Raj Shah, former head of the Defense Innovation Unit; and Joe Felter, former US Army Special Forces and Director of West Point’s Combating Terrorism Center. Guest speakers have included former Sec Def Ash Carter, former Sec Def Jim Mattis, GEN. (ret) Jack Shanahan, and many others. Steve Blank’s annotated notes on their remarks are a treasure (as is his Secret History of Silicon Valley).

Bottom line:

Good teams will kill great ideas, no matter how smart the strategy, how enticing the technology, or how much leaders insist on innovation. That’s the adoption problem. If large teams don’t solve for that, in the public or private sector, they will fail. The adoption problem doomed IBM, Nokia, and nearly sunk the Allies in WWII. We have a chance, a limited window, to avoid the same mistakes.
The crux of the problem is the tension between the creatives working on new ideas (loonshots) and the soldiers focused on delivering on time, on budget, on spec (franchises). A special forces unit of highly-trained, neutral mediators can bridge the divide. Think of the complex machinery inside a car. To move, the car needs pistons and it needs wheels. But without oil it will not run at high speeds. Those special forces are that oil. They fight the internal battles organizations must win to survive rapid change. They solve the beautiful baby problem (video).

* * *

Highlights from comments to Stanford’s Technology, Innovation and Modern Warfare class (edited):

Invention versus innovation

I’m using the word invention deliberately — not innovation — because invention and innovation are different things. That difference lies at the heart of the problem with innovation inside so many organizations.

An invention is a new idea that works. For example, in the 1920s, when Robert Goddard showed that we could propel metal tubes by exploding liquid fuel inside them, he invented jet propulsion. That was a great invention. It didn’t become an innovation until it was developed and deployed at scale. In the case of jet propulsion, it wasn’t the U.S. that innovated. It was Nazi Germany with the V1 and the V2 missiles, and the Messerschmitt 262, the first jet aircraft.

For defense, innovation means deployed at scale. For the private sector, Jeff Bezos captured the distinction nicely: “Inventions are not disruptive. The only thing that’s disruptive is customer adoption.”

So what’s the core of the problem for national security organizations? What’s stopping them from innovating faster and better? Strategy is not the problem. The 2018 National Defense Strategy is one of those rare analyses that is clear, forceful, well regarded, and immediately actionable. Technology is not the problem. The military has 76 innovation labs and access to every technology hub in the country. Leadership is not the problem. Leaders across every service are pounding the table for innovation.

The military has put in place three of the four pieces of the puzzle: strategy, access to technology, leadership. The fourth, however, is missing. That’s organization design. Good teams will kill great ideas, no matter how smart the strategy, how enticing the technology, or how much leaders insist on innovation. Why that’s the case is a longer story, which I wrote about in Loonshots (video) but the bottom line is that we need to design our organizations to solve that problem —the adoption problem. If we don’t do that, we will lose.

Don’t be IBM

Here’s an example from the private sector: For close to 60 years, IBM dominated the IT industry. The industry was known as IBM and the Seven Dwarfs because IBM’s competitors were so far behind. If there was a superpower in any industry, it was IBM in IT. There were a couple of tiny competitors in the 80s who didn’t seem like much. A small company in Seattle called Microsoft. When they did their first partnership with IBM, they had just 32 employees. There was another little company in Santa Clara called Intel. They were running out of cash.

The story of a superpower ignoring distant threats from seemingly weak competitors should sound familiar. It’s US vs China.

At IBM, invention was not the problem. Many widely used technologies originated at IBM. Leadership was not the problem. Leaders at IBM have been pounding the table for innovation for years. But IBM today is one tenth the value of Microsoft. It’s half the value of Intel. Good teams kept killing great ideas. That’s the adoption problem.

Imagine a very large cake with sparkly candles on top and delicious nuts on the outside. It doesn’t matter how bright the leadership candles sparkle, or how tasty your technology nuts have made the icing, if your cake, on the inside, is dry and brittle.

You won’t solve the problem by adding more nuts or candles, just like IBM and DOD won’t solve the problem by adding more innovation labs or preaching more loudly about innovation. You need to change the recipe.

Getting that right, as I wrote in Loonshots, means solving for structure rather than culture (video, blog post). It’s the missing fourth leg, once you’ve got strategy straight, leadership aligned, and access to technology.

So what can we do?

Six problem patterns

Let’s start with six patterns I’ve observed across the service branches that block genuine innovation. These aren’t unique to the military; they’re the same inside many large companies.

1. A preference for big versus small. Bigger jets, bigger engines, bigger ships, as opposed to the small changes that can make an enormous difference…
2. A preference for product over strategy. For example, the tank was invented in the mid-1910s and used in World War I. But it wasn’t the technology, by itself, that allowed Nazi Germany to take over Western Europe in a matter of weeks. It was their strategy, the Blitz. Same with submarines. They were deployed in WWI. But it was the wolfpack strategy, developed by the German Navy, that nearly strangled the Allies during WWII.
3. A focus on technology as opposed to transfer. Leaders will invest a ton of time, energy, and money in acquiring sexy new technology. They’ll spend much less on navigating the internal barriers to adoption. By which I mean assuming that good ideas and technologies will win the day, by themselves, on their merits, and neglecting the often-hidden sources of internal resistance, agendas, misaligned incentives, and legacy stakes that kill those ideas.
4. A focus on prototyping as opposed to pretotyping (neologism from my friend Alberto Savoia, of Google and Stanford). Pretotyping is about what to do before the minimum viable product. How to test hypotheses incredibly fast. In one day for $100. Doing it well bakes into the system a preference for hypotheses rather than opinions; fast experiments rather than big plans; and testing ideas and strategies, not just products and technologies.
5. A focus on big plans rather than small experiments. Plans are comfortable because they can be debated and settled in committees, where no one person may be directly accountable. Many small experiments means lots of failures tied to directly accountable people. Yet it’s only through rapid experimentation that we quickly and efficiently explore the boundaries of what’s possible.
6. A focus on minimizing as opposed to maximizing risks. By which I mean maximizing the intelligent risk-taking you need to discover important breakthroughs.
  
  I see this all the time in mission-driven, as opposed to profit-driven, organizations; by mission-driven I mean hospitals and some nonprofits in addition to the military. When lives are at stake, there’s an enormous focus on reducing risk. You don’t want a lot of risk in your parachutes, or in your nuclear silos. But if you want to discover a new technology before your competitor, you want risk. You want to fail. A lot. If you only try things that don’t fail then you won’t discover the truly important breakthroughs, the ones where everyone gave up because they didn’t think it could be done.
  
  And who will discover them? Your adversary, who is taking those risks, who is working through the nine failures to get to the tenth iteration, the one that works. And you’ll see that tenth iteration when it’s too late. When it’s a bullet coming at your head.

Three solutions

1. Measurement. As all good leaders know, if you can’t measure it, you can’t manage it. Conversely, the things you measure well, with easily understood and visible metrics, tend to improve without much extra push. A senior officer at the Joint Chiefs of Staff told me, “We have no tangible way of knowing how we’re doing on innovation across the service branches. Absolutely no idea.”
  
  The absence of measurement is widespread at private sector companies as well. So many CEOs have told me, “We’re not doing well on innovation; we need to get better.” Well how do you know? And how will you know if you are getting better? So often there’s no answer … Good measurement requires tracking a carefully designed experiment matrix. One with simple – but not simplistic – categories…
2. Rewards. Here’s a quote from a major in the Air Force: “You get promoted in the Air Force by not screwing up. Trying something new means risking failure, scaring people around you, and therefore risking advancement. Do what the guy did before you and train those below you to do what you do.”
  
  What are the implications of a mindset where the top priority is to not screw up? How much risk will people take? I mean among the people that you want to be taking intelligent risks, the ones that you want to discover new technologies and strategies before your adversaries do? Do you want the top priority in their minds to be how to play it safe?
3. Special forces. We need a special forces for innovation – not another innovation lab, but a new functional command of program champions who can identify the internal barriers to adopting new technologies, come up with solutions, and get the job done. With representatives from each of the service branches. If “reforming the acquisitions process” in the military, or a massive mergers and acquisition campaign in the private sector, are the long-slog, full-army approaches to improving innovation, then special forces are the surgical strike. Fast, focused, powerful.

Let’s dig into #3 more. The need for special forces gets to the heart of why so many innovation labs fail, in the private sector as well as the military. It’s because the weak link in innovation is almost never in the supply of new ideas. It’s in the transfer to the field.

(Below is a summary; more details and examples in Loonshots.)

Good inventor vs. good champion

There’s a widespread failure to understand that being a good inventor and a good champion are vastly different skill sets. In the same way that being a good trombone player won’t make you a good basketball player, and vice versa.

For example, the idea of radar was discovered by a pair of scientists in the Naval Research Labs 18 years before World War II started. They were great inventors, but lousy champions. The idea sat there for a decade until a naval officer named Deak Parsons discovered it, went around to every bureau chief, pounded the table on why it mattered, boom, boom, boom, until he got them to cough up a check for $5,000 to fund the project. Robert Goddard was another great inventor, but lousy champion. It’s because there was no good champion for that idea here in the U.S. that the Nazis – who got the idea from Goddard’s papers – developed missiles and jet aircraft first.

You heard from General Shanahan about a bullheaded Colonel named Drew Cukor who pounded the table to stand up Project Maven and JAIC to bring AI to the military. Cukor is the most recent in a long line of internal champions, like Deak Parsons or Vannevar Bush, or Schreiver with ICBMs, or Moffett with aircraft carriers, or Rickover with the nuclear Navy (book, film). They were all great champions. Not inventors.

We need a new functional command to attract, train, and deploy great champions. To develop the next generation of Drew Cukors or Deak Parsons or Bernard Schrievers, rather than hoping and praying that maybe we’ll get lucky and another disrupter will come along in time and modernize the military. We no longer have that luxury. We cannot afford to start our conflicts with yesterday’s technology and hope that we will catch up. Not in the era of data and algorithms. Chris Brose was quoting John McCain when he said, “Hope is not a strategy.”

We need a separate command for the same reason that we need cyber or special operations as a separate command. The problem is endemic to all the service branches, not specific to just one. And there’s a unique skill set that needs to be developed. Good champions need to be mediators, buffers between technologists and soldiers. They need to be bilingual, to speak the language of each side fluently. They need to understand product-market fit: why some ideas will get internal traction while others won’t. They need to identify hidden organizational barriers and come up with solutions. They need to understand horizontal influence: how to influence people over whom they have no direct authority. Those are specialized skills, with best practices and useful lessons to be learned from years of examples across the different branches. Yet no such training exists today.

Think of it like this: A car needs pistons and it needs wheels. But without oil it will not run at high speeds. The special forces unit is that internal oil.

An elite unit, trained in those internal skills, will lessen the inevitable friction between the creatives working on new ideas and the soldiers focused on delivering on time, on budget, on spec. Those two groups generally don’t understand each other and often don’t like each other. In the private sector, the group that’s making the money rarely likes the group that’s spending the money, and vice versa. (More on the skills needed, design, and case for a new command here.)

Google, Microsoft, and a handful of leading private sector companies have figured this out. They’ve understood what it takes to do it well. They create a career ladder to retain people in the role, to build experience and skill, to convey prestige and respect. They keep the role neutral, like Switzerland, neither on the research side, nor on the operation side, but in between, like a mediator needs to be.

And there’s a bonus: If you create a joint special forces for innovation Sherpas, for program champions, you not only gain the ability to innovate faster and better as an organization. You improve your ability to attract, retain, and motivate talent.

When I put out that War on the Rocks article, I got emails from very impressive former service members, with an entrepreneurial bent, who clearly wanted to contribute. They said, “If that division was there, sign me up.”

The bottom line on why our nation’s defense organizations need this: because hope is not a strategy.

Or as another great leader once said: Luck is the residue of design.

* * *

My site doesn’t have comments, and alas my inbox is flooded, so if you’d like to engage, please link to this post on Twitter or Linkedin and tag me and #loonshots so I don’t miss it.

Yours in the pursuit of crazy ideas,
Safi

Further reading

WWII: chapters 1 and 2 of Loonshots (detailed references in the notes)
China and AI: AI Superpowers, Kai Fu Lee
The end of China’s foreign policy restraint: Foreign Affairs
Prior letter: from vacuum robots to balls through windows to mRNA: the story of the first covid vaccine

Video clips