Rule 1 – Separate your Soldiers and Artists
Theodore Vail took over AT&T with the promise that soon, anyone in the US would be able to call anyone else, from New York to San Francisco. Everyone, including most of AT&T, thought he was crazy. Which was a perfectly rational conclusion, given that it was 1907 and we’d only discovered the electron 10 years earlier.
In order to achieve his goal, Vail’s company would need technologies that did not yet exist, based on science that was not yet known.
If you’ve been following the stories this summer around the 50th anniversary of the Apollo program, that should sound familiar. In order to get a man to the Moon and back, NASA had to invent technologies based on science that was not yet known.
Both AT&T and NASA succeeded by separating and sheltering radical ideas from the day-to-day operations. Vail created a department of loonshots, run by loons, free to explore the bizarre. They were not subject to the management structure that keeps the “mission execution” side consistent. This meant that they were not killed early due to not being perfect.
Ed Catmull calls early-stage innovations “Ugly Babies.” They don’t even vaguely look like what they’ll be; they are a caterpillar that the visionaries believe can be a butterfly, but to get there they need strong cocoons.
At some point, the Artist does have to “take his baby to Fight Club (as we say in my circles),” but that exposure needs to be carefully managed; the Artists need to be separated from the Soldiers. Bahcall phrases this in his talks as “you may think your baby is beautiful and has great potential, but when you hand it to a Soldier, he or she is doing to see dirty diapers.”
When Two Tribes Go To War
In most IC agencies, the innovators/artists are dramatically outnumbered by the executors/soldiers. The overall structure of the agency is around repeatable mission execution; it’s the Ice side. For this reason, we’ve been calling our innovators “pirates in the navy” – the innovators have a lot of the same skills, but a flatter hierarchy, a looser dress code, and they work better under uncertainty than conformity.
Steve Jobs referred to his innovators as “pirates” also, but he made the mistake of favoring them over the “mundanes.” Like any tribal othering anywhere, that’s just dumb. Scientist, engineer, or technologist: all are problem solvers and want to get stuff done. Separating the pirates from the navy, the artists from the soldiers, is not about saying who is better – both are crucial, just at different times.
What Separation Means
It doesn’t mean that pirates and navy can never mingle. In fact, the mingling is absolutely required. The Artists need to understand the unmet needs of the Soldiers, and the Soldiers need multiple chances to interact with what the Artists are working on in order to buy down risk.
What it does mean is that the Artists need a place where they can be artists. Each IC agency, and probably the IC in general, should have a loonshot nursery. Vannevar Bush quarantined the team working on the technology for long-distance telephony in an office building in lower Manhattan. The Pentagon wasn’t able to just drop in and kill it. Microsoft quarantined the Xbox team in a separate facility, where the Windows OS management couldn’t find it and kill it.
Pixar (and here I’m leaning on Ed Catmull’s book Creativity Inc., rather than Loonshots) alternates loonshots with franchise movies. They make a sequel to pay for the next unique, creative, but more uncertain, release. This systems thinking approach draws from Jim Collins’ Flywheel (he just published this as a monograph if you don’t want to re-read From Good to Great).
Loonshots also looks at the movie industry, pointing out that studios that did not separate artists from soldiers and focused only on franchise passed on a movie adapted from a book series about a metrosexual British spy who saves the world, and also passed on a script titled The Adventures of Luke Starkiller. This happens today as well: movie studios are happy to make Transformers 8, because they can calculate the return-on-investment (ROI). It’s whatever Transformers 7 made, minus a bit. Calculating how well Slumdog Millionaire or Moonlight will do is much less certain.
But, if all you do is churn out declining repeats, you’ll miss the wins that come from loonshots. That’s bad for business, but it can be existentially devastating for a defense, intelligence, or other national security agency.
If Vannevar Bush hadn’t separated the OSRD from the military, and if Vail hadn’t stood up the separate Bell Labs, it’s very likely the US would not have succeeded in WWII. Even if the Nazis hadn’t won due to other factors, it’s possible that we wouldn’t have the vacuum tube, the transistor (which is kind of important for the Information Age), the solar cell, the CCD chip, the continuously operating laser, Unix, or C.
But what does “separation” really entail? We’ll get into more detail with the next rule.
Rule 2 – Tailor the Tools to the Phase
So, ideally you’ve physically separated loonshot nurseries and Artists from franchise production and Soldiers.
However, just having a different building or office isn’t the key. It’s, in fact, a classic trap. Almost every large organization will set up a snazzy “innovation office” somewhere that will proceed to produce nothing. Everyone outside the office understands that they have no stake, and focuses on achieving rank, which doesn’t come from innovating. Those in the office lack authority or sufficient weight to balance out the entire rest of the organization.
Whenever I’ve seen a continuous exchange between the loonshots and the franchise, the franchise kills the loonshots, often without having to provide a reason, sometimes stomping it into the ground. Rather than Chaos, as Bahcall says, it’s a meat grinder. Managing and limiting that exchange is essential.
Tailoring the Tools
In addition to physical separation, the Artists must also have a separate management structure, one tailored to the needs of their process. This means no quad charts, no charters, no prescriptive mandates for what gets delivered when. No governance boards – governance in general kills innovation. Six Sigma or TQM might help franchise/execution projects, but they will suffocate loonshot/search. Any attempt to impose the strict franchise/execution structure on the Water/Artist/Loonshot/Prototype side will ensure that nothing survives. Rather than a Jim-Collins-style flywheel effect, you wind up with a Doom Loop, where nothing can get funded because it isn’t already funded.
Rather, Artist management must set smart guardrails so that the artists can drive as fast as possible around the curves. They must expect and get evidence from the artists of what they’re doing and why. KPIs and metrics must be qualitative as well as quantitative. The management in charge of making decisions about the projects must have full access to all the information about the projects.
The private sector calls this Product Management; it’s like being a Program Manager, but for something that doesn’t exist yet. Product Managers take products from concept through pilot and help them transition into larger programs. This isn’t a role the government has; in government HR, only things that already exist have managers and managers only have tools and support to manage existing, understood, programs.
The tools and processes needed by the Artists are very different and must be consciously established and maintained.
In addition, the incentive structure for Artists must be different than what the Soldiers have. On the franchise/execution side, failure is bad; on the loonshot/search side, failure is crucial. Just as we wouldn’t give an award to an ops team for failing to collect intel, we shouldn’t punish an innovation team for failing to collect intel when they are actually working on designing a better cover strategy. Yet, “collection of intel” is likely to be a core performance rating for all officers.
The Artists don’t really need an org chart, but rather an algorithm for who is good at what. This side should work more like the elite Joint Special Operations Task Force (JSOTF) in the Middle East under McChrystal, as explained in Team of Teams, or a “teal” organization as described in Reinventing Organizations, by Frederic Laloux.
They need a different career service, maybe (although ideally officers would move in and out of both sides throughout their careers) but for sure a specialty that bases promotions on loonshot work.
They need slack time, and the ability to work when needed. Hourly billing or the 8-hr/5-day workweek is stupid for all knowledge workers, but extra stupid for innovators.
Finally, the Water/Artist/Loonshot/Prototype side needs money. Not a lot, honestly, because otherwise the focus could slip from what’s truly needed…but some. Right now, government budgets are locked into the costs for executing repeating mission. Where there is R&D funding, how it’s ventured is not transparent, and we don’t use an approach that buys down risk early on. A better approach would hold contributions from the folks with the unmet needs, and use that to fund prototyping (Artists only) and piloting (the dynamic equilibrium between Artists and Soldiers, aka a refactory). That way, the recipients have skin in the game.
As a rule of thumb, more of the funding will go to the refactory than to the prototyping, which should be done cheap and fast, but that, of course, depends on the project. Industry recommendations are to put 70% of your innovation budget against Horizon 1 (H1) projects, those that get into production fast with little risk, but that also deliver small value. H2 projects get 20%, and H3 get 10%, yearly. An H3 project might take years to deploy, with a commensurate total development cost.
That’s 30% on “loonshot” innovation – what percentage of your agency’s budget goes to that now?
But what about the Soldiers?
I chose to focus on what a tailored toolset for the Artists would look like, because by and large we don’t have that. There are some elements of some agencies that work this way some of the time, but, especially around budget and incentives, they are still using Soldier tools. We’ve got executing repeating mission sorted; what we don’t have is a parallel structure that will keep us from becoming obsolete.