Think In Systems: The Skill Most People Never Learn (Full Course)

Main ideas / lessons (Systems Thinking course)

• Systems thinking explains recurring problems as products of underlying structures, not as isolated personal failings.
• Core claim:
  • A problem that keeps returning is being produced by a structure.
• Most structures can be understood through four recurring elements:
  1. A “tank” that drains
  2. A reinforcing “loop” that amplifies/drains
  3. A “delay” that hides the connection
  4. A “lever” you keep pulling that isn’t the strongest one
• Goal of the course: use a progression of ~20 tools to place any repeating problem on one page and identify the correct leverage point.

---

Course workflow & method (taught repeatedly)

Preparation concept

• Before starting, keep in mind three repeating problems (the “class material”) and apply tools to them.

---

Part 1 — Why problems keep coming back (seeing the system)

Tool 1: The System Test (identify what you’re really looking at)

• Purpose: distinguish a heap (one-off accumulation) from a system (interacting parts producing reliable output).
• Run it (3 steps):
  1. Pick one life area → list its parts.
  2. Draw connections (what feeds what).
  3. Name the actual output it reliably produces (the system’s real purpose).
• Key lesson: if output is something you hate, blame the structure’s designed output, not a single component.

Tool 2: The Iceberg (go from events → patterns → structures)

• Iceberg layers:
  • Event: visible occurrence (“skipped gym,” “bought too much”)
  • Pattern: repetition of events (“happens every deadline month”)
  • Structure: incentives/information/habits that generate the pattern (what makes it scheduled and inevitable)
• Run it (3 steps):
  1. Write one recent frustrating event in one sentence.
  2. Ask when it has happened before; keep asking until you describe repetition.
  3. Ask what arrangement would generate the same pattern on schedule; write that sentence.
• Key lesson: as you go deeper (event → pattern → structure), leverage increases and blame shifts:
  • Event → “character”
  • Structure → “design” (changeable with calm analysis)

---

Part 2 — Learning to see structure (diagramming it)

Tool 3: Variables and Links (causal loop diagram basics)

• Variables: nouns that can rise/fall (e.g., stress, savings, backlog).
  • Rule: use nouns only, not verbs.
• Links: arrows between variables with a sign:
  • + means same direction effect
  • - means opposite direction effect
• Run it (3 steps):
  1. Pick one recurring problem → write 3–5 noun variables.
  2. Draw arrows for direct influences; label each plus/minus.
  3. Read every arrow as a sentence (if it sounds wrong, the link is wrong).
• Key habit: the diagram should represent what happens, not what “should” happen.

Tool 4: Reinforcing Loop (R)

• Definition: a closed circle where arrows reinforce each other; each pass amplifies outcomes.
• Key lesson: unfairness in life outcomes often comes from small early differences that snowball.
• Run it (3 steps):
  1. Find a circle where “more → more” and diagram it.
  2. Label it R and note direction (growth or decay).
  3. Make one small deliberate push in the desired direction.
• Key warning: reinforcing loops amplify whatever they’re given—feed what you want, starve what you don’t.

Tool 5: Balancing Loop (B) and Set Point

• Definition: loops that defend a target level (like a thermostat).
• Examples:
  • Diet snapback (weight returns)
  • Money spending drifting back after crackdowns
• Willpower can “win” briefly, but balancing loops keep correcting toward normal.
• Strategy: move the set point by changing what counts as “normal” in your environment.
• Run it (3 steps):
  1. Spot snapback: results rubber-band to the same level.
  2. Find the set point being defended.
  3. Change surroundings that define normal → make the new behavior “path of least resistance.”

Tool 6: Loop Dominance (who’s winning right now?)

• Real life has multiple loops, and they trade dominance over time.
• Example: honeymoon growth phase (reinforcing dominates) then plateau phase (balancing dominates).
• Run it (3 steps):
  1. Draw both loops on one page (R growing it, B slowing it).
  2. Add delay marks (time gaps between cause and effect).
  3. Identify which loop is winning this month; feed the engine during R dominance, tend constraints during B dominance.
• Key lesson: people often burn out by pushing the wrong lever during the handoff.

---

Part 3 — The plumbing (what sets levels)

Tool 7: Stock and Flow (bathtub model)

• Stock: a measurable “amount” (level in a tank) at a moment in time.
• Flows: inflow/outflow rates that change stocks.
• Valve detail: on pipes sits a valve controlling the rate.
• “Clouds” on the diagram indicate where tracing stops (blind spots).
• Run it (3 steps):
  1. Pick the key stock you’re low on and draw it.
  2. Draw named taps (inflows) and drains (outflows).
  3. Circle the valve you’ve never consciously watched.
• Key consequences:
  • Levels change only as fast as flows allow (change is slow).
  • When inflow = outflow, the stock holds steady.

Tool 8: Valve Rule (you can’t act on stocks)

• Core rule: you cannot directly order a level; you act on flows/valves.
• Run it (3 steps):
  1. Find “stock commands” in your plans (goals phrased as a state you’ll possess).
  2. List the flows that feed/drain the stock.
  3. Rewrite the goal as a valve setting (a rate with a number and day).
• Key lesson: speeches and wishes don’t move levels; rate changes do.

Tool 9: Buffer (why small shocks become crises—or not)

• Buffer: a stock deliberately kept above daily need.
• Calm vs snapping is often about buffer levels (money, sleep, etc.).
• Run it (3 steps):
  1. Identify your thinnest buffer (usually cash or sleep).
  2. Set a floor (minimum acceptable level).
  3. Use refill-first: when below floor, refilling outranks new commitments.
• Key consequence: buffers convert emergencies into manageable events, not cascading crises.

Tool 10: Rate vs Level (judge direction, not snapshots)

• One snapshot can mislead because it shows level, not rate.
• Key idea: change is assessed by the slope (trend across time), not a single moment.
• Run it (3 steps):
  1. Pick a metric you track (weight, backlog, pages read, etc.).
  2. Record it this week and again in 2 weeks and 4 weeks.
  3. Decide using the slope between points—not the snapshot.
• Key lesson: quitting often happens when people abandon a good trajectory during an early low-change period.

---

Part 4 — Timing illusions (why delay fools you)

Tool 11: Delay

• Cause and effect are separated in time; the brain wrongly assumes causes are close to outcomes.
• Run it (3 steps):
  1. Before a change, write:
     • start date
     • earliest honest payoff date (domain expert estimate)
  2. Put the payoff date where you’ll see it; don’t evaluate early.
  3. When mid-delay despair hits, explicitly blame “the delay” (as a scheduled experience), not personal failure.

Tool 12: Overcorrection and Oscillation

• With delays, people overreact: they crank the wrong lever, wait too little, then swing back.
• Fix: reduce gain and wait out a full delay before correcting again.
• Run it (3 steps):
  1. Spot a swing pattern (crackdown → blowout → crackdown).
  2. Apply half the correction you feel in the moment.
  3. Use the honest delay as the review interval; don’t change until it arrives.

Tool 13: Bullwhip (amplification along a chain)

• Small upstream wobble becomes large downstream crisis due to stacked delays.
• Run it (3 steps):
  1. Trace a recurring blowup backwards until you find a small early source input.
  2. Steady that source for ~2 weeks (regular, modest adjustment).
  3. Judge success by reduced downstream swing size.
• Key strategy: smooth the source, never the symptom.

---

Part 5 — Standard shapes (archetypes)

Tool 14: Shifting the Burden

• A “quick fix” relieves symptoms quickly but weakens root capacity over time.
• Run it (3 steps):
  1. Name your most used quick fix.
  2. Name the root capacity it substitutes for.
  3. Cap the quick fix and schedule the slow root fix as a committed commitment with a start date.

Tool 15: Fixes that Fail (boomerang cost)

• Fix works short-term but creates a larger problem on delay.
• Run it (3 steps):
  1. Pick a repeated fix.
  2. Price the boomerang: cost in ~4 weeks (energy, money, trust).
  3. Decide with full price; keep the fix only if it’s worth it, ideally with constraints.

Tool 16: Limits to Growth

• Growth stalls because the system hits a constraint; the engine is not always the problem.
• Run it (3 steps):
  1. When progress flattens, write: “The engine is fine; something is being used up.”
  2. List what is being consumed (common: recovery, attention, partner patience).
  3. Relax the binding constraint directly; don’t keep feeding the engine past the plateau.

Tool 17: Tragedy of the Commons

• Shared resources degrade through reasonable individual withdrawals when no one is assigned to refill.
• Run it (3 steps):
  1. Name the commons (shared stock under strain) and who the “invisible refiller” is if any.
  2. Make the level visible (dark kills commons).
  3. Create one shared refill rule owned by everyone.

---

Part 6 — Matching solutions to problem type (Cynefin)

Tool 18: Cynefin Sort

• Using the wrong method on the wrong problem is the “most expensive mistake.”
• Five “worlds” determine which actions work:
  1. Clear: checklist works; cause/effect obvious.
  2. Complicated: expert needed.
  3. Complex: solutions emerge; run safe-to-fail probes and iterate after observing effects.
  4. Chaotic (crisis): act to stabilize first; diagnose later.
  5. Confused: danger from premature application of favorite methods regardless of terrain.
• Run it (4 steps):
  1. Ask if a checklist could fix it → if yes, apply.
  2. If checklist can’t, but an expert could → buy/borrow analysis.
  3. If experts can’t know reliably → use three small safe-to-fail probes with review dates.
  4. If on fire → stabilize and then sort later.

---

Part 7 — Where to push (leverage points)

Tool 19: The Leverage Ladder

• A ranked scale of where interventions work best (weakest → strongest). Five rungs:
  1. Numbers (change a quantity, but underlying loops may remain)
  2. Structure (change physical setup; environment shapes behavior)
  3. Information (make hidden stocks/levels visible to change behavior without lectures)
  4. Rules (change what system aims to achieve; e.g., refill rules, timing rules)
  5. Story / assumptions (deep assumptions underpin everything; highest leverage but slowest)
• Run it (3 steps):
  1. Locate your current attempt’s rung honestly (often rung 1).
  2. Design an intervention one rung higher.
  3. Run it ~2 weeks and compare against what lower rungs produced.

Tool 20: Niche Strategy (how locked-in life actually changes)

• Big changes fail because they assault the “incumbent regime” (current structure of daily life).
• Transition concept:
  • Landscape: slow external shifts (industry changes, tech, values, etc.)
  • Regime: current dominant arrangement defended by routines and expectations
  • Niches: small protected experiments that mature without competing immediately
• Run it (3 steps):
  1. Name your regime honestly (daily shape to defeat).
  2. Build a small niche (protected, scheduled, low-stakes, permission-light).
  3. Identify the next plausible window (landscape shift) and decide what the niche claims when that window opens.

---

Capstone protocol — “Read the system” (4 steps, one page)

The 4-step one-page method

1. Name the stock that’s straining (what level is low or draining).
2. Trace the loop that drains it (reinforcing/balancing dynamics and their direction).
3. Find the delay that’s fooling you (cause/effect time gap).
4. Climb one rung up (stronger leverage point instead of pushing harder on symptoms).

Example interpretations used in the course

• Office example (manager’s collapsing weeks):
  • Stock: energy dropping
  • Loop: stress → shallow work → rework → backlog → more stress
  • Delay: short nights Monday show up as Thursday blame
  • One rung up: move toward protected structure + visibility (e.g., protected mornings + workload board)
• Kitchen example (parent’s empty evenings):
  • Stock: patience/available goodwill (low early evening)
  • Loop: tiredness → takeaway/scroll → sleep degradation → next day tiredness
  • Delay: evening choices affect tomorrow
  • One rung up: fridge list + rule change (screens down at 10; change set point via structure)

---

Closing lessons / final framing

• People act rationally based on what their local system signals, delays, and incentives deliver.
• Therefore:
  • No character transplant is required
  • Fix the structure and leverage, and behavior changes without demanding impossible willpower.
• The course argues it’s often not “you”; it’s the system around you.

---

Speakers / sources featured

• MIT scientist / systems education references: described as the author of the defining book on systems thinking (referenced via “Elements, interconnections, purpose”).
• Donella Meadows: referred to as the scientist who wrote this field’s defining book; author of the leverage ladder paper ranking intervention points.
• Elinor Ostrom / ecologists and “tragedy of the commons” framing: subtitles reference ecological/governance research; specific name “Ostrom” is not present, but the concept is attributed to ecologists.
• Cynefin framework creator (Welsh researcher): name not provided in subtitles.
• A Dutch scholar (transition/niche strategy): name not provided in subtitles.