Spies, Marines, Army and Police

Reid Hoffman has this bit in Blitzscaling that has stuck with me. It’s an idea much deeper than he gives space to—maybe worth an entire book. It can help us understand common pathologies in organizations, and when spoken out loud, can help remedy them. The gist of the idea is:

Organizations require different types of people at different points in an organization’s life or activities. These types are: Spies, Marines, Army and Police.

Marines storm the beaches and secure the ports. They are comfortable with ambiguity. They prototype ideas, find product market fit, write imperfect code in high-level languages, have woeful test coverage, and are happy to drop a project that isn’t showing promise.

The Army goes into the hinterland and fights the guerillas. They bring in heavy artillery and industrial-scale logistics. They are comfortable with ambiguity but ultimately want to tame it. They root out edge cases and write in C++. They write checklists, systematizing things. They create the conditions in which police can govern.

Finally the Police enter. They are deeply integrated in the community, uphold rules that the army might ignore, and maintain law and order. They abhor ambiguity, and require well-defined rules to implement. They thrive in scoped roles in which their performance can be fairly measured.

One valuable archetype that Hoffman does not include but which I think is worth mentioning are the Spies.

Spies provide intelligence to the system to help inform the decisions by marines, army and police. They actively seek out ambiguity and are curious, creative and analytical. Their roles change from day to day and defy descriptions. They help inform which markets to enter and which strategies are worth employing. Importantly, the payoffs to their work can be observed only occasionally, and can be disastrous or incredible. When a policeman does a poor job the result might be a few murders; when a spy does a poor job you get Gallipoli or the Iraq war.

Personalities

The suitability of people to spy/marine/army/police-type roles depends a lot on what’s going on in their lives that might change over time—career stage, young children, insecurity at home, financial resources and so on—but probably on relatively fixed personality traits (especially big-5 Openness) and inclinations also.

Spies

• Very high openness and analytical intelligence. This might present as high curiosity and creativity.

• High sociability, which can increase information flow.

• Often bad at management and bureaucracy, and often need an interpreter to work well with army and police.

• Classic VC, researcher, board member.

Marines:

• High openness, extraversion, not high agreeableness, low neuroticism. Very high intelligence—but an ability to not get caught up in the weeds.

• Actively seek out ambiguity and have a bias towards action (“seek forgiveness”).

• Works well with other marines but viewed by army and police as being undisciplined or abrasive.

• Classic founder.

Army:

• Moderate openness, very high conscientiousness.

• A “team player” with ability to discern when to be disagreeable and when not to.

• Capable of “getting” marines and police, and so acts as social glue within the org.

• Classic executive. Brings order to chaos.

Police

• Low openness, high conscientiousness. Perhaps high neuroticism.

• Craves excellence in daily activities; loves a checklist.

• Works well with other police and respects good army personnel, but can have mutual antipathy with marines and spies because of perceived inflexibility—”I’m just following the rules, why can’t you?”

• Classic executive in operations, legal, finance.

Excellent organizations need top spies, marines, army and police—but sometimes end up lopsided

Organizations need all types. Without good spies, an organization solves the wrong problems or enters the wrong markets; without good marines it fails to innovate, or optimizes prematurely at enormous cost; without a great army it struggles with the many wrinkles scale brings; and without excellent police, the wheels come off. Yet few organizations manage to be best-in-class across all types. Isn’t that …weird? Why do large network TV companies struggle to hire the marines that might help them compete with Amazon or Netflix? It’s not like they don’t have the money. Why do successful startups struggle to attract top army and police?

My theory is that organizations with strong types across the board—healthy organizations—are ones where its members (a) know their type, (b) work in roles that agree with their type, and (c) genuinely respect the value of the other types and understand their inclinations. Job satisfaction and performance seems so linked to the fit between individual type and role need. And so it’s worth paying attention to the forces that drive discord between types. There are two forces in particular that worry me: valorizing certain types, and hiring.

Some types get valorized, skewing behavior: police trying to be marines, or spies LARPing as army officers. Popular culture exacerbates this by heaping attention on the founders and executives who tend to be marines or army. Internal incentives can also valorize certain types. If marines perceive that only army officers get promoted, they morph into second-rate army officers or leave—diminishing the organization’s ability to solve new problems.

Left unaddressed, the problem compounds over time through hiring. People are attracted to people who share similar strengths to themselves, and maybe more importantly, those whose weaknesses won’t make their own jobs harder. Spies value creativity and divergent thinking in others. Marines value a “just ship it” attitude. Army soldiers value others who tame the wild. And Police officers value those who obey the rules.

Note that the Spies and Marines are in some sense “upstream” of the Army and Police, in that they can make the Army/Police’s lives a lot more difficult, while Army and Police make life easier for the Spies and Marines. This asymmetry is the nub of the problem. It’s why older organizations tend not to hire Spies and Marines. As a firm grows, it hires Army and Police. But after some time, both recruitment teams and selection panels are staffed by people who want to hire people who make the organization work more harmoniously—Army and Police. Zoom forward a few decades and once-exciting startups lose their ability to do anything new. It’s why you can’t name an established company that has a genuinely new and innovative product that wasn’t an acquisition or copycat effort. [NB this doesn’t mean they don’t make a lot of money for a long time!]

How to prevent discord and build balance

• Name yourself
  
  Most people can read the descriptions of the types above and immediately see themselves as one of them. Having each member on the team identify their type and talk about it as a group helps highlight the balance and recognize what future tensions will look like. It breeds a healthy respect across the team for their differences.

• Hire and promote all types intentionally
  
  Naming the types—and needs for them at all levels—helps in creating job descriptions and hiring committees for new staff. But you also need to make sure that promotions don’t skew too much in any one direction. If junior staff don’t see people of their type in more senior roles, they know there is no future for them. You really do need to do it intentionally.
  

• Finally, learn from Focused Research Organizations, DARPA and US Presidential Administrations and put an end-date on things
  
  US Presidential Administrations are filled with Spies and Marines, despite the White House being an old institution. It’s obvious why this is: turnover between presidents means the hiring dynamic never gets into gear (unlike in the public service…). Similarly, DARPA’s program managers—archetypal spies—serve on a fixed term. The power of finite terms to create a bias towards creative talent is one of the reasons I’m excited about Focused Research Organizations (FROs), a new research model in which small groups of researchers run a focused research initiative that starts its life with a scheduled end-date.
  
  Obviously not all problems are amenable to hiring people with an end-date in mind. Yet many are—far more than currently structured that way. Building a muscle for spinning up new time-bound teams to work on a project outside standard bureaucracy is a way of attracting precisely the talent most scarce in mature organizations.