Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Articles The Flow System: Leadership for Solving Complex Problems

The Flow System: Leadership for Solving Complex Problems

This item in japanese

Key Takeaways

  • Most of the complex problems that leaders and teams deal with in organizations are human derived.
  • Complex problems require different techniques than those used for simple and complicated problems.
  • Systems thinking works best mainly for closed systems, and complexity thinking works best mainly for open systems.
  • Middle management can be repurposed into a new role known as “Boundary Spanners” that enable the leadership of multiteam systems, the key to scaling agility.
  • Leadership is not reserved for those leading an organization; it’s distributed.

The book The Flow System: The Evolution of Agile and Lean Thinking in an Age of Complexity by John Turner, Nigel Thurlow, and Brian Rivera supports organizations that operate in complex environments to increase agility. Built on a foundation of the Toyota Production System, it elevates Lean Thinking in an age of complexity by combining complexity thinking, distributed leadership, and team science into what the authors call the Triple Helix of Flow, which organizations can use to become more innovative, adaptive, and resilient.

InfoQ readers can download an extract of The Flow System

This book Q&A with John Turner, Nigel Thurlow, and Brian Rivera is published in two parts. Part 2 (this article) dives into the three helixes of complexity thinking, distributed leadership, and team science. Part 1, Getting Fast Customer Feedback and Managing Flow, explored the importance of quality, getting fast feedback from customers, the concept of flow, and The Flow System. 

InfoQ: What are complex problems and how can we address them?

John Turner: Complex problems are different from traditional, predictable, complicated problems that we deal with almost daily. Complex problems involve multiple agents, organisms, or systems interacting in different ways and at differing levels. Complex problems introduce incongruence, a level of not knowing, and unpredictability resulting in a constant feeling of uncertainty throughout the whole process.

The literature has identified three types of problems: simple, complex, and wicked. Churchman initially identified these types of problems in 1967. Simple problems have a consensus on the problem and solution; complex problems involve agreement on the problem, but not on the technique for solving the problem. Wicked problems include those in which there is no consensus on the problem or the solution. An example of a simple problem includes an algebraic solution (3X + 2 = 5). Dealing with a pandemic’s impact provided all of us with a real example of a complex problem. We knew what the problem was, but there was much scientific and political debate on managing the problem and what solution was necessary to resolve the problem on a global scale. An example of a wicked problem includes global warming. 

Unable to be compartmentalized into smaller sub-problems because there are too many unknowns, complex problems require different techniques than those used for simple and complicated problems.

This begins the process that we call complexity thinking.

Step 1: Understand the characteristics of complex systems.

Step 2: Have a worldview or perspective that systems, entities, or events are complex adaptive systems.

Step 1 gives one the knowledge to acknowledge the type of environment or problem one is dealing with. We find that the Cynefin framework is one of the best tools to identify the problem one is facing. The Cynefin framework and the folks at Cognitive Edge (Dave Snowden & colleagues) also provide steps for working in the complex domain.

Step 2 involves recognizing the many components involved in complex adaptive systems. These components are non-linear, unpredictable, and unrepeatable; they include precise patterns relevant only to themselves that are not present in many other problems. Dealing with complexity involves experimentation and explorative activities. Sensemaking is required to learn about the complex adaptive system’s unique components. 

It is necessary to mention that most of the complex problems that leaders and teams deal with in organizations are human-derived. We create most of the complex problems that we have to face. This is why it is vital to integrate the Triple Helix of Flow components; addressing only one helix will not prevent complexity. Only after integrating the three helixes across the organization is flow achieved, resulting in value for the customer. 

The Triple Helix of Flow requires the integration of the three helixes (complexity thinking, distributed leadership, team science) to achieve seamless flow from design to delivery. This example, one that repeatedly plays in today’s organizations, identifies how flow is achieved using the Triple Helix of Flow concept and integrating each of the three helixes holistically (for more information see The Evolution of Lean Thinking). Flow cannot be achieved without this integration.


InfoQ: What problems do you see when using systems thinking to understand organizational systems?

Turner: No problem, when dealing with mostly closed systems. Systems thinking mostly applies to closed systems. When dealing with complex environments and problems, the tables shift from mostly closed systems to mostly open systems. Complexity thinking is designed for working with open and complex adaptive systems. Systems thinking is primarily designed for working with closed systems. 

There is an area of overlap between closed systems and open systems. This area is best described as the Liminal Boundary in the Cynefin framework. Systems that are more closed than open are best managed using systems thinking techniques. In contrast, systems that are more open than closed are best managed using complexity thinking techniques. It is a matter of identifying what type of problem you are dealing with first, then identifying the proper tools and techniques for that type of problem. This is the general idea behind complexity thinking. First, identify what type of problem/environment you are facing. 

However, some systems thinking strongholds believe all systems can be managed using systems thinking techniques and completely dismiss complexity techniques. These two paradigms, systems and complexity thinking, evolved around the same time in different disciplines. They remained separate over the years and have only just begun to merge some techniques. I have identified this merger as components of complex adaptive systems where ST contributes some components and CT contributes others, and both share a few components. I published this overlap in an article, but there is still much work required to identify which techniques originated from which paradigm ( I am working on this research, albeit slowly. Nevertheless, for this conversation, we did discuss this briefly in the book. We placed this overlap in the Cynefin framework’s liminal boundary, positioning ST with the complicated domain (closed systems) and CT with the complex domain (open systems).  

This question also hints at the boundaries or constraints placed on a system. When a system is bounded to the point that the complexity elements have been removed, systems thinking techniques will work - as long as the boundaries remain intact. Complexity thinking techniques are in order when systems are bounded with the components that lead to complexity. When dealing with complexity, you do not want to frame the system to the point that the complexity elements have been removed. In doing so, if optimized in its bounded state, the system will fail to function as desired when reintroduced to complexity because the optimization took place without the factors that caused complexity. How boundaries or constraints are placed around a system is essential when dealing with systems and complexity.

InfoQ: What are the limitations of today's leadership practices?

Turner: One of the most significant limitations in today’s leadership practices is the lack of development. Most leadership training is disguised as leader education. These training efforts also do not include time for emerging leaders to practice their newly learned leadership skills. Without practice and the freedom to fail during the developmental stages, it is nearly impossible for emerging leadership to master skill. Another problem with leadership development is that most programs deliver training to everyone the same way. Most leadership development programs were initially designed as “one-size-fits-all” training. In The Flow System, we make great efforts to design leadership and team development around the contextual setting. 

We view leadership as a collective construct, not an individual construct. We incorporate the team as the model of leadership, and individual team members as leaders using a shared leadership model. This collective becomes the organization’s leadership model, from the lower ranks up to the executive level. This is how leadership becomes more distributed within an organization, focusing on distributed leadership rather than several other individualistic leadership theories. 

The team as the leadership model is just the first step in the overall distributed leadership transformation, but this concept is essential for people to realize. Leadership is collective rather than individualistic. This is also necessary when dealing with complex problems. Diversity of skills and knowledge is required to address complexity, often by using teams to achieve such diversity. There is no room for individualistic efforts. There is no single person with the skills to address complexity. There is no individual leader who can manage complexity. It is essential to distribute leadership when dealing with the complex domain.

Brian Rivera: To add to John’s points above and from my recent experience in several FORTUNE 50 companies and the U.S. Navy, I think the biggest limitation of today’s leadership practices is the focus on the individual. This focus on improving the individual agent in a complex adaptive system is wrong. Instead, we need to focus on improving interactions and this is accomplished through focusing on teamwork at every level of the organization. Let’s face reality here: most “leaders” are promoted on their technical skills, not their leadership skills.  

Compounding this problem is the Agile community’s fixation and jargon that focuses business leaders to apply “Agile” at the “team level.” To be perfectly clear, those who need to be working as interdependent teams are not always those who are closest to the customer. The reality is that those people who need to work as interdependent teams are those managers and leaders nearest to and in the C-Suite and Board. 

Nigel Thurlow: The intrinsic motivation for most people who term themselves “leader” is self worth and value. That is, “What do I get out of it?” The focus is more “me” than “us”, leading to narcissistic behavior characteristics emerging. This is then the “leadership” the company follows.

Changing the focus to customer first value goes some way to addressing this, but it is not enough. Leadership needs to change and adopt a collective leadership model defined in the work we present in distributed leadership. This does not mean the lunatics start running the asylum, so control is still needed, which we provide for with enabling constraints as part of managing constraints; guardrails, if you like.

Leadership are teams. Leaders need to operate as teams and not say they’re a team, but continue to operate in silos. A key test of a team is that each member knows what each member is doing at any one time. My own work has shown most leaders gather together in a meeting and then go back to running their own world with little reference to the other members of the “leadership” team.

Understanding the concepts behind the Triple Helix of Flow will enable future leaders to emerge at all levels in an organization, and for that “Collective Leadership” to optimize the “Flow of Value” to the customer. 

InfoQ: How can we develop leadership in organizations?

Turner: Using our distributed leadership model, we begin by focusing on self-leadership and self-efficacy skills at the individual level. At the team level, we concentrate on developing teams to be self-organizing and autonomous entities using a shared leadership model. Functional leadership develops boundary spanners at the multiteam system level. At the executive level, we embed existing leadership functions with those found in strategic, instrumental, and global leadership. In its totality, this makes up the distributed leadership model that we introduced in The Flow System.

Leadership is contextual at all levels. In being contextual, the necessary skills and competencies required for each level (individual, team, MTS, Executive) would be identified and listed in the boxes shown above in the figure. Once these necessary skills are identified for a given organization, leadership development is designed around developing these skills at each level. It is critical for developing leadership and teamwork skills (interpersonal skills) at all levels in an organization. Leadership is emergent and occurs at all levels within an organization. Emergence occurs when all levels are involved in the leadership process. Like a stack of blocks or a set of Jenga blocks, if you remove one level of blocks, all remaining blocks above the level removed fall. Ignoring developing leadership and teamwork skills at any level introduces inhibiting constraints within an organization, reducing the delivery of value (flow) to the customer. 

InfoQ: What advantages does shared leadership bring?

Turner: Shared leadership provides a psychologically safe environment where members are free to share responsibilities, collaborate, make decisions collectively, and provide supporting functions during taskwork. Shared leadership does not just occur by putting individuals together and calling them a team. Team members must be trained in teamwork leadership skills before achieving a state in which authentic shared leadership occurs within a team. 

One advantage of achieving a state where teams can operate using the shared leadership model is that there is no need for team leaders or external leaders; the team leads as a collective. This reduces the total number of management or oversight positions as the team’s next level is the boundary spanner. Organizations only need team members and one boundary spanner for each multiteam system. 

When operating in complexity, no one leader can direct teams effectively. By utilizing shared leadership, with team members trained in the requisite skills, teams function as a diverse collective, better able to address complexity than a team led by a single team leader. 

Creativity is a team construct in which team members interact with one another freely to exchange ideas. When external leaders, or a team leader, disrupts this free exchange of ideas through attempts to direct the process, creativity becomes inhibited and fails to be complete. Shared leadership provides a mechanism for teams to freely exchange ideas, providing a better structure for teams to be more creative in their outcomes. 

InfoQ: What's the difference between taskwork and teamwork?

Turner: Team functions require both teamwork and taskwork. In the most basic form, teamwork deals with team members’ interpersonal activities, and taskwork deals with team members’ technical aspects to complete their tasks. 

Most teams are composed of team members capable of performing the necessary tasks to complete the assigned team’s job. What is not always present, and where many teams become disrupted, is that these same team members are rarely trained on functioning as a team. Teamwork training is the key differentiator in research. Studies have shown that teams trained in teamwork skills outperform teams not trained in teamwork skills, everything else being equal.

InfoQ: How can teams increase their effectiveness?

Thurlow: We derived a formula for team effectiveness that highlights the essential components for team effectiveness. 

This formula is provided: TE = (TW + IP) + TK f (TP + AP) + PF + CV

Team Effectiveness = (Teamwork + Interpersonal Process) + Taskwork f (Transitional Phases + Action Phases) + Performance + Customer Value 

(TW) Teamwork involves identifying the (AP) action phases, (TP) transition phases, and (IP) interpersonal processes. Team members should be capable of self-managing the team’s core processes (coordination, cooperation, cognition, conflict, communication, coaching).

(TK) Taskwork involves the knowledge, skills, and abilities required of team members to complete their assigned task. Teams need to be composed (designed) of team members who have the requisite knowledge and skills necessary for the team’s task.

(PF) Performance is an outcome measure involving a number of different matrices (e.g., quality, quantity, time).

(CV) Customer value represents the value delivered to the customer. However, this also includes team member satisfaction as team members are an internal customer.

InfoQ: How can we set up and lead organizations consisting of multiple teams?

Turner: Even though this question focuses on multiple teams, the long answer involves the Triple Helix of Flow. Each of the three helixes is essential for an organization to function with multiple teams successfully. We call this team-based structures, where organizations are structured around multiple teams. This effort involves individuals with the knowledge and skills to differentiate methods and techniques depending upon the type of problem they are facing. Teams can function as a self-contained autonomous unit, with leadership structures that utilize the capabilities afforded by functional leadership and boundary spanners, and with executive leaderships’ support and resources.

The short answer: the key is structuring the multiple teams around multi-team systems (MTS) and incorporating the boundary spanner’s role to manage the multi-team system and coordinate activities with organizational goals. This will require most organizations to restructure their teams using a multiteam system structure; how this is achieved varies on the organization, its products, and the number of teams involved - it’s contextual. 

Using the shared leadership model and the boundary spanner’s role will be the most significant change for most organizations. This autonomy and self-organization level is hard to achieve when some leaders insist on having control over these entities. The biggest hurdle is to get leadership and managers to the point where they trust the teams and multiteam systems to function and deliver value. Once the right training, practice, support, and resources are in place, the organizational system will operate while being adaptive and resilient to external threats (change).

Thurlow: Most organizations that attempt to use a multi-team system (MTS) implement some form of agile scaling approach. This is typically delegated to an IT department to oversee. HR is rarely ever involved until problems arise perhaps around role changes or definitions. An MTS is more than just multiple teams. It is a structural change on how we do work and how value is delivered. For Lean practitioners, this is the basic concept behind a value stream, a soup to nuts container of all the people and tools needed to deliver value.

Agile scaling approaches deal with role functions and delivery processes, almost exclusively software, but fail to address leadership changes required, although some mention them but don’t describe how to do it effectively, nor do they do not address the structural changes necessary across the organization to make the new MTS effective.

A refocus on goals is necessary, and the 2020 Scrum Guide gives a nod in this direction with the implementation of the “Product Goal” which is defined in the literature and the book as a “Distal Goal” with team goals defined as “Proximal Goals”. Goals are what determine value, but alas, many organizations lose focus above the individual team level or some vision/mission statement on a website.

The biggest challenge in scaling Scrum or Agile in most organizations is the fact that they seek to maintain their traditional management approaches, and worse still, their hierarchies. They also do not define any roles for managers and supervisors, and of course Scrum and its siblings just ignore the managers and use wise words like, “Er, um, well they could be product owners”. This is where it inevitably fails. Management hasn't changed, yet the operational workers are moving to team-based self-guiding self-managing structures (Complex Adaptive Systems), yet the “leadership” and management still cling to their legacy styles. Enter distributed leadership.

The Boundary Spanner role John mentions is crucial to changing this, and I define this as Repurposing Management. It is the answer to all those executives who don’t know what to do with all those managers when they “go Agile”!

Repurposing middle management is essential to avoiding the resistance and fightback of the machine. As Peter Drucker said, “Any innovation in a corporation will stimulate the corporate immune system to create antibodies that destroy it”. Repurposing middle management is a major key differentiator in Flow as companies battle with that challenge in change initiative. You can now offer those you coach a way to have those managers be valuable assets to solve problems and become engaged and participatory in optimizing the flow of value.

Repurposing management into new Boundary Spanner roles will change their behaviors, their focus, and their targets. This will result in emergent leadership, and a new culture of distributed leadership will start to form. The Boundary Spanner will become key to guiding and controlling multi-team systems. The core concept is that the boundary between teams needs to be filled to facilitate inter team relationships, problem solving, ensuring alignment to goals, focus on flow of value, etc.

As we scale into multiple MTS’, Boundary Spanners will work together to guide these combined MTS’. Scrum has suggested this is the role of the Scrum Master or an Agile Coach, but in reality these roles have no empowerment in most organizations, and they simply grow frustrated leaving the teams demoralized and adrift in the corporate misery. 

Part 1 of the interview on The Flow System, Getting Fast Customer Feedback and Managing Flow explores the importance of quality, getting fast feedback from customers and the concepts of flow and the flow system.

About the Book Authors 

John R. Turner, Ph.D. is an assistant professor at the University of North Texas, USA. He is the co-author of The Flow System: The Evolution of Agile and Lean Thinking in an Age of Complexity, The Flow System Guide, and The Flow System: Key Principles and Attributes. His research interests are in the interconnectivity between leadership, teams, and complexity. He can be reached at

Nigel Thurlow is the creator of the award-winning Scrum the Toyota Way training course, and the co-creator of The Flow System. He previously served as the first ever chief of Agile at a global Toyota company leading Lean and Agile practices at Toyota Connected, and is a recognized expert in the Toyota Production System, the Toyota Way, and various Agile approaches. He is a well-known keynote speaker and as a professional Scrum trainer, and has trained over 7,500 people worldwide as of 2020. His YouTube channel can be found at

Brian Rivera, callsign “Ponch,” is a recovering Navy TOPGUN and F-14 Demonstration  Team Member with extensive experience at the operational and strategic level of warfare.  He is the co-author of The Flow System: The Evolution of Agile and Lean Thinking in an Age of Complexity and The Flow System Guide. Ponch is a keynote speaker, and CEO and founder of AGLX Consulting.


Rate this Article