Blog / Leadership / Systems Thinking: Seeing the Pattern Before You Ac

Systems Thinking: Seeing the Pattern Before You Act

Most smart mistakes come from applying the wrong kind of thinking to the wrong kind of system. A four-type model — Clear, Complicated, Complex, Chaotic — plus the DART diagnostic for telling them apart before you act.

Shreedhar Kodate · 03 September 2026 · 8 min read
leadershipsystems-thinkingdecision-making

Source video: How To Think SO Clearly People Assume You're Brilliant — Sandeep Swadia (YouTube: theMITmonk) Credit: All core framework ideas below (the four system types, the "DART" diagnostic, and the platform-perspective idea) are drawn directly from this video. Section 6 adds outside context to connect these ideas to established literature.

1. Summary

  • A system is a set of connected parts that keeps producing a repeating pattern (a company, a career, a marriage, a coffee shop).
  • Most "smart mistakes" happen because someone applied the wrong kind of thinking to the wrong kind of system.
  • There are four types of systems — Clear, Complicated, Complex, and Chaotic — and each demands a different response.
  • Three things make systems thinking hard: (1) not knowing which system you're in, (2) the incentive/cobra effect (people game misaligned rewards), and (3) delayed feedback loops (the consequence shows up long after the action).
  • A four-step diagnostic — DART (Deconstruct, Analyze, Recognize, Test) — helps you figure out which system you're actually in before you act.
  • Every system you live inside is also quietly training you, and you usually can't see the direction from inside it — you need mentors, data, or time to get an outside ("platform") view.
  • Many "either/or" choices (Ferrari vs. Toyota, luxury vs. mass-market) are limits of system design, not limits of reality — Apple's mass-produced luxury iPhone is the counterexample.

2. Thinking Framework

The Four Types of Systems

System Cause → Effect Relationship Example from video
Clear Obvious, directly observable Van Halen's "no brown M&Ms" contract clause; a recipe; surgical scrub-in protocol
Complicated Exists but hidden — needs analysis/expertise to uncover ER patient with chest pain; choosing a mortgage
Complex Only visible in hindsight; emerges over time Post-acquisition culture integration; raising a teenager
Chaotic Broken / impossible to know in the moment 1982 Tylenol cyanide poisonings

DART — The Diagnostic Framework

  • D — Deconstruct: Break the problem into sub-parts. Are they stable or shifting?
  • A — Analyze: What's the cause-effect relationship? Obvious → Clear. Discoverable via analysis → Complicated. Emergent, seen only in hindsight → Complex. Broken → Chaotic.
  • R — Recognize: Have you seen this pattern before, in this system or a different one?
  • T — Test: Run the smallest possible test before committing (skip this step only in a chaotic system — there's no time to test).

The Platform Perspective

You cannot see which direction a system is taking you from inside it (like not knowing if your train or the one next to it is moving). Three ways to get outside perspective:

  1. Mentors — someone outside your story who can see your "train" from the platform.
  2. Data — numbers don't care about your narrative.
  3. Time — compare yourself to a year/month/week ago.

3. Act / Applying the Framework

  1. Ask the three diagnostic questions anywhere — at a coffee shop, in a meeting, in your career: What are the hidden parts? How are they connected? What patterns keep repeating?
  2. Run DART before you act on any non-trivial problem. Don't skip straight to a solution — deconstruct and analyze first to identify the system type.
  3. Match your response to the system type:
  • Clear → build/follow a checklist. Don't improvise.
  • Complicated → slow down, bring in the right specialist (not just any expert).
  • Complex → run small experiments, stay directionally right, course-correct — don't expect a fixed playbook.
  • Chaotic → stabilize first, act immediately, analyze later. Avoid analysis paralysis.
  1. Watch for the cobra effect in anything you design — before setting an incentive or metric, ask "how could someone technically satisfy this without achieving the actual goal?"
  2. Build in a feedback-loop check — for slow-moving risks (health, culture, financial habits), don't wait for the damage to be visible; look for early leading indicators.
  3. Schedule regular "platform time" — a recurring check-in with a mentor, a data review, or a look-back at where you were 6–12 months ago.
  4. Question your own binary constraints — when you're told "you can only have A or B," ask whether that's a fact about the world or just a limitation of the current system design.

4. Details of Topics Discussed in the Video

  • The firefighter analogy: experienced firefighters read smoke color as information (black = fuel/plastics/chemicals; white/gray = moisture or oxygen-starved fire) — this is pattern recognition from observable parts, the essence of systems thinking.
  • The coffee shop example: customer → cashier → barista → customer is a simple, observable system used to illustrate "parts → connections → patterns."
  • The cobra effect: British colonial bounty-per-dead-cobra policy in Delhi backfired because people bred cobras for profit — a textbook incentive-misalignment failure.
  • Delayed feedback loops: 20th-century cigarette culture — pleasure arrived in seconds, damage arrived in decades, which is why the system was hard to correct in real time.
  • Clear system example — Van Halen's M&M clause: a trivial contract detail (no brown M&Ms backstage) functioned as a proxy signal for whether a venue had read the entire safety contract carefully.
  • Complicated system example — the ER "chest pain" scenario: one symptom, ~20 possible causes; requires expert differential diagnosis, not a checklist.
  • Complex system example — a real acquisition: the speaker (as COO) led an acquisition where the products fit well on paper, but cultural incompatibility (formal/hierarchical vs. informal/fast-moving) caused leadership attrition and product shutdowns within 90 days — cause and effect were only clear in hindsight.
  • Chaotic system example — the 1982 Tylenol poisonings: Johnson & Johnson couldn't analyze first; they pulled 31 million bottles immediately (stabilize first, understand later).
  • The "binary choice is a system-design limit" idea: Apple manufacturing ~350 iPhones/minute — a luxury product at mass-market scale — shows that "high margin OR high volume" was never a law of physics, just an unsolved system-design problem until Apple solved it.
  • Personal narrative: the speaker's own path from a "running away from home to become a monk" story (which he later realized, via honest feedback, was really running away from his father) to recognizing that self-narrative is itself a redesignable system.

5. Diagrams

The Four System Types and Their Response

flowchart TD classDef node fill:none,stroke:#2b6cb0,stroke-width:1.5px,color:#2b6cb0; A[Problem / Situation] --> B{Cause & Effect<br/>relationship?} B -->|Obvious, observable| C[Clear System] B -->|Exists but hidden| D[Complicated System] B -->|Only visible in hindsight| E[Complex System] B -->|Broken / unknowable| F[Chaotic System] C --> C1[Follow a checklist<br/>Do not improvise] D --> D1[Slow down<br/>Bring in the right expert] E --> E1[Run small experiments<br/>Stay directionally right] F --> F1[Stabilize first<br/>Act, then analyze] class A,B,C,D,E,F,C1,D1,E1,F1 node;

The DART Diagnostic Loop

flowchart LR classDef node fill:none,stroke:#2b6cb0,stroke-width:1.5px,color:#2b6cb0; D1[D — Deconstruct<br/>Break into sub-parts] --> A1[A — Analyze<br/>What links cause & effect?] A1 --> R1[R — Recognize<br/>Seen this pattern before?] R1 --> T1[T — Test<br/>Run the smallest test] T1 -->|Chaotic: skip test| ACT[Act / Decide] T1 --> ACT class D1,A1,R1,T1,ACT node;

Getting the Platform Perspective

flowchart TD classDef node fill:none,stroke:#2b6cb0,stroke-width:1.5px,color:#2b6cb0; Inside[Inside the System<br/>Cannot see direction of travel] --> M[Mentors<br/>Outside view, no stake in your story] Inside --> Da[Data<br/>Numbers vs. your narrative] Inside --> T[Time<br/>Compare to past self] M --> Out[Platform Perspective<br/>See which way the train is moving] Da --> Out T --> Out class Inside,M,Da,T,Out node;

6. Learnings from Additional Sources (Connecting the Concepts)

  • This maps closely to the Cynefin Framework (Dave Snowden, IBM/Cognitive Edge, formalized ~1999–2007), a well-known sense-making model in organizational and management theory that uses nearly identical categories: Clear (formerly "Simple/Obvious"), Complicated, Complex, and Chaotic, plus a fifth "Disorder" state for when you don't know which domain you're in — which is exactly the gap DART is designed to fill. Cynefin is widely used in agile software development, public policy, and crisis management.
  • The cobra effect is a specific case of Goodhart's Law: "When a measure becomes a target, it ceases to be a good measure" (economist Charles Goodhart, 1975). Both describe the same failure mode — optimizing for a proxy metric instead of the underlying goal.
  • **Donella Meadows' *Thinking in Systems: A Primer*** (2008, based on her decades of systems-dynamics research at MIT) is the classic text on this topic — it formalizes concepts like stocks, flows, reinforcing/balancing feedback loops, and delayed feedback (directly relevant to the cigarette example), and shows how leverage points differ by system type.
  • The chaotic-system response ("stabilize first, analyze later") echoes the OODA loop (Observe–Orient–Decide–Act), developed by military strategist John Boyd, which similarly prioritizes fast action and iteration over complete information in fast-moving, high-uncertainty situations.
  • Complex-system advice ("run small experiments, stay directionally right") parallels agile/lean methodology and the "Cynefin complex-domain" guidance to probe–sense–respond rather than plan-and-execute — small, safe-to-fail experiments generate the hindsight needed to navigate emergent systems.
  • The "platform perspective" idea overlaps with the concept of the "outside view" popularized by psychologist Daniel Kahneman and used in forecasting — using external data/reference cases instead of your own internal narrative to judge a situation.

7. References

  1. Sandeep Swadia (theMITmonk), "How To Think SO Clearly People Assume You're Brilliant", YouTube.
  2. Kurtz, C. F., & Snowden, D. J. (2003). "The new dynamics of strategy: Sense-making in a complex and complicated world." IBM Systems Journal — origin of the Cynefin Framework.
  3. Meadows, D. H. (2008). Thinking in Systems: A Primer. Chelsea Green Publishing.
  4. Goodhart, C. (1975). Original statement of what became known as "Goodhart's Law."
  5. Wikipedia — "Cobra effect" (background on the historical Delhi bounty policy referenced in the video).
  6. Wikipedia — "1982 Chicago Tylenol murders" (background on the chaotic-system example).
  7. Boyd, J. — originator of the OODA Loop concept, U.S. Air Force strategist.