translate Automatic translation · View original
Systems thinking: Where is the real problem when something keeps failing in your team?
Canal De Escalado Y Liderazgo
The symptom is not the problem (even though it looks like it)
It's Monday morning and your team arrives at the Daily Scrum with the same tired faces as always. Three people report that they "almost finished" the tasks they've been "almost finishing" for a week. Someone mentions that they need access to a test environment that has been blocked since the previous sprint. The Scrum Master writes the impediment on the board, next to five others that have been there since February.
After the meeting, the Product Owner tells you: "We need to improve the team's commitment." The delivery manager suggests running a estimation workshop. HR proposes training in time management.
All of these responses start from the premise that the problem is in the team. But what if the team is simply showing a problem that exists outside of it?
Peter Senge, in his seminal work The Fifth Discipline (1990), describes this phenomenon as "shifting the burden": we focus our attention on solving symptoms because they are visible and measurable, while the structural causes remain hidden and unaddressed. The difficulty is that solving symptoms provides temporary relief, enough to validate the intervention, but it doesn't eliminate the root cause, so the problem resurfaces as soon as we stop applying the patch.
Donella Meadows, one of the most influential voices in systems thinking, put it clearly: "The behavior of a system is a consequence of its structure" (Thinking in Systems, 2008). If we want to change recurring behaviors, we need to modify the structure that generates them. But first we need to see it.
Invisible structures that generate visible dysfunctions
Systemic structures are not organizational charts or documented processes. They are patterns of incentives, implicit norms, information flows, and power relationships that shape people's behavior within a system. They are there, operating constantly, but they are rarely made explicit.
When the system rewards the opposite of what it asks for
Let's take a common example: an organization that says it values innovation but evaluates teams exclusively on meeting deadlines and fixed scope. What invisible structure is at work here?
- Formal incentive: "Be innovative, propose improvements, experiment"
- Real incentive: "Deliver exactly what was promised, with no surprises and no delays"
- Resulting behavior: Teams minimize risks, avoid experiments, and deliver the bare minimum agreed upon
Chris Argyris called this "theories-in-use" versus "espoused theories" (Organizational Learning II, 1996). What organizations say they value (their espoused theory) often contradicts what they actually reward or punish (their theory-in-use). Teams are not naive: they read the real structure, not the rhetoric.
This mismatch generates what Argyris calls "single-loop learning": the organization gets better at doing the same thing more efficiently, but never questions whether what it does makes sense. The recurring problem, in this case a lack of innovation, won't be solved with creativity workshops if the structure keeps punishing risk.
The bottleneck nobody sees (because it's in the org chart)
Another common pattern: a team systematically depends on approvals from a committee or an external area that operates with different logic and timelines. The agile team works in two-week sprints; the security committee meets once a month. The team is organized by product; architecture is organized by technology and has its own priorities.
This is not an "alignment problem" that gets solved with a meeting. It is a structural design problem: two subsystems optimized for different goals are trying to collaborate without anyone having designed the coordination mechanism.
Russell Ackoff, a pioneer of systems thinking, argued that most organizational problems are problems of interaction between parts, not problems of the parts themselves (Systems Thinking for Curious Managers, 2010). A team can function flawlessly in isolation and still experience continuous dysfunction if its structural environment is not designed for the collaboration expected of it.
Many agile scaling frameworks (SAFe, LeSS, Nexus) address precisely this: redesigning organizational structures so that agility is not confined to islands of teams surrounded by bureaucracy. But you cannot scale agility without changing structures. Teams cannot "become more agile" within structures that slow them down.
Diagnosing the system: what's the use of seeing the problem if you can't change it?
Suppose you have already identified the invisible structure. You have detected that the recurring problem is not a "lack of team commitment" but misaligned incentives, unresolved structural dependencies, or information that doesn't flow between areas. What now?
This is where systems thinking becomes uncomfortable. Because seeing the system does not automatically give you the power to change it.
Levels of influence: what you can change from where you stand
Meadows proposes a hierarchy of "leverage points" in systems: places where a small intervention can produce significant changes (Leverage Points: Places to Intervene in a System, 1999). But the paradox is that the points of greatest impact (changing paradigms, modifying the rules of the system) are usually out of reach for those who suffer the dysfunctions most.
If you are a Scrum Master for a team, you can change how you facilitate retrospectives, but you cannot change the corporate incentive structure. If you are an Agile Coach, you can make systemic dysfunctions visible, but you don't always have the mandate or the political influence to change them. If you are a transformation leader, you can change structures, but you don't always have the executive backing to do it without fierce resistance.
Gene Kim, in The Unicorn Project (2019), captures this tension masterfully: protagonists who clearly see the dysfunctional structures but lack the formal power to change them. They then resort to "subversive improvements": informal changes, coalitions, pilot experiments that demonstrate another way is possible. Sometimes it works. Sometimes it doesn't.
Signs that the problem is structural (and you need executive support)
How do you know if you are facing a systemic problem that requires intervention from higher levels? Some signs:
The problem reappears with different teams. If you rotate people and the pattern persists, it is not a problem of individual competence.
Local solutions work temporarily and then revert. A team implements improvements, but when organizational pressure increases, it goes back to its previous behaviors.
People describe the problem as "that's just how things work here." When dysfunctional behavior is normalized, the structure is sustaining it.
Any proposed change requires approvals from areas that have no incentive to facilitate it. If every improvement needs to be negotiated with silos that lose power or resources by approving it, the structure is blocking by design.
In these cases, making the system visible is an act of professional honesty. A Scrum Master who identifies a structural dysfunction and expects the team to solve it internally is shifting responsibility downward. A coach who diagnoses misaligned incentives and doesn't escalate them to those who can act on them is being complicit in the system.
Possible interventions when you can't change the structure (yet)
Recognizing that a problem is systemic doesn't always mean giving up. Sometimes it means working strategically within existing constraints while building the conditions for structural change.
Making the cost of the current system visible
Dysfunctional systems persist because their costs are hidden or spread out. If nobody measures how much time teams lose waiting for approvals, the organization doesn't perceive any urgency to change the approval process.
The State of DevOps report (DORA, 2023) shows that high-performing organizations systematically measure the costs of dependencies and waiting times. Not to punish areas, but to make it evident where the system generates unnecessary friction.
Making things visible is not complaining. It is instrumenting: measuring lead time per process phase, identifying where bottlenecks accumulate, quantifying the blocked work in progress. Data that enables conversations about redesign, not about blame.
Designing low-risk experiments that test alternatives
If the formal structure can't be changed yet, can informal structures be created that operate in parallel? Teams that test different ways of working, agreements between areas to speed up approvals in certain contexts, pilots with metrics that demonstrate results.
John Kotter, in Accelerate (2014), proposes operating with a "change network" parallel to the formal hierarchy: people from different levels who experiment with new ways of working without needing to change the whole org chart at once.
It doesn't always work. Sometimes successful pilots stay as pilots because scaling them requires dismantling power structures. But it is a more effective strategy than waiting for a transformation big bang that never arrives.
Protecting the team from the system (without pretending the system doesn't exist)
There are Scrum Masters and coaches who act as buffers: they filter irrational organizational pressures, negotiate impossible deadlines, and shield the team so it can work with some sanity. This doesn't solve the system, but it creates spaces where people can function without destroying themselves.
The risk is that this strategy becomes unsustainable. The buffer ends up burning out because it absorbs all the dysfunction. Or the team stays protected from organizational reality until the protector leaves and everything collapses.
Heidi Helfand, in Dynamic Reteaming (2020), suggests that in organizations with highly dysfunctional structures, a viable strategy is to rotate people between teams so that the system's toxicity doesn't always concentrate on the same people. It's not ideal, but it's a way to distribute the damage.
When the only possible intervention is to leave the system
This doesn't usually appear in articles about agile transformation, but it is a pragmatic truth: not all systems can be changed, and it isn't always worth trying.
If you work in an organization with incentives fundamentally aligned against collaboration, information hoarded as a source of power, and leadership with no genuine interest in changing any of this, you can spend years burning out without achieving sustainable improvements.
It is not a personal failure. It is recognizing limits. As Jerry Weinberg says in Quality Software Management (1992): "No matter how well you do your job, you cannot change a system that does not want to change."
Systems thinking is not neutrality: it implies choosing what to change
Seeing the system is a political act. Because as soon as you make a dysfunctional structure visible, someone benefits from it. The area that withholds information to maintain power. The committee that justifies its existence by generating bureaucracy. The middle management that blocks changes because it fears losing relevance.
Edgar Schein, in Humble Inquiry (2013), notes that diagnosing organizational cultures involves navigating power dynamics: who has an interest in certain things being made visible and who has an interest in them remaining hidden.
That is why systems thinking is not a sterile analytical tool. It is a diagnostic tool that forces you to ask yourself: What am I willing to name? Whom am I willing to make uncomfortable? What consequences am I willing to accept?
Many organizational dysfunctions are known to everyone and discussed by no one. Because naming them means confronting power structures. And not everyone has the political capital, job stability, or emotional energy to do so.
There is no formula for this. But there is one useful question: What would change in this organization if people could freely name the structures that generate the dysfunctions they experience?
Practical tools for diagnosing systemic structures
Although systems thinking is often presented as an abstract mindset, there are concrete techniques that help make the invisible visible.
Mapping feedback loops
A feedback loop is a causal chain where the result of an action affects the action itself. They can be reinforcing (the effect amplifies the cause) or balancing (the effect counteracts the cause).
Example of a problematic reinforcing loop:
- Pressure to deliver leads to accumulating technical debt
- Technical debt slows down future deliveries
- The slowdown increases the pressure to deliver
- Result: a downward spiral
Example of a balancing loop that blocks changes:
- A team improves its velocity through automation
- Higher velocity attracts more demand from stakeholders
- More demand saturates the team
- The team slows down to manage the demand
- Result: the improvement cancels itself out
Identifying these loops allows you to intervene at the right point. Sometimes, breaking a negative loop requires changing incentives external to the team, not improving internal practices.
The "five whys" applied at a systemic level
The five whys technique, popularized by Toyota, also works for organizational diagnosis, but it must be applied carefully.
Misuse:
- Why doesn't the team meet deadlines? Because they don't estimate well.
- Why don't they estimate well? Because they lack experience.
- Why do they lack experience? Because they are juniors.
This kind of analysis blames people further down and stops before finding the structure.
Good use:
- Why doesn't the team meet deadlines? Because unplanned tasks keep appearing.
- Why do unplanned tasks keep appearing? Because stakeholders request changes mid-sprint.
- Why do they request changes mid-sprint? Because they don't take part in planning.
- Why don't they take part in planning? Because they have executive meetings on Mondays that clash with it.
- Why isn't the agenda coordinated? Because each area works with independent calendars.
Now you've reached something structural: organizational silos that don't coordinate.
The trick is to keep asking until the answer points to a norm, an incentive, or an organizational structure, not an individual deficiency.
Organizational congruence analysis
Nadler and Tushman proposed in the 1980s a diagnostic model that remains useful: analyzing the congruence among four components of an organization:
- Work: What must be done (goals, tasks, flows)
- People: Who does it (competencies, culture, incentives)
- Formal structure: How it is organized (teams, hierarchies, processes)
- Informal structure: How it actually works (relationships, communication, power)
Dysfunctions usually lie in the incongruences between these elements. For example:
- The work requires multidisciplinary collaboration, but the formal structure is organized by specialty.
- People have individual incentives, but the work requires collective results.
- The informal structure (who consults whom) contradicts the formal structure (who reports to whom).
Identifying these incongruences helps diagnose why certain problems persist: because the system is fundamentally misaligned.
Communication architecture: Conway's law
Melvin Conway formulated in 1968 an observation that still holds: "Organizations design systems that mirror their communication structure." If your company has separate frontend, backend, database, and QA teams, you will probably end up with a layered system that reflects that separation.
Matthew Skelton and Manuel Pais explore this further in Team Topologies (2019): team structures determine what kind of software is possible to build. If you want decoupled architectures, you need autonomous teams. If you maintain structural dependencies between teams, you will have technical dependencies.
When a team has recurring problems with integration, coordination, or technical dependencies, it's worth asking: is it a technical problem or the inevitable reflection of how the company is organized?
Sometimes the solution is not to improve engineering practices. It is to redesign teams.
How to bring this into real conversations (without getting fired)
Identifying dysfunctional structures is one thing. Generating conversations that lead to changing them is another.
Framing findings as opportunities, not accusations
Nobody wants to hear "the problem is that you organized this badly." It works better to say: "We've noticed a pattern: when X happens, it always results in Y. What would happen if we tried Z?"
The difference is not just semantic. It is strategic. You present data, identify a pattern, and propose an experiment. You don't point fingers.
Using metrics that make the systemic cost evident
As we mentioned earlier, measuring waiting times, blockers, and rework allows you to make the conversation objective. It is no longer "I think we have a problem." It is "40% of the team's time is spent waiting for approvals."
Mik Kersten's Flow Framework (Project to Product, 2018) proposes measuring four key metrics:
- Flow Velocity: How much value is delivered
- Flow Time: How long it takes to deliver it
- Flow Efficiency: What proportion of time is active work versus waiting
- Flow Load: How much work is in progress
These metrics make visible the impact of structures that generate blockers, dependencies, or excessive multitasking.
Looking for allies at different levels
Systemic change is rarely driven by a single person. It requires coalitions: people from different areas who see the same problem and have incentives to address it.
A Scrum Master can only make things visible. A Product Owner can prioritize solutions. An engineering leader can authorize changes. An executive can modify incentives. You need several of these actors aligned to change structures.
This is what Kotter calls "genuine urgency": not panic, but convergence of key actors who recognize that the status quo has intolerable costs.
Knowing when to stop
Not every battle can be won. Sometimes you have made the structure visible, proposed alternatives, gained allies, and the organization still decides not to change. Because the political cost is too high, because there are conflicting priorities, because it simply isn't the right time.
In those cases, persisting can become a form of professional self-destruction. We are not saying you should give up. We are saying you should consciously choose where to invest your energy, knowing that not every system wants to be transformed.
Seeing the system doesn't make you solely responsible for fixing it
One of the dangers of systems thinking is assuming that, because you now see the whole structure, you should be able to solve it. But that would fall into another trap: the fantasy of the superhero who fixes organizations.
Complex systems are not changed by a single person. They are changed by coalitions, external pressure, crises that make the status quo unsustainable, and leadership that decides to bet on something different.
Your job, whether you are a Scrum Master, Agile Coach, Product Owner, or team lead, is not to fix the entire organization. It is to:
- Make visible what is generating recurring dysfunctions
- Protect your team from the most toxic parts of the system while you work on changing them
- Experiment with alternatives in the spaces where you have influence
- Build alliances with those who have the power to change structures
- Know when to escalate a problem and when to recognize that it is beyond your reach
And perhaps most importantly: know when the system will not change and you need to leave it before it destroys you.
Because systems thinking also includes recognizing that you are part of the system. And that taking care of yourself is, at times, the only strategic intervention possible.
References
Ackoff, R. L. (2010). Systems Thinking for Curious Managers. Triarchy Press. ISBN: 978-0956263865
Argyris, C. (1996). Organizational Learning II: Theory, Method, and Practice. Addison-Wesley. ISBN: 978-0201629831
DORA (2023). Accelerate State of DevOps Report 2023. Google Cloud. https://cloud.google.com/devops/state-of-devops
Helfand, H. (2020). Dynamic Reteaming: The Art and Wisdom of Changing Teams. O'Reilly Media. ISBN: 978-1492061298
Kersten, M. (2018). Project to Product: How to Survive and Thrive in the Age of Digital Disruption with the Flow Framework. IT Revolution Press. ISBN: 978-1942788393
Kim, G. (2019). The Unicorn Project: A Novel about Developers, Digital Disruption, and Thriving in the Age of Data. IT Revolution Press. ISBN: 978-1942788768
Kotter, J. P. (2014). Accelerate: Building Strategic Agility for a Faster-Moving World. Harvard Business Review Press. ISBN: 978-1625271747
Meadows, D. H. (2008). Thinking in Systems: A Primer. Chelsea Green Publishing. ISBN: 978-1603580557
Meadows, D. H. (1999). Leverage Points: Places to Intervene in a System. The Sustainability Institute. http://donellameadows.org/archives/leverage-points-places-to-intervene-in-a-system/
Schein, E. H. (2013). Humble Inquiry: The Gentle Art of Asking Instead of Telling. Berrett-Koehler Publishers. ISBN: 978-1609949815
Senge, P. M. (1990). The Fifth Discipline: The Art & Practice of The Learning Organization. Doubleday. ISBN: 978-0385517256
Skelton, M., & Pais, M. (2019). Team Topologies: Organizing Business and Technology Teams for Fast Flow. IT Revolution Press. ISBN: 978-1942788812
Weinberg, G. M. (1992). Quality Software Management, Vol. 1: Systems Thinking. Dorset House Publishing. ISBN: 978-0932633224
Agile Club: Do you want to go deeper into organizational diagnosis and systems thinking applied to agile transformations? Join Agile Club, where we share real cases, analysis tools, and conversations about what to do when you identify structural problems. Access here.
Take part in the poll
This article includes an active poll in the Scrum Manager community.
poll View the poll in the community