Systems Thinking

Systems thinking begins by resisting the urge to solve the problem too quickly.

What appears to be a simple failure may be the visible result of relationships, incentives, accumulated decisions, delayed information, and different people acting reasonably from different positions. More effort applied to the visible event may produce temporary movement without changing what keeps producing it.

The practice is to widen the inquiry before narrowing the intervention.

Start with the boundary

Every explanation draws a boundary. It decides what belongs to the situation and what can be treated as background.

If a team repeatedly misses deadlines, a narrow boundary might include the people assigned to the work and their daily execution. A wider boundary might include how commitments are made, how priorities change, when dependencies become visible, who can make trade-offs, and whether the date was credible in the first place.

Neither boundary is automatically correct. Both are choices. Each one makes some relationships visible and leaves others outside the map.

I use questions such as Part/Not Part and Thing/Not Thing to make those choices explicit. What exactly are we examining? What are we treating as separate that may belong to the same system? What have we excluded, and what changes if we bring it back into view?

A boundary is necessary. It is not neutral, and it is rarely final.

Look beneath the event

The Iceberg model helps move the inquiry through several depths: the event we can see, patterns across time, structures that may be producing those patterns, and the mental models through which people interpret the situation.

This does not mean every problem has one hidden root cause. In a complex situation, several conditions may reinforce one another. A useful question is not only “Why did this happen?” but also “What keeps making this outcome possible?”

A customer complaint is an event. Similar complaints across several months form a pattern. Product incentives, hand-offs, information gaps, or approval rules may be part of the structure. Beliefs about speed, quality, authority, or what customers should tolerate may shape how that structure persists.

Moving down the Iceberg does not make the visible event unimportant. It helps us respond without mistaking the latest incident for the entire system.

Listen before drawing

A system map made from one desk usually becomes a map of one person’s assumptions.

Stakeholder Interviews, Empathy Walks, Dialogue Walks, and the 4 Levels of Listening help me encounter the system from positions other than my own. The purpose is not to collect polite agreement. It is to notice where accounts differ, where one group carries a cost another group cannot see, and where the same event acquires different meaning from another point in the system.

This is also where perspective work matters. DSRP, developed by Derek and Laura Cabrera, offers four connected ways of examining thought: Distinctions, Systems, Relationships, and Perspectives. In practice, I use them as questions:

• What is this, and what is it not?
• What whole is it part of, and what parts does it contain?
• What affects it, and what does it affect?
• From which point is this view being formed?

I also use point/view perspective prompts to make the observer visible inside the inquiry. What is the point from which this view is being formed? What becomes visible from there, and what disappears? The prompt is useful because a perspective is not merely another opinion. It changes which parts, relationships, and consequences can be seen.

Changing perspective does not require pretending that every account is equally accurate. It asks us to understand what each account can reveal, then return to evidence.

Observe behaviour over time

Interconnectedness is easy to declare and difficult to examine. Feedback loops make the claim more specific.

A reinforcing loop amplifies movement. Success may attract resources that create more success. Repeated delay may reduce trust, which produces more checking, which adds further delay. A balancing loop works against movement, as when capacity limits constrain growth or a control responds to deviation.

The loop that matters most can also change with time. A practice that supports early growth may create fragility at a larger scale. A useful buffer in an uncertain period may become unnecessary cost when conditions stabilize. Systems thinking asks us to observe these shifts instead of converting yesterday’s intervention into permanent doctrine.

Stocks and flows add another distinction. Trust, cash, skill, attention, backlog, and technical debt can accumulate or decline. The event that attracts attention is often a change in flow, but what people experience is the accumulated condition.

Delays complicate the picture. The effect of a decision may arrive after the person responsible has concluded that nothing happened. Several delayed responses can create oscillation, overcorrection, or blame even when no participant intends the outcome.

These ideas are central to Donella Meadows’ treatment of systems, feedback, stocks, flows, delays, rules, information, goals, and leverage points. Her caution matters as much as the framework: leverage points are not a formula for controlling a complex system. They are an invitation to examine where intervention might alter its behaviour.

Treat the map as a hypothesis

A map is not the system. It is a claim about the system.

The arrows on a causal map may come from experience, interviews, records, observation, or inference. They should not all receive the same confidence merely because they appear in the same diagram. A relationship noticed by several stakeholders may still need evidence. Correlation over time does not by itself establish causality. A clean visual can conceal a weak assumption.

For me, systems mapping is useful when it exposes these assumptions. It gives people something specific to question. Which relationship is missing? Which arrow points in the wrong direction? Where is a delay being ignored? What evidence would make us revise the map?

Synthesis comes after this examination. It is not the production of a larger diagram. It is the work of forming a sufficiently coherent view from multiple observations, perspectives, and forms of evidence so that a decision can be made.

Intervene and learn

Systems thinking does not remove uncertainty. It should improve the quality of the experiment we choose under uncertainty.

A useful intervention names the boundary, the expected effect, the evidence we will observe, and the consequences we may create elsewhere. It is small enough to learn from when possible, but not so small that it avoids the structural issue.

This is where I prefer to test an idea by becoming the guinea pig myself. The test does not prove a universal law. It produces contact with reality. What changed? What resisted? Which consequence arrived late? Who experienced an effect that the original map did not include?

The model must remain revisable. Systems can adapt to the intervention, people can change their behaviour when a measure becomes a target, and an improvement in one part can move the cost somewhere else.

The value of systems thinking is not that it allows us to see everything. It reminds us that we do not.

Before acting, widen the inquiry. After acting, watch what the system does. Then revise the map.

How I use these practices

The practices in this work come from more than one systems-thinking, inquiry, and listening tradition. I use them together as a working sequence:

1. Observe the event and its behaviour over time.
2. Listen to the people experiencing different parts of the situation.
3. Make boundaries, distinctions, parts, relationships, and perspectives explicit.
4. Map feedback, accumulation, delay, and possible causal relationships.
5. Form a synthesis sufficient for a decision.
6. Make a bounded intervention and observe what follows.
7. Revise the map when reality contradicts it.

This sequence is my synthesis of the practices, not a claim that they constitute one universally validated method. Empathy Walks, Stakeholder Interviews, Dialogue Walks, the 4 Levels of Listening, the Iceberg model, DSRP, systems mapping, and causal inquiry do different jobs. Combining them does not erase their different origins or limitations.

The map remains a working interpretation. It should make assumptions easier to inspect, not make them look authoritative.

Sources and lineage

• Derek and Laura Cabrera’s DSRP work supplies the Distinctions, Systems, Relationships, and Perspectives structure used here: https://www.cabreraresearch.org/
• Daniel H. Kim’s Introduction to Systems Thinking is the source used here for the Iceberg presentation of events, patterns, systemic structures, and mental models: https://thesystemsthinker.com/wp-content/uploads/2016/03/Introduction-to-Systems-Thinking-IMS013Epk.pdf
• Donella H. Meadows’ Thinking in Systems and “Leverage Points: Places to Intervene in a System” inform the treatment of stocks, flows, feedback loops, delays, information, rules, goals, and intervention: https://donellameadows.org/archives/leverage-points-places-to-intervene-in-a-system/
• Otto Scharmer and the Presencing Institute provide the 4 Levels of Listening practice referenced here: https://www.presencing.org/offerings/apply-levels-listening