The most innovative companies tend to push decisions as far down in the organization as possible, giving people at all levels the opportunity to move fast, utilize their creativity, apply their intellect, and assume responsibility.

—Collins, Jim. [1]

Principle #9 – Decentralize Decision-Making

Surviving and thriving in today’s business environment requires quick and efficient decision-making. Disruptive technology, high interconnectedness, and intense competition mean opportunities are fleeting, problems escalate rapidly, and information moves faster than traditional organizational structures can handle. Creating complex Solutions that capitalize on business opportunities requires swift exploration and experimentation within a short timeframe. This necessitates an approach to decision-making that can handle complexity while prioritizing speed.


In many contexts, centralized authorities take decisions that should instead be made by knowledge workers who have both the local information and the technical skills to decide. Often this is due to historical top-down cultures, a lack of trust, and an emphasis on control and perceived efficiency over agility and responsiveness. But when managerial input is necessary for most decisions, it causes a queue of decision requests that block progress and hinders and delays—rather than facilitates—the performance of the people doing the work (Figure 1).

Figure 1. Excessive centralization of decisions creates queues and bottlenecks and causes delays
Figure 1. Excessive centralization of decisions creates queues and bottlenecks and causes delays

In addition, this approach often fails to yield optimal decisions due to the limited context of those making the centralized decisions and the lack of local information available, which the passage of time can further compound. Since solution development requires multiple decisions to be made on an ongoing basis, this centralized approach ultimately impairs business agility throughout the entire organization. Therefore, an alternative approach is needed. The remainder of the article outlines proven and fundamental concepts for decentralizing decision-making in a SAFe Enterprise.

Decentralized Decision-Making Accelerates Flow

Solving the decision bottleneck requires restructuring the decision-making process itself. In his book, Team of Teams [2], General Stanley McChrystal states, “Individuals and teams closest to the problem, armed with unprecedented levels of insights from across the network, offer the best ability to decide and act decisively.” This is the essence of decentralized decision-making and involves transferring responsibility and authority to the worker doing the task where the most in-depth knowledge resides.

Decentralizing decision-making offers many benefits:

  • Responsiveness and speed – The pace of decision-making impacts the speed at which the organization can deliver value to its customers. This is often the most significant advantage of decentralized decision-making in complex solution development tasks.
  • Better decisions – Decentralized decisions are made by those directly exposed to the context of the problem they are trying to solve. This is critical to correctly assessing the situation and making the best decision possible.
  • Employee motivation – Decentralized decision-making encourages involvement and a sense of ownership among knowledge workers. It fosters engagement and enables a significant degree of autonomy.
  • Alignment– Decentralized decision-making enables teams and individuals to obtain a more comprehensive view of the mission and focus on developing the best solutions to achieve it rather than following orders.
  • Collaboration – Many tactical decisions happen between individuals and teams working towards a common goal. Such decisions are often better resolved by those involved than by a centralized role.
  • Innovation – Self-directed workers are more likely to discover innovative solutions to problems. This is due to their higher autonomy and a better understanding of the context of their work.
  • Joy of work – People who build the world’s most important systems are intrinsically motivated to do so. That is their purpose. Being empowered to make the decisions they know they are capable of making is personally satisfying and helps restore the joy in the work.

Enabling Decentralized Decision-Making with Competence and Clarity

In his book ‘Turn the Ship Around’ [3], retired United States Navy Captain David Marquet shares the invaluable lessons he learned as Commander of the USS Santa Fe submarine. Once a laggard of the fleet, this vessel underwent a drastic transformation into a standout success story, all thanks to the pivotal shift of moving decision-making power to its crew members. Marquet emphasizes the importance of creating an environment that enables the effective delegation of such responsibilities. To achieve this, he highlights two vital components: technical competence and organizational clarity.

Figure 2. Increasing technical competence and organizational clarity enables decentralized decision-making
Figure 2. Increasing technical competence and organizational clarity enables decentralized decision-making

Cultivating Technical Competence

Marquet notes, “Have you tried to divest control without first ensuring your organization is competent to handle more decision-making authority? I learned the hard way that control without competence is chaos.” Decentralizing decisions when lacking technical competence often leads to costly errors, frustration, and demotivation. A leader that aims to empower decentralized decision-making must address the following questions:

  • Which areas in the organization lack the necessary competence?
  • What resources are required to grow competence in those areas, and what support must I provide?
  • What impediments stand in the way of acquiring competence, and what do I need to do to eliminate them?

Growing competence requires skills training while also benefiting from mentorship, knowledge exchange, and the immediate utilization of newly acquired knowledge in practical scenarios. Agile’s incremental and iterative nature also helps advance technical competence and quickly validate new learnings.

In SAFe, competency can be measured using the seven core competency assessments (See Measure and Grow). Practices like collective ownership, pair work, and peer reviews assist on this path (See Built-In Quality). Additionally, growing T-shaped skillsets and leveraging Communities of Practice furthers the competence of knowledge workers.

Providing Organizational Clarity

When decision-making authority is distributed, it becomes necessary for everyone to understand their roles and responsibilities, as well as the organization’s values, purpose, and mission. Without such clarity, executing with competence and yet moving in the wrong direction is possible. Marquet emphasizes that leaders should communicate the organization’s intent to enable every employee to grasp the overall objectives and how their roles contribute to achieving them.

Many SAFe elements support organizational clarity, but some are particularly instrumental in achieving it.

  • Strategic Themes: expressed as Objectives and Key Results (OKRs), provide the business context for decision-making.
  • Solution Vision and Roadmap: drives energy, clarity, and alignment across all levels of the organization.
  • PI Planning: aligns teams and stakeholders to a shared mission and vision.
  • Backlogs: providing a prioritized list of work that might be done, they ensure focus and a sense of direction

While leaders play a significant role in creating organizational clarity, it is not solely their responsibility. In a complex organization where multiple teams collaborate towards a shared goal, those teams must collaborate and communicate directly with the teams around them.

The Cascading Nature of Decentralized Decision-Making

Any significant decision will trigger the need for further decisions at multiple levels in the organization. For instance, a strategic decision made by the portfolio to utilize artificial intelligence for customer engagement will require decisions by Product Management, the System Architect, and, ultimately, the teams in developing a set of features to realize this goal (Figure 3). The result is a “daisy chain effect” where a decision taken at any level appears decentralized to the level above it but centralized to the level below. Indeed, as the old saying goes, ‘One person’s ceiling is another person’s floor.’

Figure 3. Decentralized decision-making requires communication of intent
Figure 3. Decentralized decision-making requires communication of intent

As Figure 4 illustrates, a critically important aspect of decentralized decision-making is the communication of intent. Simply put, the decision maker must be clear to those impacted by the decision about the ‘what’ and the ‘why’ behind it.

But the responsibility doesn’t stop there. Those impacted by the decision will, in turn, make further decisions based on this information. They then have the same obligation to communicate that intent of the decision up and down the chain. In this way, all parties are informed of the intent of all decisions and are aligned around the agreed-to purpose behind these decisions.

Anatomy of a Decision

Understanding how decisions are made helps knowledge workers approach the decision-making process more clearly and intentionally. And while it’s motivating for a knowledge worker to be empowered to make the decisions needed to move forward, empowerment can be sustained only when those who make these decisions act responsibly.

Figure 4 illustrates the thought processes and actions involved when an individual is faced with a decision. From CEO to rank-and-file employee, the process is the same:

Figure 4. Anatomy of a Decision
Figure 4. Anatomy of a Decision

When considering a decision, the decision-maker must consider five factors which will assist them in taking the appropriate action. These are:

  • Information. Do I have insufficient information on which to base a sound decision? Do I have the right stakeholders to inform this decision?
  • Authority. Do I have, or can I get, the authority (legal, policy, hierarchical, etc.) to make the decision?
  • Competence. Do I have, or can I access, the technical competence needed for this decision?
  • Clarity. Do I have sufficient clarity of the organizational purpose for this decision? If not, can I access it?
  • Impact. What will be the impact of this decision? Am I prepared to address the concerns of those who may disagree with the decision? Do I understand the economic framework that’s relevant to this decision? Is this decision associated with safety, privacy, policy, security, or other significant factors?

Based on the above, there are four possible actions a decision-maker can take, and each is described below.

Escalate the Decision

If, after considering all these factors, the decision maker is uncomfortable making the decision and thereby assuming responsibility for the outcomes, it is appropriate to escalate the decision to those with greater authority and responsibility.

Perhaps contrary to some thinking, escalation is not a failure pattern. Instead, it is an essential mechanism in a functioning organization. It keeps the process moving and highlights effective decentralized decision-making by illustrating that not every decision is appropriate for decentralization at any node in the chain. Too much escalation however, signals a weak decision-making process and a lack of delegated authority. No escalation is not a sign of perfection; rather, it may indicate that people may be making decisions beyond their purview or are simply taking orders that are not to be questioned.

Defer until the Last Responsible Moment (LRM)

Sometimes, the needed information is not readily available to the decision-maker. In this case, deferring a highly impactful decision may be appropriate. Indeed, the practice of Set-Based Design and SAFe Principle #3 – Assume variability, preserve options, gives insights into the economic basis for deferring a decision. This is all well and good up to the point of the LRM— the moment when the cost of not taking a decision exceeds the cost of making a potentially wrong decision. But decision-makers should be aware that value flow may be blocked while a decision is deferred.

Take or Delegate?

Two choices remain, and these are the primary focus of decentralizing control, which is whether to take or delegate the decision. Simply put, the decision-maker can take a ‘centralized decision’ on behalf of those below them in the chain or ‘decentralize the decision’ to a level below. Reinertsen [4] emphasizes the importance of “balancing centralization and decentralization.” Derived in part from that work, the considerations in Table 1 can help with identifying the best course of action.

Based on these considerations, a thinking tool is shown in Figure 5. As an exercise, consider three decisions that you are currently facing, rate them based on the frequency, time-criticality and economies of scale, assigning values of 0, 1 or 2. A value of 1 may be useful where a decision is not entirely ‘Yes’ or ‘No’ in each category, but rather somewhere in between. The next step is to add the values and look at the total . A score of 0 to 3 would suggest a centralized decision (take the decision) and a score of 4 to 6 a decentralized decision (delegate the decision).

Figure 5. A simple decision-making ‘take vs delegate’ thinking framework
Figure 5. A simple decision-making ‘take vs delegate’ thinking framework

In the example above, the decision ‘on the test approach for a feature’ would be a good candidate for delegation since those types of decisions are frequent and time critical. The remaining two decisions are less frequent and have economies of scale. These may be appropriate for the decision-maker to ‘take’ to achieve those benefits.

Take the Decision

If, after consideration of all these factors, the decision maker is comfortable making the decision themselves, the decision maker should take the decision. ‘Take’ is the act of selecting a specific course of action. It is the unique choice that neither defers, escalates, or delegates. In this case, ‘the buck stops here’ and a specific course of action is decided. Taking the decision triggers two important responsibilities.

  • Take responsibility for outcomes. It’s an empowering thing to be able to take the decisions appropriate to fulfilling and increasing one’s duties. But the action requires assuming responsibility for the outcomes the decision ultimately creates.
  • Communicate the decision and intent to affected stakeholders. The decision affects many stakeholders, both within and outside of the decision chain. Some stakeholders may disagree with the decision even if their input has been heard. Understanding intent will help them move forward with the actions they need to take.

Delegate the Decision

Escalate, Defer, and Take are three important options. But decentralized control requires moving the decision to where the information is (moving the meter in Figure 2). Therefore, delegating the decision to others closer to the task is the critical element of decentralizing decision-making. However, delegating the decision does not abrogate the decision maker’s responsibility for the decision. As Marquet notes

“While the commanding officer may, at his or her discretion… delegate authority to subordinates for the execution of details, such delegation of authority shall in no way relieve the commanding officer of continued responsibility for the safety, well-being, and efficiency of the entire command (Section 0802 United States Navy Regulations”- [3]

While few of us are involved with the extreme assignment of operating nuclear submarines, the point is the same. If the delegator shirks that responsibility, then when a mistake or failure does occur, those with more authority are likely to react in such a way as to increase, not decrease, centralized control.

Enabling Decentralized Decision-making with Psychological Safety

How an organization makes decisions, responds to them, and handles their aftermath is determined by its culture, and therefore transitioning towards decentralized decision-making is a cultural transformation.

The organization’s handling of risk and error is central to its decision-making ethos. Consequently, the most challenging obstacle to decentralizing decision-making is a culture that discourages open and honest communication of problems or even penalizes mistakes. In such an environment, employees avoid taking on decision-making responsibilities entirely or misuse them, which exacerbates problems. Decentralization requires psychological safety, meaning employees can voice their opinions, question beliefs, express concerns, or make mistakes without fear of retribution or embarrassment.

In her book, “The Fearless Organization” [5], Amy Edmondson notes: “Low levels of psychological safety can create a culture of silence. They can also create a Cassandra culture – an environment in which speaking up is belittled, and warnings go unheeded.

It is important to note that mistakes from poor decisions can have severe consequences for the customer and the business. Viewing them as ‘learning opportunities’ does not diminish their impact. Appropriate guardrails and checks and balances must be in place. However, a common problem is that these mistakes are not surfaced due to a fear of retribution. It’s often the hidden mistakes that cause a plane to crash or an oil rig to blow up.


Delivering value in the shortest sustainable lead time requires decentralized decision-making. Decentralizing decision-making reduces delays, improves product development flow and throughput, and facilitates faster feedback and more innovative solutions. Higher levels of empowerment are an additional, tangible benefit.

Creating the right environment is critical to unlocking this new approach. Leaders must ensure that those in decision-making positions have the clarity and competence needed to make the best choices for the organization and that the decision-makers recognize their responsibility in communicating the intent of each decision made.

Furthermore, new information constantly emerges in solution development, presenting challenges and opportunities to add value. This dynamic nature of the work affects decision-making by altering the available options, their impact, and the situation’s urgency. As a result, the balance between centralized and decentralized decision-making must be adjusted regularly.

Learn More

[1] Collins, Jim and William Lazier. BE 2.0 (Beyond Entrepreneurship 2.0): Turning Your Business into an Enduring Great Company. Penguin Random House, LLC. 2020.

[2] McChrystal, General Stanley; Collins, Tantum; Silverman, David; Fussell, Chris. Team of Teams: New Rules of Engagement for a Complex World. Penguin Publishing Group, 2015.

[3] Marquet, L. David. Turn the Ship Around! Penguin Publishing Group, 2018.

[4] Reinertsen, Donald G. The Principles of Product Development Flow: Second Generation Lean Product Development. Celeritas Publishing, 2009.

[5] Edmondson, Amy C. The Fearless Organization. Wiley, 2019.


Last update: 14 August 2023