I’ve been reading Loonshots, which means I’m simultaneously arguing with bits of it in the margins and highlighting other bits to remember and study. This is why I can’t do audiobooks, especially for non-fiction.
Loonshots, by Safi Bahcall, is a business book written by a physicist. Unlike most other innovation books, this one really digs into the underlying structures and systems thinking that’s required to do innovation in a large organization. Bahcall brings a scientist’s critical thinking and eye for clarity to the topic.
There’s nothing new in here to the thoughtful innovators toiling inside IC agencies. We know all of these points; we’ve experienced them firsthand. We’ve tried to convey them to other practitioners, to management, and most importantly, to leadership. To date, we’ve been only marginally successful, but Bahcall has put all the points together in one very readable volume, so I am fully happy to leverage his privilege.
To do that, I’m going to write a blog post around each of what he calls The Bush-Vail Rules, going into detail about what that rule looks like from my perspective on the ground and front lines of innovation inside a large organization. I’ll bring my two decades of experience, but I want to hear yours also: how have you seen this manifest where you are? What might work where you are?
Ultimately, the obstacles we face in the national security world with doing innovation while maintaining our mission are pretty much the same (although not fully the same) as any large organization. That’s good, because that means we have lots of examples to learn from. Where we differ, I submit, is in impact. If Kodak or Nokia fail, that’s unfortunate for employees, shareholders, and maybe Finland’s GDP. If the US IC fails, that’s bad for the nation, for the world and life as we know it.
First, a primer.
Before we get into The Bush-Vail Rules, I need to define some of Bahcall’s terms (I did say this book was written by a physicist, didn’t I?) so that I can use them and you can understand them in future posts.
If you’ve been in one of my innovation courses or otherwise worked with the innovation network in the IC, you’ll have encountered one or more of the concepts in the Loonshots primer, but we’ve often framed them differently. I figured I’d do a little terminology comparison before getting into The Bush-Vail Rules, because I believe that will illustrate the differences in how Bahcall is thinking about it versus what our lived experiences have led us to understand.
Not to be confused with a “moonshot,” which is a high-risk/high-reward innovation effort, we’ve learned that inside an organization, anything new, no matter how obviously useful or small, will encounter friction when its creators try to get it into use. A lot of folks call this “corporate antibodies.”
Bahcall tends to look at game-changing technologies rather than incremental innovations, which is fine, but the same antibodies that keep the cure for cancer from getting deployed also keep any other innovation from getting deployed.
At one agency, using Mozilla Firefox as a browser rather than Microsoft’s IE was considered a loonshot even though it’s what we would term an H2 – Horizon 2 – innovation. The guy who proposed it and demo-ed it was called crazy and got walked out as a security risk (he was back the next day). Within most of our agencies, any idea that would significantly change our business processes or capabilities is met with the label “crazy.” Small tweaks and improvements are the only acceptable innovations.
However, the fact that Bahcall focuses on BIG IDEAS is fully valid, because it’s those big ideas that make big differences. To use the innovator’s risk assessment math, the uncertainty of success is the same, but the impact is staggeringly different. If the DoD killed an idea in the 1930s for ensuring that all mess halls have coffee stirrers, that’s one thing. If the DoD killed an idea in the 1930s for this new tech called “radar,” that’s another.
Stake and rank
We haven’t framed it as neatly as “stake and rank,” but we should start using this framing. We discuss the incentive structures a lot, particularly those under which middle management operate. Incentives in government agencies are an impedance mismatch with innovation (and with some actual mission, but that’s beyond the scope of this blog). Far from having a stake in the success of a new capability, our officers are incentivized to avoid contact with it if humanly possible, in case it fails…and 90% of new ideas do. It’s far better to maintain the status quo than to risk taking on something new.
We’ll end up talking a LOT more about incentives in this blog series, because incentive structure is a main theme of Loonshots…but think of incentive structure like a game. You win Monopoly by charging rent to players that land on spaces you own. Even if you don’t understand that from reading the rules, you’ll get it after playing the game once. After that, if you don’t buy properties early and often, you’re playing stupid.
To say that officers in a large organization should be focused on either the success of new capabilities or refining the systems underlying their production, rather than focusing on their personal rank, is to say that officers should look at how the game is played and won and then play stupid.
We don’t expect our workforce to play stupid.
Which means we have to change the rules, or split our agencies into two or three different games.
Or, as it’s commonly known here, the Valley of Death.
While we’ve occasionally been able to leverage large groups of people for innovations, mostly innovators are told to work by themselves. Sometimes with resources. The onus for doing everything from coming up with the idea to prototyping it to testing it to enrolling support to marketing it to getting user adoption to getting program adoption is laid on whoever brings up the idea. No one has all those skills, or time, or ability to stop doing the mission execution stuff they are rewarded for. (Well, okay, there are a few, maybe 1 out of every 20,000 people. We call them “unicorns” and if they aren’t hunted down for their magic blood, they realize that they just sacrificed everything for no reward and fade into the forest.)
In fact, sometimes we “reward” the person who has the idea by saying “you get to build it!” That is not a reward for an idea person; it’s more like a punishment.
There is no structure in any IC agency for enrolling the large number of people needed for a transition, managing them and the work, etc. We don’t even have refactories, which are a staple of Silicon Valley. For the IC, it’s “Lab —> [magic happens here] —> Factory.” Thus, many promising prototypes, pilots, or acquisitions never scale.
Any exceptions are due to heroics or a dedicated team of top-level executives spending way too much time cracking the ice.
Does an organization deliver on time, on budget, and on spec? Or do they continually develop new capabilities in order to remain relevant? They should do both, but the environment each requires is toxic for the other. We’ve been calling this “Search vs. Execute” (to use Steve Blank’s terminology that’s gaining traction in the innovation world), or sometimes “Reach/Execute.” To quote Vannevar Bush: “The essence of a sound military organization is that it should be tight. But a tight organization does not lend itself to innovations…and loosening it in a time of war would be fraught with danger.” But, Bush continued, there “should be close collaboration between the military and [some] organization, made loose in its structure on purpose.”
A 2004 Harvard Business Review article on this coined the term “ambidextrous organization.” It seems obvious, but it’s really hard to do; Haanaes, Reeves & World argue that only 2% of the companies are part of the elite group of organization who understand that you have to excel at both efficiency and innovation.
Efficiency is much easier to measure, and an organization’s existing mission is much better defined, than innovation. Innovation exists, by definition, in conditions of extreme uncertainty. Unfortunately, the methods by which innovators pay down risk (Human-Centered Design, Lean Startup, DevOps, Agile Development) are exactly the opposite of the metrics and measurements used to manage existing operations.
Without a lot of work by either one high-powered individual with POTUS as his top cover (like Vannevar Bush) or a team of dedicated and resilient individuals with the CEO (or D/CIA or DIRNSA etc) as their top cover, the Execute side will stifle and kill the Reach side and the organization will stagnate.
I don’t want to worry you, but when that happens, the organization goes out of business (Kodak), or suffers huge losses (Pearl Harbor), or millions of people die when a cure is held up in bureaucracy.
OTOH, when phase transition is managed and balanced, we win wars and transform the world.