Learning guide and study notes
How do the parts interact? Use it when a local fix creates new problems.
Systems thinking examines how interconnected parts, reinforcing and balancing feedback loops, and time delays produce emergent behavior across a whole environment—especially when local fixes backfire.
A system's behavior can emerge from relationships, feedback and delays, not just one faulty component.
Work through these steps in order, applying the tools above. Write down the answer to each step before moving on.
Plot what changes over time and define which actors and variables are in scope. [1] Source: The Open University — Causal loop diagrams (opens in a new tab)
Connect variables and mark reinforcing, balancing and delayed effects; tell the causal story aloud. [1] Source: The Open University — Causal loop diagrams (opens in a new tab)
Test a leverage point and watch for side effects in other parts of the system. [2] Source: UK Government — Systems thinking toolkit (opens in a new tab)
Draw one reinforcing and one balancing loop behind a recurring service problem; mark one delay and one measurable intervention.
Situation: A service slows during traffic spikes.
How to apply: Trace latency → retries → more load → higher latency; also check downstream capacity.
Outcome: Test backoff and rate limits alongside capacity changes, then watch the whole loop.
Situation: After a failed release, leadership adds three manual approval gates to reduce deployment risk.
How to apply: Map the delayed loop: heavier gates → fewer, larger batched releases → harder diagnosis → more rollbacks.
Outcome: Replace manual batch gates with automated pre-flight checks and smaller, reversible releases.
Situation: On-call engineers raise queue-depth alert thresholds to stop noisy pages during busy hours.
How to apply: Trace how delayed visibility allows worker saturation to cascade into upstream connection timeouts.
Outcome: Fix worker throughput and backpressure signals instead of hiding the queue buildup.
Illustrative examples showing how the lens changes decisions; not measured outcomes.
Watch out: A diagram is a hypothesis, not proof that every arrow is causal.
Combine this lens with others rather than relying on it alone. Browse all thinkings
Repeated behavior often comes from relationships, delays, and feedback rather than one isolated component.
Use a problem of your own. Write a short answer to each prompt, then return after trying the method.
Answer from memory before opening the suggested answer.
Demand may rise in response, or the bottleneck may move downstream. Examine inflow, processing, feedback, and delay before assuming capacity is the only constraint.
It states directional relationships, reinforcing or balancing effects, and delays that can be checked against observed behavior. It is a hypothesis, not proof.
These sources support the method; the examples and exercises are original teaching illustrations.