Back to prompts
Decision MakingPremiumadvanced
4.7

Apply Second-Order Thinking to Any Decision

Go beyond 'what happens next' to predict the downstream consequences most people miss

Copy & Paste this prompt
I'm considering [ACTION/DECISION/POLICY].

Context: [SITUATION, CONSTRAINTS, GOALS]

Apply second-order (and third-order) thinking:

1. FIRST-ORDER EFFECTS
   What happens immediately?
   - List 3-5 direct, obvious consequences
   - Who is affected first?
   - What changes right away?

2. SECOND-ORDER EFFECTS
   What do those first-order effects CAUSE?
   - For each first-order effect, trace 2-3 downstream consequences
   - Who is affected that wasn't obvious at first?
   - What behaviors change in response to the first-order effects?
   - What systems or incentives are altered?

3. THIRD-ORDER EFFECTS
   Go one level deeper:
   - What feedback loops are created? (reinforcing or balancing)
   - What unintended consequences emerge over 6-12 months?
   - Could this trigger a cascade or tipping point?

4. HISTORICAL PARALLELS
   - Find 2-3 situations where someone made a similar decision
   - What second-order effects surprised them?
   - What can we learn from their experience?

5. PERVERSE INCENTIVES
   - Does this decision create incentives that work AGAINST the intended goal?
   - Could people game or exploit this in ways I haven't considered?
   - What's the "cobra effect" scenario? (where the solution makes the problem worse)

6. TIME HORIZONS
   Map the likely outcomes at:
   - 1 week: [immediate effect]
   - 1 month: [short-term adjustments]
   - 6 months: [medium-term consequences]
   - 2 years: [long-term systemic changes]

7. REVISED DECISION
   Now that you see the cascade:
   - Should I still do this? Why or why not?
   - How should I modify the decision to capture benefits while avoiding negative second-order effects?
   - What should I monitor to catch unintended consequences early?

Think like a chess player — it's not about the next move, it's about the position it creates.
#decision-making#strategy#apply#second-order#thinking

Works with

chatgptclaudegemini

💡 Pro Tips

  • •Always ask 'and then what?' at least three times — most people stop after the first answer
  • •Look for perverse incentives — the cobra effect (where the solution worsens the problem) is more common than you think
  • •Second-order thinking is especially critical for policies, pricing changes, and anything that changes incentives

✨ Example Output

Decision: "Offer a 30% discount to win back churned customers"

FIRST-ORDER EFFECTS:
1. Some churned customers return → short-term revenue boost
2. Marketing team has a campaign to run → clear action item
3. Revenue per returning customer is lower (30% off)

SECOND-ORDER EFFECTS:
1→ Active customers who are paying full price find out → feel penalized for loyalty → some threaten to cancel to get the discount too
2→ Team celebrates "win-back" numbers → anchors future strategy on discounts → erodes pricing power
3→ Returned customers churned for a reason → if root cause isn't fixed, they'll churn again → now you've spent marketing dollars AND given a discount for a 60-day customer

THIRD-ORDER EFFECTS:
→ A discount culture develops: customers learn to churn and wait for the win-back offer → churn INCREASES because churning is now incentivized
→ Support team gets frustrated re-onboarding the same customers → morale drops
→ Competitor notices your discounting → prices aggressively → race to the bottom

PERVERSE INCENTIVE: 🐍 COBRA EFFECT DETECTED
You're literally paying customers to churn. The rational customer strategy becomes: subscribe → cancel → wait for 30% off → resubscribe → repeat.

REVISED DECISION:
Don't offer a blanket discount. Instead:
1. Segment churned customers by reason (price vs. product vs. support)
2. For price-sensitive: offer an annual plan at 20% off (locks them in)
3. For product issues: show them the new features they missed (no discount)
4. For support issues: personal outreach from a senior rep
Monitor: Track if win-back customers churn again within 90 days. If >40% do, the strategy is failing.