Aug. 5, 2019 · 6 min read
Turn the Ship Around!
On his first day in command of USS Santa Fe, L. David Marquet gave an order that didn’t exist. He called for “ahead two-thirds,” a speed setting from his previous submarine. The Santa Fe didn’t have one. His officer of the deck relayed the order anyway. The helmsman tried to comply. Nothing happened. The officer knew the command was impossible and issued it regardless, because the captain had spoken (Marquet, 2013).
That moment defined what Marquet spent the next year dismantling. He called it the Leader-Follower model. Authority concentrates at the top. Decisions flow back down. It works for physical labor. It fails for intellectual work.
Santa Fe had the worst retention and performance record in the Pacific fleet when Marquet arrived. Three years later it was ranked first. Reenlistments went from three to thirty-three in a single year.
What Is the Leader-Leader Model?
Leader-Follower concentrates authority at the top. Information moves up. Decisions move down. Performance is tied to whoever is at the top. When they leave, the organization regresses.
Marquet’s alternative is the Leader-Leader model. Authority moves to where the information lives. Every person leads within their domain. The leader’s role shifts from decision-maker to conditions-setter.
Empowerment is a top-down concept. A leader grants power to a subordinate. That act reinforces the hierarchy it claims to dismantle. Emancipation starts from a different premise. The team already has the capability. The structures prevent its use. The leader removes the structures.
Westrum found the same pattern. Organizations where information flows freely and authority matches expertise outperform hierarchical ones on performance and safety (Westrum, 2004).
Control, Competence, Clarity
The Leader-Leader model rests on three conditions. Remove any one and the model breaks.
Control. Decision-making authority must move to where the information lives. On the Santa Fe, Marquet shifted authority over crew leave from the executive officer to the chiefs. Small change, large effect. The chiefs were now accountable for the watch bill.
Competence. Control without competence produces chaos. Distributing authority only works when people have the knowledge to act on it. Marquet shifted the crew from passive training to active learning. Before operations, teams were certified rather than briefed. A certification requires participants to demonstrate that knowledge before the operation proceeds. If they cannot, the operation waits.
Clarity. Distributed decisions align only when everyone understands the mission well enough to make consistent trade-offs. Clarity means goals, not methods. On the Santa Fe, a fire drill had one objective. Get a pressurized hose to the fire in under two minutes. Freed from rigid procedures, crews self-organized and rearranged bunk assignments to optimize response time.
The Mechanism: “I Intend To”
The most concrete change in the book is a phrase. Crew members stated their plan directly. “I intend to submerge the ship. The water depth is sufficient and the watch team is ready.” “Very well,” the leader responds.
The speaker must formulate a plan and a rationale before speaking. Psychological ownership transfers from leader to doer. The leader becomes a check on the organization rather than its engine. To construct a credible case, the speaker must think at the level of the person above them.
The crew began taking initiative before problems escalated. Communication became more specific. Officers stopped asking for permission and started informing superiors of plans already in motion.
Research on autonomy and intrinsic motivation supports what Marquet found in practice. When people direct their own work, engagement and performance improve (Pink, 2009). “I intend to” creates that condition while preserving accountability.
Common Mistakes
Empowerment without structural change. Most organizations announce empowerment without altering the decision rights that make it necessary. If authority still flows from the top, the initiative reinforces the model it claims to replace.
Control before competence. Distributing authority to people who lack the knowledge to act on it produces bad outcomes and erodes trust. Marquet’s sequence was deliberate. He built technical competence and knowledge redundancy before accelerating control distribution. The order matters.
Treating “I intend to” as ceremony. The phrase only transfers ownership if the speaker genuinely thinks through the plan and the rationale. Used as a ritual to satisfy a process, it produces the right words with none of the effect.
Measuring leadership by the leader’s presence. In Leader-Follower organizations, performance drops when a strong leader departs. This is typically read as evidence of effectiveness. In Leader-Leader organizations, the departure changes nothing. The test is what happens after. High-performing technical teams show the same pattern. Elite performers deploy continuously and recover from failures without a single point of human control (Forsgren et al., 2018; DORA, 2023).
Confusing flat hierarchies with distributed authority. Removing layers from an org chart does not distribute decision rights. Team cognitive load and communication boundaries reveal where authority actually sits, regardless of what the hierarchy diagram shows (Skelton and Pais, 2019).
Put It Into Practice
Identify one decision that currently requires your approval but shouldn’t. Ask whether the person doing the work has the information and competence to make the call themselves. If yes, transfer the decision right. Do not announce it as empowerment. Remove yourself from the loop.
Introduce “I intend to” to your team. When someone comes to you for an approval they should own, ask what they intend to do. Let them answer. Respond to the plan. The habit takes weeks to establish and months to become natural.
Measure progress by what happens when you step back. If performance holds or improves, the conditions are working. If it degrades, that reveals where competence or clarity gaps remain. Address those before distributing more control.
References
DORA (2023). State of DevOps Report 2023. Google. https://dora.dev/research/2023/dora-report/
Forsgren, Nicole, Jez Humble, and Gene Kim (2018). Accelerate: The Science of Lean Software and DevOps. Portland: IT Revolution Press. https://itrevolution.com/product/accelerate/
Marquet, L. David (2013). Turn the Ship Around! A True Story of Turning Followers into Leaders. New York: Portfolio/Penguin. https://davidmarquet.com/books/turn-the-ship-around-book/
Pink, Daniel H. (2009). Drive: The Surprising Truth About What Motivates Us. New York: Riverhead Books. https://www.danpink.com/books/drive/
Skelton, Matthew, and Manuel Pais (2019). Team Topologies: Organizing Business and Technology Teams for Fast Flow. IT Revolution Press. https://itrevolution.com/product/team-topologies/
Westrum, Ron (2004). “A typology of organisational cultures.” BMJ Quality & Safety, 13(suppl 2): ii22–ii27. https://qualitysafety.bmj.com/content/13/suppl_2/ii22
Outtakes
The USS Benfold. Commander D. Michael Abrashoff transformed the USS Benfold into what he called the best ship in the Navy using almost identical principles, years before Marquet and without knowing his work. He wrote about it in It’s Your Ship (Abrashoff, 2002). Two commanding officers independently converged on the same solution to the same problem.
Toyota’s andon cord. Any worker on a Toyota production line can signal a problem by pulling the andon cord. If it goes unresolved by the next cycle boundary, the line stops. Distributed authority to halt production has been operating at industrial scale since the 1950s, decades before Marquet named it (Ohno, 1988).
Changelog
2026-05-16 Reformatted from reading notes. Rewrote as full post.
2019-08-05 Initial draft.