The essence of decision-making: when information is never complete and the future is never certain,
use a system to raise the “hit rate” and lower the “cost of being wrong.”
This is the only “engine” in this book — it drives every other layer.
1. Theoretical Foundations (Papers / Ideas)
1.1 Herbert Simon — “Bounded Rationality” (Nobel Prize)
- Humans are not omniscient; decisions aim for a satisficing solution, not an optimal one
- Method: set the criteria first → execute the first option that meets them → don’t chase perfection
- Takeaway: waiting forever for the “optimal solution” = never deciding at all
1.2 Kahneman — System 1 / System 2 (Nobel Prize)
- System 1: fast, intuitive, automatic (effortless, but error-prone)
- System 2: slow, rational, effortful (accurate, but energy-intensive)
- Use System 2 for major decisions, while staying alert to System 1 biases:
- Anchoring (being swayed by the first number you see)
- Confirmation bias (only seeking evidence that supports your view)
- Overconfidence (overestimating the accuracy of your own judgment)
1.3 Klein — Naturalistic Decision Making: the RPD Model
- An expert’s “intuition” under pressure = pattern recognition (built on extensive experience)
- Beginners shouldn’t rely on intuition; veterans can move fast in familiar domains
- Method: recognize the situation → match a pattern → run a mental simulation → act
1.4 Taleb — Antifragility
- In uncertainty, don’t “predict” — “prepare”
- Keep redundancy (cash / backup suppliers / backup plans)
- Make volatility work for you (barbell strategy: extremely conservative + extremely aggressive)
1.5 Shell — Scenario Planning
- Don’t make a single forecast; build 2–4 scenarios (best / worst / base case)
- Pair each scenario with a contingency plan
- Case: in the 1970s Shell used scenario planning to anticipate the oil crisis and rose from last place in the industry to first
1.6 Ren Zhengfei (Huawei founder) — “Gray Philosophy”
- Don’t agonize over black-and-white choices — start moving, then calibrate
- “Direction roughly right, organization full of vitality”
- Tolerate trial and error, tolerate gray areas — but never tolerate stagnation
2. A Decision Method You Can Actually Use: The Vex Five-Step Method
① Classify the Decision Type
Bezos: one-way doors vs. two-way doors
| Type |
Characteristics |
Decision speed |
| Two-way door (reversible) |
Cheap, can turn back |
Fast (decide with 70% of the information) |
| One-way door (irreversible) |
Expensive, hard to reverse |
Slow (verify more, consult more) |
- Example: switching suppliers (reversible) → move fast; building a new factory (irreversible) → move slow
- ⚠️ Most CEOs drag reversible decisions into slow motion and rush irreversible ones
- What information is missing? Cost of getting it vs. cost of waiting?
- Decide with 70% of the information (Bezos): waiting for 100% usually means missing the window
- Decision matrix:
Information sufficient Information insufficient
Low cost of action → Just do it Small experiments
High cost of action → Analyze fully, then act Scenario planning + options thinking
③ Hedge Uncertainty with Rules
Set decision principles in advance (when emotions spike, follow the rules — don’t improvise):
- “Never bet on a single customer” (customer concentration red line)
- “Cash flow red line” (no expansion when cash on hand covers less than X months of expenses)
- “Never invest in domains I don’t understand”
- “Every major outlay gets a stop-loss line”
④ Small Experiments / Options Thinking
Break big bets into reversible small bets:
- New market → pilot on a small scale first
- New product → validate with an MVP first
- New customer → start with a small order
- Scale up only after validation (real options: buy the “right to decide later,” not a one-time all-in bet)
⑤ Set Checkpoints (Bayesian Updating)
Agree upfront: under what conditions will we reassess / stop-loss / double down:
- “If the new customer contributes < X after 3 months, kill it”
- “If gross margin falls below X%, raise prices / switch suppliers”
- Revise your judgment when new evidence arrives — never refuse to admit a mistake out of pride or sunk cost
3. How Decision-Making Connects to the Rest of the System
Strategy sets direction (Ch. 01)
↓
Budget allocates resources (Ch. 01)
↓
Decisions happen constantly in execution (this layer) ← needs an information pipeline (Ch. 05 "Data")
↓
Review & retrospective (Ch. 01) → corrects the next decision
The information pipeline (the foundation of decisions):
- Operating cockpit: daily revenue / gross margin / cash flow / inventory / work-in-progress
- Management accounting: true profit by product / by customer / by project
- Meeting cadence: monthly operating review (numbers) + quarterly retrospective (direction)
4. CEO Decision-Making Field Checklist
Ask yourself 5 questions before every major decision:
- Is this a one-way door or a two-way door? (sets the pace)
- What information am I missing? Is it worth waiting for? (decides whether to gather more first)
- What’s the worst case? Can I survive it? (decides the size of the bet)
- Is there a smaller, more reversible way to validate? (decides the path)
- When will I reassess? (decides the checkpoint)
Recognize your own biases:
- Am I only looking at evidence that supports me? (confirmation bias)
- Am I anchored to the first number I saw? (anchoring)
- Am I overestimating my odds? (overconfidence)
- Am I refusing to cut losses because I’ve already invested too much? (sunk cost)
5. Case Library
| Company / Person |
Decision |
Methodology |
Lesson |
| Jeff Bezos |
AWS, from internal need to public service |
Two-way door + 70% information |
Move fast on reversible decisions |
| Ren Zhengfei |
Gray philosophy; “let those who hear the gunfire call for artillery” |
Gray + delegation |
Direction roughly right is enough |
| Wang Chuanfu |
Betting big on EVs in 2003 |
Technology fish pond + resolute bet |
Bets must be survivable |
| Shell |
1970s scenario planning |
Scenario planning |
Use scenarios, not forecasts, for uncertainty |
| Toyota |
Stops production at the first sign of a quality problem |
Andon + stop-the-line principle |
Rules hedge against emotion |
| Luckin Coffee |
2020 fraud crisis → rapid restructuring |
Admit mistakes fast + rebuild |
Cut losses decisively |
| Nokia |
Refused to pivot to smartphones |
Held hostage by sunk cost |
Cut what must be cut |
📌 One-line summary: CEO decision-making = classify reversibility × identify information gaps × hedge with rules × validate in small steps × set checkpoints.