Systems Thinking Evaluation Skill
Apply a rigorous systems lens to evaluate what's working, what isn't, and why in any system — technical, organizational, product, or process.
When to Use This Skill
Trigger this skill whenever the user asks you to:
- Evaluate or assess a system, product, process, strategy, or architecture
- Understand why something isn't working as expected
- Identify bottlenecks, failure points, or risks
- Suggest where to intervene for the most impact
- Review a design or plan before execution
Core Mental Models to Apply
1. Identify the System Boundary
- What's inside the system (in scope)?
- What's outside (environment, dependencies, constraints)?
- Where does the system begin and end?
2. Stocks and Flows
- Stocks: What accumulates over time? (users, debt, trust, knowledge, bugs, revenue)
- Flows: What increases or decreases those stocks? (acquisition, churn, learning, entropy)
- Where are flows blocked, broken, or leaking?
3. Feedback Loops
- Reinforcing loops (R): Self-amplifying dynamics — virtuous cycles or vicious spirals
- Example: More users → more content → more users (growth flywheel) - Example: More bugs → less trust → fewer contributors → more bugs
- Balancing loops (B): Self-correcting dynamics — goal-seeking behaviors
- Example: High load → auto-scale → stable performance - Example: User complaints → support → resolution → satisfaction
- Ask: Which loops dominate the system's current behavior?
4. Delays
- Where are there time lags between cause and effect?
- Delays often cause oscillation, overcorrection, or invisible failures
- Example: Hiring takes 3 months → team overloads → burnout → more attrition
5. System Archetypes (common failure patterns)
Match observed behavior to known archetypes:
| Archetype | Pattern | Signal |
|---|---|---|
| Limits to Growth | Growth hits a constraint and stalls | Plateau despite investment |
| Fixes that Fail | Quick fix creates new problems | Recurring issues after "solutions" |
| Shifting the Burden | Symptomatic fixes erode fundamental ones | Team always firefighting |
| Tragedy of the Commons | Shared resources are depleted | Quality/performance degrades over time |
| Escalation | Competing actors amplify each other | Bidding wars, arms races |
| Drifting Goals | Performance gap closed by lowering standards | "Good enough" keeps declining |
| Accidental Adversaries | Well-meaning actors undermine each other | Misaligned incentives between teams |
6. Leverage Points
Rank interventions by impact (from lowest to highest leverage):
- Numbers (parameters, budgets, quotas) — low leverage
- Buffer sizes and stock capacities
- Flow rates and delays
- Feedback loop strength
- Information flows (who has access to what, when)
- Rules and incentives
- Goals of the system
- Power to change the system's structure
- Mindsets and paradigms — highest leverage
Evaluation Output Format
When evaluating a system, structure the response as follows:
🟢 Where the System Works
- Identify functioning feedback loops, healthy stocks, aligned incentives
- Call out genuine strengths (not to be polite — to understand what to protect)
🔴 Where the System Breaks Down
- Point to broken loops, leaking flows, missing feedback, or misaligned incentives
- For each issue, name the archetype if one applies
- Identify delays that hide the problem
⚠️ Key Risks and Failure Modes
- What could cause the system to tip into a bad equilibrium?
- What reinforcing loop could go negative?
- What constraint will be hit next?
🎯 High-Leverage Interventions
- Ranked list of where to intervene
- For each: what changes, what loop or flow it affects, expected result
- Flag quick fixes that might backfire (Fixes that Fail archetype)
📊 System Diagram (optional, when helpful)
Describe or sketch a causal loop diagram in text:
[Variable A] → (+) [Variable B] → (+) [Variable A] ← Reinforcing loop R1
[Variable A] → (+) [Variable C] → (-) [Variable A] ← Balancing loop B1Tone and Approach
- Be direct about what's broken — systems evaluation is not diplomacy
- Use concrete examples tied to the user's specific context
- Prioritize systemic causes over symptoms — don't just describe what's wrong, explain *why the system produces that outcome*
- When a problem "keeps coming back," suspect a reinforcing loop or Shifting the Burden archetype
- Always suggest at least one high-leverage intervention — not just diagnosis
Example Application Triggers
- "Evaluate our onboarding funnel" → apply stocks/flows to conversion, identify where users leak out and why
- "Why does our team keep missing deadlines?" → look for delays, Shifting the Burden, workload dynamics
- "Is this architecture scalable?" → identify capacity limits, balancing loops under load, missing circuit breakers
- "Assess our growth strategy" → find reinforcing flywheels, limits to growth constraints, escalation risks
- "What's wrong with our deploy process?" → trace flow from commit to production, find delays, balancing loops
- "Should we change our pricing model?" → map revenue stocks, customer feedback loops, competitive dynamics