Turn the Ship Around!

The problem it starts from

In 1999 David Marquet, a US Navy officer, had spent a year preparing to command the nuclear attack submarine USS Olympia. Weeks before he was due to take over, he was sent instead to USS Santa Fe, whose captain had left unexpectedly. The Santa Fe was a different class of boat, and Marquet did not know its systems in detail. It also had one of the worst records in the fleet for retention and operational performance.

Soon after he arrived, during a drill, Marquet gave an order that was impossible on that class of submarine. The officer of the deck repeated it to the crew, and nobody objected. When Marquet asked afterwards why he had passed on an order he knew could not be carried out, the answer was that the captain had given it. The episode showed Marquet the trap he was in. The navy’s system of leadership depended on the captain knowing more than anyone else. On this boat he knew less than many of his crew, and they had been trained to follow him anyway.

Turn the Ship Around!, published in 2012, is his account of what he did instead.

The practices

Marquet calls the traditional arrangement leader-follower: one person thinks and decides, and the others carry out. Its weakness is that the organisation’s intelligence is limited to the leader’s, and that its performance collapses when the leader leaves. His alternative, leader-leader, depends on three things, which he treats as the pillars of the book.

Give control. The first and most distinctive practice is a change of language. Rather than asking permission (“Request permission to submerge the ship”), officers were to state their intention (“Captain, I intend to submerge the ship”). Marquet would then ask what they thought he was thinking, and over time officers began to answer that question before he asked it, laying out the safety checks they had made. Eventually he would only say “very well”. The phrase “I intend to” moved the officer from asking to owning. Other practices served the same end: pushing authority for specific decisions down the chain, eliminating layers of sign-off, and ending the habit of top-down monitoring in favour of officers reporting what they were doing.

Build competence. Control given to people who do not know enough is dangerous, especially on a nuclear submarine. Marquet’s answer was to replace briefings, in which a senior officer told the team what would happen, with “certifications”, in which the team showed the senior officer that they understood what they were about to do. He asked crew to take deliberate action, pausing to vocalise and gesture before operating a control, so that mistakes were caught before they happened rather than after. Learning became continuous and everyone’s job, not something done in a training block.

Build clarity. People can only make good decisions if they understand what the organisation is for and what is most important at any moment. Marquet worked with the crew on a shared statement of the boat’s guiding principles, used them in daily decisions, and encouraged the crew to talk openly about their purpose. He also insisted on telling the crew the reasoning behind decisions, so that they could apply the same reasoning when he was not there.

The phrase that sums up the book’s approach is that authority should be moved to where the information is, rather than information moved up to where authority sits.

The idea underneath

Marquet’s real target is a set of assumptions about people. Douglas McGregor argued in 1960 that managers who assume people dislike work and avoid responsibility (his “Theory X”) build controls that make people behave exactly that way, while managers who assume people will seek responsibility given the right conditions (“Theory Y”) make it more likely. The Santa Fe crew was not passive by nature. It had been made passive by a system that rewarded compliance.

There is an idea about knowledge as well. The economist Friedrich Hayek argued that the knowledge needed to run an economy is dispersed among many people and can never be gathered by a central planner. The site’s idea page The Knowledge No One Holds explores it. A submarine is far smaller than an economy, but the problem is the same in miniature: the people closest to each system know things the captain cannot, and an organisation that requires every decision to pass through the top throws that knowledge away.

Marquet also gives a test for success that most leadership books avoid: what happens after you leave. He reports that the Santa Fe went on performing well under later captains, and that a disproportionate number of its officers were later selected for command. He presents this as the measure of leader-leader. An organisation that depends on its leader’s brilliance has not been led well.

Since it was published

The book has been widely adopted outside the military, especially in software teams, where the gap between the knowledge of the people doing the work and the people approving it is often large. Marquet’s later book Leadership Is Language (2020) extends the “I intend to” idea into a broader account of how the words managers use shape whether teams think or merely comply. His approach fits with Stanley McChrystal’s Team of Teams (2015), which reached a similar view of shared understanding and devolved authority from a very different setting.

A first step

For the next two weeks, when someone on your team asks permission, ask them to rephrase it as “I intend to…” and to tell you what they have checked. Answer with as few words as possible. Notice which decisions you could stop being asked about altogether, and tell the team so.