PDF Summary:Turn the Ship Around, by L. David Marquet

Turn the Ship Around! by L. David Marquet tells how a captain turned the US Navy’s worst-performing nuclear submarine crew into one of the best. Marquet transformed the demoralized crew into an empowered, motivated fighting force by replacing the traditional “leader-follower” structure with a “leader-leader” model that gave crew members control over their work and taught them to think and act proactively. Marquet argues that most organizations would benefit from a similar change, especially those that still operate on outdated models designed for coordinating physical labor, not the knowledge work that defines modern jobs.

Marquet’s model distributes decision-making authority throughout the organization. In this guide, we’ll explore how you can apply Marquet’s principles for giving people control, building their competence, and creating clarity around shared goals. We’ll also connect his ideas to research on leadership and motivation, trace how traditional management emerged from industrial thinking, and examine how to adapt his military-tested approach to different workplaces.

(continued)...

Organizations also reinforce the model through their structures and incentives. The Navy, for instance, emphasizes the total accountability of the commanding officer. This makes sense for maintaining clear responsibility, but it creates pressure for the CO to control everything. (If anything goes wrong, it’s the CO’s fault regardless of who made the decision.) Marquet notes that many civilian organizations operate similarly, holding managers accountable for their team’s results without giving them any incentive to develop their people’s decision-making capabilities. The emphasis is on short-term performance, not long-term leadership development.

How Leadership Accountability Varies Across Cultures

The accountability structure Marquet describes reflects specific organizational and cultural assumptions. In individualistic cultures like the US, leaders receive disproportionate credit when their teams succeed (reinforcing the idea that strong leaders drive performance), but often aren’t held as accountable for failures. In collectivist cultures, like those of many Asian, African, and South American countries, accountability patterns differ. Leaders typically share credit for success with their teams but are held personally responsible for failures.

For example, after the 2011 Fukushima nuclear disaster, TEPCO’s leader took personal responsibility and resigned, even though the earthquake and tsunami that caused it were beyond anyone’s control. In contrast, after BP’s 2010 Deepwater Horizon oil spill—which a US court ruled resulted from “gross negligence”—the CEO eventually resigned but never accepted personal responsibility. But the difference in accountability doesn’t necessarily make collectivist cultures more compatible with Marquet’s leader-leader model. Many collectivist cultures emphasize hierarchy and respect for authority, which suggests they often maintain hierarchical decision-making structures similar to the leader-follower model.

The implications for implementing Marquet’s approach may vary by context. In organizations with a very individualistic culture, the challenge may be counteracting incentives that reward individual achievement and discourage leadership development. In collectivist organizations with strong hierarchies, the challenge may be different: overcoming cultural norms around hierarchical deference and obedience to authority, even while leveraging existing cultural emphasis on group harmony and collective responsibility.

Empowerment Programs Don’t Fix the Problem

​​Many organizations recognize that workers are disengaged and try to address the problem through empowerment programs. Leaders give speeches about “taking ownership” and “being proactive.” But Marquet argues these efforts fail because they try to layer empowerment onto a leader-follower structure without changing the structure. He learned this firsthand on the USS Will Rogers, where he served as an engineer before taking command of the Santa Fe. The CO of the Will Rogers ran a strict top-down operation, but Marquet tried giving his crew more autonomy by explaining objectives and letting them figure out how to achieve them.

The change backfired. The crew missed deadlines, fell behind, and made dangerous mistakes, in one instance improperly installing critical bolts to save time. The problem wasn’t a lack of ability or bad intentions, but that the sub’s entire culture—all the systems and procedures, and everyone’s training—still reinforced “leader-follower” thinking. When he tried giving people authority to make decisions, they didn’t know how to use it well because everything else in the organization told them their job was to follow orders. Marquet explains that you can’t empower people within a system designed to keep them powerless.

The Leadership Challenges of a Nuclear Submarine

The workplace Marquet was trying to change was that of a ballistic nuclear submarine, which is powered by an onboard nuclear reactor to generate electricity and propulsion. The ballistic missile submarine fleet was built during the Cold War as part of America’s nuclear deterrence strategy.

The conditions onboard such submarines created an extremely high-stakes workplace—the crew was submerged for months alongside both nuclear weapons and a nuclear reactor, and mistakes could kill the entire crew and trigger environmental or geopolitical catastrophe. To deal with these risks, the Navy designed an intensely centralized system. Every detail of reactor operation was controlled through standard procedures, and shore-based authorities maintained oversight at all times. When the Navy started using nuclear subs, one admiral, Hyman Rickover, selected every officer in the program and insisted on absolute procedural compliance. This worked: His fleet operated 152 reactors over 30 years without a single radioactive emission.

This context helps explain why Marquet’s efforts on the Will Rogers failed. The crew had been trained to follow procedures without deviation because in their environment, deviation from procedure could be catastrophic. Research on submarine crews suggests that extended isolation in confined, artificial environments also degrades people’s psychological functioning, which makes adaptation to new expectations even more difficult. The entire nuclear submarine culture, built over decades to ensure safety, actively discouraged the independent thinking that good decision-making requires.

Why the “Leader-Leader” Model Works Better

Marquet contends that the “leader-leader” model addresses the flaws of the “leader-follower” model by changing how the entire organization thinks about work and responsibility. First, it engages everyone’s intellectual capacity. In a “leader-follower” system, workers can do their jobs without really thinking—they wait for instructions, execute them, and report back. But when they have authority over decisions in their work, thinking becomes unavoidable. They have to analyze situations, consider options, anticipate consequences, and take responsibility for the outcome. They’re no longer just mindlessly executing on someone else’s thinking—they’re doing critical thinking for themselves.

(Shortform note: When we stop actively thinking at work—whether others micromanage us or technology does the thinking for us—we risk losing our capacity for independent thought. Studies show that people in intellectually demanding jobs score higher on cognitive tests decades later, suggesting that their work builds durable mental reserves. Conversely, chronic workplace boredom leads to depression, anxiety, and stress. AI may exacerbate this problem: By outsourcing our thinking to AI, we risk creating a feedback loop that narrows our thinking patterns, and if AI replaces many jobs involving basic intellectual labor, as some experts predict, it’s an open question how we’ll maintain our cognitive capacity.)

Second, Marquet explains that the “leader-leader” model works better than the “leader-follower” model because it develops leaders throughout the organization. In a “leader-follower” system, you only need to understand your specific task, while your supervisor worries about the bigger picture. But when you have authority to make decisions, you have to weigh whether it’s the right thing to do given the broader context and think through the questions your supervisor is likely to ask—considering competing priorities, resource constraints, organizational goals, and risks. In effect, you’re practicing leadership at the next level before you get promoted to it.

(Shortform note: When Marquet notes that workers learning to be leaders do so in part by learning to anticipate their supervisor’s questions, he’s describing the active processes of learning by doing—and by developing a mental model of expert decision-making. This also aligns with “cognitive apprenticeship,” a method for teaching complex thinking skills. In traditional apprenticeships, you watch a carpenter build a cabinet or observe a tailor sew. But in knowledge work, thinking processes are invisible, so you have to make them visible to help others learn. To do this, an expert models their thinking; provides scaffolding, or support, as learners practice; fades that support as skills improve; and coaches them through the process.)

Third, the “leader-leader” model also creates lasting organizational excellence. Marquet explains that in “leader-follower” organizations, performance depends on having an excellent leader at the top and often collapses when that leader leaves. But when decision-making authority is distributed throughout the organization—when the systems, procedures, and culture all reinforce people taking ownership—performance doesn’t depend on any one person. The capability is embedded in how the organization operates.

How Distributing Authority Also Distributes Knowledge

Marquet’s insight about lasting organizational excellence connects to a problem familiar in startups: the founder’s curse. This occurs when a charismatic founder becomes the sole source of innovation and decision-making in a company, rewarding people who execute their vision efficiently while discouraging independent thinking. When this happens, the startup grows rapidly in the short term but collapses after the founder leaves. By then, everyone has learned how to execute the founder’s decisions, but all the crucial decision-making knowledge stays locked in the founder’s head.

Organizations preserve different types of knowledge in different ways: Explicit knowledge like documented procedures transfers easily through manuals, but the most valuable knowledge is often tacit: the judgment developed through experience that’s difficult to articulate or transfer. In organizations that are overly dependent on their founder, this knowledge never gets distributed. Marquet solves this by distributing not just authority but the knowledge people gain as they make decisions and articulate their reasoning. Research confirms that this matters: Startups become 50% less likely to successfully adapt after losing their founder, but when knowledge is distributed, the ability to adapt persists even as people come and go.

The “Leader-Leader” Model on the Santa Fe

Marquet explains that his experience on the Santa Fe demonstrates all three advantages of the “leader-leader” model. When he took command in January 1999, the submarine was the worst performer in the fleet, and he had six months to prepare it for a high-stakes combat exercise. Instead of introducing stricter discipline or closer supervision, he gave decision-making authority to the chiefs and crew, changed how people communicated, and made them responsible for tracking their own work. Marquet explains that the transformation was dramatic: Crew members who had operated in “whatever they tell me to do” mode became active problem-solvers who noticed inefficiencies and proposed improvements.

Reenlistments also skyrocketed from three to 36 in one year—people wanted to stay because their work finally required and valued their full capacity. The submarine went from worst to first in the fleet, earning the Arleigh Burke Fleet Trophy for the unit that improved in battle efficiency the most in a year. It produced far more promoted officers and senior enlisted personnel than comparable ships, and 12 years after Marquet took command, one of his former weapons officers became the Santa Fe’s commanding officer, while several other officers went on to command their own submarines.

After Marquet left, the Santa Fe continued performing at the highest level, earning the Battle “E” award for most combat-effective submarine three times in 10 years.

Different Types of Excellence May Require Different Leadership Structures

Marquet’s results on the Santa Fe demonstrate operational excellence: getting complex equipment to run reliably, maintaining combat readiness, and keeping people engaged in executing well-defined procedures. But achieving different types of organizational goals may require different leadership structures. Consider Bell Labs, the industrial research laboratory that earned 10 Nobel Prizes and invented the transistor, laser, and solar cell. Bell Labs’ work was fundamentally different from the operational procedures Marquet’s crew performed. Neither Bell Labs’ researchers nor its development engineers were executing established protocols—they were creating new knowledge and translating it into new products.

Bell Labs’ director Mervin Kelly structured the organization accordingly. Kelly gave extraordinary freedom to an elite group of researchers who received no deadlines, filed no progress reports, and had the freedom to choose their own methods and collaborators. Kelly would identify important problems within the field of telecommunications, then give researchers complete freedom to solve those problems, sometimes only checking back years later.

Outside of this research group, many Bell Labs employees were development engineers who translated research discoveries into manufacturable products. These engineers worked within more traditional hierarchies, with managers who assigned projects and set goals. Yet within their assigned work, these engineers had significant technical autonomy. They made complex decisions balancing functional performance, manufacturing costs, and service requirements, and could freely consult expertise across the organization.

This suggests that “distributing authority” isn’t an all-or-nothing choice—it’s multiple choices about different types of authority. You can distribute technical problem-solving (how to do the work) while centralizing strategic decisions (what work to do). You can give people access to organizational expertise while maintaining hierarchical reporting structures. The question isn’t whether to distribute authority, but which types of authority to distribute, to whom, and under what conditions.

How to Implement the “Leader-Leader” Model

Changing from a “leader-follower” model to “leader-leader” instead requires changing three fundamental aspects of how your organization operates. Simply urging people to “take ownership” won’t do it—the structure itself must change. Marquet identifies three elements that must all work together: control, competence, and clarity. Giving people control without competence creates chaos. Giving people control and competence without clarity means they’ll make individually good decisions that don’t serve the organization’s mission. Building competence and clarity without giving actual control wastes people’s potential. In this section, we’ll take a closer look at how to equip workers with all three.

1. Give People Control: Push Authority to Where Information Lives

First, Marquet argues that decision-making authority should rest with the people doing the work. In many organizations, information flows up to decision-makers, then orders flow back down. Instead Marquet recommends giving decision-making authority to the people who have access to the best information—which requires changing the rules that dictate who makes which decisions. On the Santa Fe, Marquet gave the chiefs (senior enlisted personnel equivalent to middle managers) real authority over their teams. Previously, approving a crew member’s leave request required 14 steps, but Marquet gave authority to the chief responsible for that member. This gave chiefs control over their crews’ schedules, watch rotations, and training timelines.

(Shortform note: Research suggests that pushing decision-making authority to those closest to the information works not because more information leads to better decisions, but because it helps avoid information overload at higher levels of the organization. Even small amounts of extra information can worsen decision quality—people struggle to identify which details matter. The problem gets worse for people with expertise on a topic, like managers making decisions about frontline work: Their existing knowledge interferes with using new information effectively. Marquet’s approach may work because it places decisions where relevant information is naturally concentrated and irrelevant details are automatically filtered out.)

To give people more control, language must also change. Marquet required crew members to state their intentions rather than ask permission. Instead of “Request permission to shift to backup power,” an officer would say, “Captain, I intend to shift to backup power.” Asking permission keeps control with the person being asked—they make the decision. But stating an intention puts control with the speaker—they make a decision and inform their supervisor. Similarly, Marquet believes leaders should resist providing solutions when problems arise, since that keeps control with the leader. Instead, Marquet would ask, “What are the options?” This gave the crew a cue to think through alternatives, rather than wait to be told what to do.

(Shortform note: Some studies affirm that the way we speak shapes our thinking, and cognitive science research suggests that the benefits of stating intentions may extend beyond the moment of communication: People who speak languages that emphasize active grammatical constructions tend to assign accountability more readily. So when workers say “I intend to,” they strengthen their sense of agency and build a mental record of themselves as decision-makers. This accumulated self-perception may be why Marquet’s changes stuck: Crew members weren’t just trained in new procedures; they developed new perceptions of their identity and ability.)

Finally, Marquet contends that to give people more control over their work, it’s crucial to invert the traditional expectations on who tracks and reports on progress. In most organizations, supervisors monitor their team’s work by checking in on projects and asking for updates. This puts supervisors in control. Marquet eliminated top-down monitoring. Instead, people tracked their own work and proactively reported progress. Department heads would tell the executive officer what they’d accomplished and what they planned to do next, rather than waiting to be asked.

(Shortform note: Research validates Marquet’s approach of having people track their own work rather than being monitored: The more frequently people monitor their progress toward goals, the more likely they are to succeed at reaching those goals. The effect is even stronger when progress is physically recorded or publicly reported. The framing of this tracking matters, too. While looking forward at how far you still have to go can feel discouraging, looking backward at how far you’ve come can motivate you to keep going. When Santa Fe crew members reported their progress—telling supervisors what they’d accomplished—they naturally emphasized their gains, which likely reinforced their motivation to keep making progress.)

2. Build Competence: Ensure People Can Use Control Well

Second, Marquet argues that as people gain decision-making authority, they must also develop the technical knowledge and the judgment to use that authority wisely. Without this foundation, distributed control becomes dangerous rather than empowering. Because of this risk, Marquet recommends teaching people to take deliberate action by pausing before acting, which consciously engages their minds. He instituted this procedure after a crew member violated a safety rule because they were operating on autopilot, requiring each person to pause and state what they intended to do. If someone said, “I’m about to turn the valve to the right,” saying it out loud required them to consciously confirm this was correct.

Importantly, Marquet found that workers’ learning must be active, not passive. The Navy traditionally uses briefings where someone reads procedures aloud, and everyone listens. On the Santa Fe, Marquet replaced these by having the team leader ask members questions about their roles, the operation, and potential problems. The team had to demonstrate their readiness for an operation, and only if the team leader was satisfied would they proceed. Building competence also requires making learning a core expectation. Marquet found that this expectation creates a reinforcing cycle: Greater competence enabled more delegation, which required more learning, which built more competence.

Finally, Marquet argues that to help workers build the competence to use decision-making authority wisely, leaders should specify goals but let people determine the methods they’ll use to achieve them. When Marquet set an objective on the Santa Fe—like extinguishing a simulated fire as quickly as possible—he didn’t prescribe how to do it. The crew organized themselves and found the most effective approach. This built competence because people had to actively problem-solve rather than follow prescribed steps.

Competence Develops in Stages

Research on how medical students actively build the competencies Marquet refers to—both technical knowledge and good judgment—suggests these abilities develop in stages. First, students have to progress from analytical to intuitive decision-making. Early in their training, students use slow, deliberate analysis for every decision, consciously working through diagnostic steps and treatment options. But as they gain experience, they build a library of “illness scripts”—saved patterns they can quickly compare to new cases. Expert physicians rely heavily on this fast, intuitive pattern recognition, which lets them make efficient and confident decisions about their patients’ care.

This progression from analytical to intuitive thinking requires psychological safety. Medical students report that although they understand that medical uncertainty exists in theory, during clinical training they fear appearing unknowledgeable, which prevents them from asking questions or admitting gaps in their knowledge. Research shows that students can better develop competence when they can practice in “low stakes” environments, where they’re expected to learn and to make mistakes. Marquet’s emphasis on creating a learning culture reflects this need—people must feel comfortable admitting uncertainty and asking questions to move from analytical novice to intuitive expert.

Finally, students must build the competence to perform under realistic workplace pressures. Even medical students who can reason through problems successfully when focused on one issue struggle when interrupted or managing multiple demands. Many students report they’ve never been trained to maintain good judgment while juggling competing priorities, despite this being constant in actual practice. This explains why Marquet’s approach of gradually increasing responsibility works—people need time to develop not just decision-making ability, but the capacity to maintain that ability under pressure.

3. Create Clarity: Align Decisions With Organizational Goals

Third, Marquet explains that everyone must understand what the organization is trying to accomplish so that they can make decisions that serve those goals. Without this shared understanding, distributed decision-making produces effort in conflicting directions. On the Santa Fe, Marquet found that a critical step to achieving this clarity was to reframe the goal from avoiding errors to achieving excellence. The crew meticulously followed checklists to avoid errors, but this left no room for exceptional performance. Marquet redefined the goal as excellence. This changed what people optimized for. Instead of asking “Did I follow the procedure correctly?”, they asked, “Did we accomplish the mission effectively?”

(Shortform note: Marquet’s approach to creating clarity aligns with research on what makes employees both happier and more productive. In The Happiness Files, Arthur C. Brooks cites research showing that two factors strongly predict workplace happiness: “organizational alignment” (when the company’s external mission matches its internal culture) and “innovation” (managers’ openness to input and ideas). Clarity helps with both. By ensuring everyone understands shared goals, you create alignment between what the company claims to value and how it operates. When you reframe the goal as excellence rather than compliance, you signal that managers want employee input and creative problem-solving, not just rule-following.)

Establishing clarity also requires sharing explicit decision-making criteria to guide people’s choices. Marquet and his officers developed guiding principles for the crew, including values like initiative, innovation, technical knowledge, and courage. These principles gave crew members concrete criteria for decisions. For instance, valuing innovation meant that when someone saw an inefficient process, they should propose improvements. Finally, Marquet contends that it’s crucial for leaders to help people envision the future they’re working toward. Marquet had officers write their own performance evaluations looking forward several years, describing what they would accomplish and giving them concrete targets to work toward.

Clarity at Different Timescales

Marquet focuses on aligning individual actions with organizational goals over a career span—typically years or decades. But some communities demonstrate what these same practices look like when oriented toward much longer timescales. For example, Inuit communities in the Arctic have developed ecological knowledge gathered through centuries of sustained observation and experience. This knowledge system provides decision-making criteria for how to interact sustainably with the environment. When Inuit hunters decide when and where to harvest marine mammals, they’re guided by accumulated observations about animal behavior, migration patterns, and ecosystem relationships.

This knowledge helps communities make choices with multigenerational consequences: decisions about sustainable harvesting practices, whether to support or oppose development projects that could impact ecosystems, and how to adapt to environmental changes while maintaining cultural practices. This helps Inuit knowledge holders orient their decisions toward a future measured in lifetimes.

For instance, they’ve long understood that bowhead whales can live twice as long as humans—which was confirmed when researchers found century-old stone harpoon points embedded in harvested whales. This understanding shapes decisions about harvesting in ways that account for whales’ extended life cycles. Individual decisions about harvesting ensure that future generations will continue to have access to traditional foods, which creates a different kind of clarity about purpose from what Marquet describes. Both approaches use guiding principles and future-envisioning to inform decisions, but the timescale fundamentally changes what those principles emphasize and what that future looks like.