The Mental Model Every Engineer Is Missing
You were trained to fix components. But the failures that cost careers, companies, and customers are never in the components — they're in the space between them. Systems thinking is the shift from seeing symptoms to seeing generators.
"94% of problems in any organization are systems problems, not people problems." — W. Edwards Deming, Architect of Japan's post-war industrial recovery.
What a System Actually Is
A pile of engine components on a garage floor is not a car. The components are all there, but the car only exists when those components are interconnected toward a specific purpose.
- 🧩 Elements: The building blocks you can see, count, or measure (Chefs, tables, code, servers).
- 🔗 Interconnections: The relationships and flows between elements (Information, data, influence).
- 🎯 Purpose: The system's WHY. The least visible but most powerful driver of behavior.
Two Flavors: Closed vs. Open
Understanding which kind of system you are building is the difference between robustness and fragility.
The System Boundary
- Nature: Self-contained, predictable, isolated.
- Analogy: A Docker container with mocked dependencies.
- Trap: Most engineers design for this but deploy to the other.
- Nature: Dynamic, exchanges energy/matter with environment.
- Analogy: A microservice in production with real traffic and APIs.
- Reality: The real world pours through the boundary constantly.
Case Study: The Gallon Fashion Disaster
A decision that looked brilliant on a P&L sheet but destroyed a brand because the model boundary was drawn too narrowly.
Boundary Analysis
- Observation: Production costs are too high.
- Action: Move production overseas.
- Result: Short-term profit, long-term brand destruction via labor scandals.
- Observation: Production costs are too high.
- Action: Map the ecosystem (Contractors, Communities, Activists, Governments).
- Result: Anticipated cascades, protected brand, sustainable growth.
Read Any System in 10 Minutes
The next time you're in a meeting, use this 4-step process to run a mental simulation of any proposed change.
The 10-Minute System Scan
List every actor, component, and resource inside or adjacent to the decision.
What does each element send and receive? Trace money, signals, data, and influence.
What is this system optimizing for based on behavior, not mission statements?
Apply the change. Which element is most stressed? Which feedback loop triggers?
Mastery Principles
Don't read the architecture docs first. Watch what the system actually produces. Behavior reveals the real purpose.
Every time reality surprises you, it crossed a boundary you drew too narrowly. Expand your model.
The gap between action and consequence is where systems oscillate. Identify the lag in your feedback loops.
Stop asking "Who caused this?" Start asking "What part of the system keeps producing this?" Fix the generator, and the symptom stays fixed.