I'm currently behind on reviews, so don't be surprised if the recent reviews are a bit sparse.

Turn the Ship Around!

Cover of Turn the Ship Around!.

Anarchic structures are great, particularly at groups under a certain size. I know this, my friends know this – I didn’t expect a US submarine captain to know and advocate it. Yet that was the pleasant surprise I got.

Marquet knowing all this is great to read, him advocating it less so – reading this book is much more fun when you read it as a cool guy telling life stories, than when you read it as the management book it’s intended to be. That’s mostly because he has a good grasp on how all of this works, because he’s done it for a long time. But he only had to introduce it once (at scale, in his own group), so it’s like his success was mostly driven by intuition and courageous poking around, and all the lessons he draws from it won’t replicate for anybody else.

He emphasises that in leader-leader structures (as opposed to leader-follower), you get much better results because everybody will support a shared goal to their best abilities, and complex operations can be prepared much more seamlessly because you don’t have to rely on a central point of failure tracking all ongoing motions. Plus, this is a self-perpetuating stable system – when an important person leaves, the system does not collapse (as can be the case in hierarchical systems), and newcomers can be introduced to the system by the responsible members.

Neither permission nor forgiveness: intent

Instead of asking for permission, or ignoring boundaries and asking for forgiveness later: declare intent. Rewording like that consistently will make you more proactive and will make it clear that you are taking responsibility.

Say clearly what you mean to do – and give as much context as appropriate. If you need permission, give as much context as your supervisor will need (e.g. the things you have checked, the alternatives you have considered, the preparations you have made). This has the advantage that you, and everybody else, has to try at least for a second and see things from the perspective of their supervisor.

This is something I see a lot while organising events – there is a relatively clear split between people who always want to request permission to do something (or say “What should I do about …”, “Do you think we should …”, “Could we maybe …”) and the people who present plans or intentions (“I plan on …”, “I intend to …”, “I will …”). Those are still up for discussion, of course, but the change in atmosphere is huge.


Additionally, Marquet makes a good point that is underrated in the hierarchy vs anarchy debate: “As authority is delegated, technical knowledge at all levels takes on a greater importance.” When you push decision making to the people actually involved and at the scene, it’s important that they have the competence to make those decisions. You have to take more care to have good education programs, pass around important information, and that everybody can handle the responsibility. He offers some thoughts on education that won’t apply outside military structures, and one thought on plans and meetings: When you have a meeting for the purpose of planning/making sure everybody is on the same page, instead of announcing the shared information and letting everybody nod, try this: Have everybody state their part and understanding. “I’m going to do X”, “then I’m going to see Y”. Less listening to plans, more tabletop RPG, nearly? I wonder if tabletop RPG players on average tend to have more initiative and less leader-follower behaviour 🤔


He makes a couple of points that I like a lot:

  • Don’t empower: emancipate. You cannot empower somebody else – that would assume that you have the authority and ability to do so, and they don’t. “Empowerment still results from and is a manifestation of a top-down structure. […] With emancipation we are recognizing the inherent genius, energy, and creativity in all people, and allowing those talents to emerge.”
  • Short-term reward systems break everything, and are particularly bad for long-term success. Same goes for top-down monitoring systems: both send the message that people don’t have responsibility for the job they do.
  • There’s a big difference between asking questions out of genuine curiosity, and testing people. Everybody can tell.
  • You need an actual shared goal that people can get behind. “Avoid errors”, for example, is no such goal – it’s not inspiring and sounds like dullness. “Be excellent”, on the other hand, can work very well.
  • Explicitly go for short, early conversations, where you think out loud to make sure you’re talking about the same thing and are going in the right direction. This requires trust on both sides: that the one person will do the right thing, and that the other person will not criticise the lack of details in a rough draft.
  • Approach anything that feels like (or is) a test/inspection/rating/judgement with the attitude that you get to learn things (particularly in formal/skilled environments). If you highlight your shortcomings and ask for advice from the experienced people judging you, and if you’re honestly interested in improving, that will benefit you and your relationship (and probably your grades/rating/…).
  • Specify goals, not methods. People will figure it out.
  • Take care of your people. Talk to them about their goals, their dreams, their concrete targets, and then do what you can to help them. Together with handing them responsibility, this is one of the best ways to show and build trust. Have people write their own evaluations (if work setting, else: diary entries? see related books) one, two, three years into the future to figure out what they want to achieve, realistically.
  • If you need awards, have awards without artificial limit. Pit your team against the world instead of each other.


Control without competence is chaos.