reversible-decisions
À propos
Cette compétence met en œuvre le cadre de Jeff Bezos pour classer les décisions comme réversibles (Type 2) ou irréversibles (Type 1), aidant ainsi les développeurs à prioriser leurs efforts. Elle permet aux équipes d'avancer rapidement sur les choix à faible enjeu et réversibles, tout en appliquant une prudence appropriée aux décisions à haut risque et irréversibles. Utilisez-la pour déléguer les décisions, éviter la paralysie par l'analyse et gérer les risques dans les flux de travail de développement.
Installation rapide
Claude Code
Recommandénpx skills add guia-matthieu/clawfu-skills -a claude-code/plugin add https://github.com/guia-matthieu/clawfu-skillsgit clone https://github.com/guia-matthieu/clawfu-skills.git ~/.claude/skills/reversible-decisionsCopiez et collez cette commande dans Claude Code pour installer cette compétence
Documentation
Reversible Decisions (Type 1 vs. Type 2)
Know when to move fast and when to move carefully. Master Jeff Bezos' framework for distinguishing high-stakes irreversible decisions from low-stakes reversible ones.
When to Use This Skill
- Prioritizing decisions to know where to invest time
- Team empowerment to understand what to delegate vs. escalate
- Avoiding analysis paralysis on decisions that don't matter
- Risk management to identify where caution is truly warranted
- Speed vs. thoroughness trade-offs in any context
- Building decision-making culture in organizations
Methodology Foundation
| Aspect | Details |
|---|---|
| Source | Jeff Bezos - Amazon shareholder letters (2015-2016) |
| Core Principle | "Some decisions are irreversible and consequential (Type 1). Most are reversible and low-consequence (Type 2). Use the right process for each." |
| Why This Matters | Most people treat all decisions like Type 1—slow, deliberate, requiring full information. This leads to paralysis and missed opportunities. The best decision-makers move fast on Type 2 and slow on Type 1. |
What Claude Does vs What You Decide
| Claude Does | You Decide |
|---|---|
| Structures content frameworks | Final messaging |
| Suggests persuasion techniques | Brand voice |
| Creates draft variations | Version selection |
| Identifies optimization opportunities | Publication timing |
| Analyzes competitor approaches | Strategic direction |
What This Skill Does
- Classifies decisions - Is this Type 1 or Type 2?
- Calibrates process to stakes - Right speed for right decision
- Enables delegation - Type 2 can be pushed down
- Prevents over-analysis - Stop treating reversible decisions as irreversible
- Improves organizational speed - Teams move faster on the right things
- Reduces decision fatigue - Don't waste energy on low-stakes choices
How to Use
Classify a Decision
Help me classify this decision:
[Describe the decision]
Is this Type 1 (irreversible) or Type 2 (reversible)?
What process should I use?
Speed Up Decision-Making
I'm spending too much time on [decision].
Apply the Type 1/Type 2 framework to help me move faster.
Build Decision-Making Process
Help me design a decision-making framework for my team.
Which decisions should require consensus vs. individual judgment?
Instructions
Step 1: Understand the Framework
## Type 1 vs. Type 2 Decisions
### Bezos' Definition
**Type 1: One-Way Doors**
"Some decisions are consequential and irreversible or nearly irreversible—
one-way doors—and these decisions must be made methodically, carefully,
slowly, with great deliberation and consultation."
**Type 2: Two-Way Doors**
"But most decisions aren't like that—they are changeable, reversible—
they're two-way doors. If you've made a suboptimal Type 2 decision,
you don't have to live with the consequences for that long.
You can reopen the door and go back through."
### The Problem
"As organizations get larger, there seems to be a tendency to use
the heavy-weight Type 1 decision-making process on most decisions,
including many Type 2 decisions. The end result of this is slowness,
unthoughtful risk aversion, failure to experiment sufficiently,
and consequently diminished invention."
### The Solution
"We must resist this tendency."
Type 2 decisions should be made quickly by high-judgment individuals
or small groups. Type 1 decisions require the full deliberative process.
Step 2: Classification Framework
## How to Classify Decisions
### The Two Questions
**Question 1: Is it reversible?**
Can you undo this decision with reasonable effort and cost?
| Reversibility | Examples |
|---------------|----------|
| **Easily reversible** | Pricing change, A/B test, new feature flag, hire (with trial), campaign |
| **Hard to reverse** | Architecture choice, brand name, key hire (C-level), market exit |
| **Irreversible** | Selling company, shutting down product, firing someone, legal action |
**Question 2: What are the consequences?**
If this decision is wrong, what happens?
| Consequence Level | Examples |
|-------------------|----------|
| **Low** | Internal process change, small experiment, minor feature |
| **Medium** | New product launch, pricing tier, team restructure |
| **High** | Major strategic pivot, large investment, partnership |
| **Existential** | Acquisition, shutdown, bet-the-company move |
### The Matrix
CONSEQUENCES
Low High
REVERSIBILITY ┌────────────┬────────────┐ High │ TYPE 2 │ TYPE 2 │ (Easy) │ (Fast) │ (Fast w/ │ │ │ monitoring)│ ├────────────┼────────────┤ Low │ TYPE 2 │ TYPE 1 │ (Hard) │ (Careful) │ (Slow) │ └────────────┴────────────┘
### Quick Classification
**TYPE 2 (Move Fast):**
- Can be undone
- Low/medium consequences
- Learning opportunity
- Failure is recoverable
- Most business decisions
**TYPE 1 (Move Carefully):**
- Can't be undone
- High/existential consequences
- Mistakes are permanent
- One-way door
- ~5-10% of decisions
Step 3: Match Process to Type
## Decision Process by Type
### Type 2 Process (70% of Decisions)
**Time:** Hours to days (not weeks)
**Who:** Individual or small group with context
**Information:** Good enough, not perfect
**Approval:** None or single level
**Documentation:** Minimal (decision log)
**The Mantra:**
"Disagree and commit" - If you have 70% of the information you wish you had,
make the decision. Waiting for 90% is usually too slow.
**Process:**
1. Identify it's Type 2 (reversible, recoverable)
2. Gather available information quickly
3. Make the call
4. Communicate the decision
5. Monitor and adjust
**Examples:**
- Feature prioritization
- Hiring most roles
- Process changes
- Pricing experiments
- Marketing campaigns
- Internal tools
- Meeting schedules
---
### Type 1 Process (5-10% of Decisions)
**Time:** Weeks to months
**Who:** Senior leadership, broad input
**Information:** As complete as reasonably possible
**Approval:** Multiple stakeholders
**Documentation:** Thorough (rationale, alternatives, risks)
**The Mantra:**
"Measure twice, cut once" - This is permanent. Get it right.
**Process:**
1. Confirm it's Type 1 (irreversible, consequential)
2. Define decision criteria clearly
3. Gather comprehensive information
4. Consider alternatives thoroughly
5. Consult relevant stakeholders
6. Document the reasoning
7. Make the decision
8. Communicate extensively
**Examples:**
- M&A decisions
- Major strategic pivots
- Leadership hires (C-level)
- Market entry/exit
- Large capital allocation
- Shutting down products
- Legal/regulatory choices
Step 4: Common Traps
## Decision-Making Traps
### Trap 1: Treating Type 2 as Type 1
**Symptom:** Analysis paralysis on small decisions
**Example:** 2-week committee review for a landing page change
**Problem:** Slows innovation, frustrates teams, misses opportunities
**Fix:** Ask "What's the worst case if we're wrong? Can we fix it?"
### Trap 2: Treating Type 1 as Type 2
**Symptom:** Moving too fast on irreversible choices
**Example:** Acquiring a company in 2 weeks
**Problem:** Permanent mistakes, existential risk
**Fix:** Ask "If this goes wrong, can we undo it?"
### Trap 3: Requiring Consensus on Type 2
**Symptom:** Everyone needs to agree before action
**Example:** 10-person meeting to decide email copy
**Problem:** Slowest person becomes bottleneck
**Fix:** Empower individuals to make Type 2 calls
### Trap 4: Not Recognizing Type 1 in Disguise
**Symptom:** Missing irreversibility hidden in details
**Example:** "Small" technical choice that creates years of debt
**Problem:** Accumulated Type 1 decisions dressed as Type 2
**Fix:** Consider second-order effects
### Trap 5: Using Decision Type as Excuse
**Symptom:** Calling everything Type 1 to avoid responsibility
**Example:** "We need more research" on every decision
**Problem:** Becomes cover for inaction
**Fix:** Default assumption = Type 2 unless proven otherwise
Step 5: Decision Process Template
## Decision Classification Template
### The Decision
[What are you deciding?]
### Classification Checklist
**Reversibility Assessment:**
□ Can this be undone in days/weeks?
□ What's the cost of reversing?
□ Are there path dependencies that make reversal hard?
□ Will this create commitments we can't exit?
**Consequence Assessment:**
□ If wrong, what's the impact on revenue?
□ If wrong, what's the impact on reputation?
□ If wrong, what's the impact on team?
□ If wrong, is the company at risk?
### Classification: ____________
| Factor | Type 1 Indicator | Type 2 Indicator | This Decision |
|--------|------------------|------------------|---------------|
| Reversibility | Can't undo | Can undo | |
| Time to undo | Months/Never | Days/Weeks | |
| Cost to undo | High/Impossible | Low/Moderate | |
| Consequence | Existential | Recoverable | |
| Frequency | Rare | Common | |
### Process to Use
**If Type 2:**
- Who should decide? _______________
- Timeline: _______________ (days, not weeks)
- Information needed: _______________
- Communication: _______________
**If Type 1:**
- Who needs to be involved? _______________
- Timeline: _______________ (weeks/months)
- What information is missing? _______________
- What alternatives exist? _______________
- What's the documentation plan? _______________
Examples
Example 1: Product Feature Decision
Input:
"Should we add a dark mode to our app? My team is debating this for two weeks."
Output:
Decision Analysis: Dark Mode Feature
Classification
| Factor | Assessment |
|---|---|
| Reversible? | YES - Can ship, can remove if bad |
| Cost to undo | LOW - One sprint to remove |
| Consequence if wrong | LOW - Users toggle it off, minor wasted effort |
| Consequence if right | MEDIUM - Better user experience, some retention |
Verdict: TYPE 2 (Two-Way Door)
Two weeks of debate on a reversible feature decision is organizational dysfunction.
Recommended Process
Time allowed: 2-3 days max Who decides: Product manager + 1 engineer Information needed:
- Quick user survey (100 responses)
- Competitive analysis (2 hours)
- Engineering estimate (30 min)
The decision:
- Is there signal users want it? (Survey >30% interest = yes)
- Is the cost reasonable? (<1 week engineering = yes)
- Does it conflict with anything? (Check roadmap)
If yes to 1 & 2, and no to 3: Ship it. If uncertain: Ship it behind a feature flag. Measure.
What NOT To Do
❌ Two weeks of meetings ❌ Executive approval ❌ Perfect user research ❌ Consensus across org ❌ Waiting for more data
The Bezos Test
"If we ship dark mode and it's wrong, what happens?"
Answer: We remove it in a sprint. A few users briefly confused. No real damage.
Conclusion: Make the call. Move on. Spend your deliberation budget on something that matters.
Example 2: Pricing Strategy Change
Input:
"We're considering changing from per-seat pricing to usage-based pricing. How should we approach this?"
Output:
Decision Analysis: Pricing Model Change
Classification
| Factor | Assessment |
|---|---|
| Reversible? | PARTIALLY - Technically yes, but... |
| Cost to undo | HIGH - Customer confusion, churn, internal systems |
| Path dependencies | HIGH - New customers on new model can't easily migrate |
| Consequence if wrong | HIGH - Revenue impact, customer churn, competitive exposure |
Deeper Analysis
Why this isn't purely Type 2:
- Existing customers expect current pricing
- Changing back signals instability
- Systems, contracts, sales comp all need updating
- Market perception shifts
Why it's not purely Type 1:
- You CAN grandfather existing customers
- You CAN run both models in parallel
- You CAN test with new customers only first
Verdict: TYPE 1.5 (Needs Careful Handling)
This is a major strategic decision with reversibility caveats.
Recommended Process
Time: 4-6 weeks of analysis Who: CEO, CFO, Head of Product, Head of Sales Information needed:
- Financial modeling of both scenarios
- Customer interviews (willingness to switch)
- Competitive analysis
- Engineering/ops assessment
- Sales compensation impact
Risk Mitigation:
- Test with new customers only first (Type 2 experiment)
- Keep existing customers on old pricing (reversibility hedge)
- Run for 6 months before migrating anyone
- Build systems for both models
Decision Framework
| Approach | Reversibility | Speed | Risk |
|---|---|---|---|
| Full switch | Low | Fast | High |
| New customers only | High | Medium | Medium |
| Parallel models | High | Slow | Low |
Recommendation: Start with new customers only (Type 2). Migrate existing only after validation (then it becomes Type 1).
Example 3: Team/Organization Decision
Input:
"We're debating whether to restructure the engineering team. Some want squads, some want feature teams. We've been discussing for a month."
Output:
Decision Analysis: Team Structure
Classification
| Factor | Assessment |
|---|---|
| Reversible? | YES - Reorgs happen all the time |
| Cost to undo | MEDIUM - Productivity dip during change |
| Consequence if wrong | MEDIUM - Slower delivery for a quarter |
| Path dependencies | LOW - People can move teams |
Verdict: TYPE 2 (But Feels Like Type 1)
This FEELS irreversible because it affects people and emotions are high. But practically? Teams restructure regularly. This is recoverable.
The Real Problem
A month of debate isn't analysis—it's conflict avoidance.
What's actually happening:
- People have opinions and aren't yielding
- No one wants to make a call and be "responsible"
- The debate is comfortable; the decision is uncomfortable
Recommended Process
Time: 1 more week, max Who decides: Engineering lead (or whoever is accountable) Process:
- Write up both options (1 page each)
- Define success criteria (what metrics improve?)
- Pick one
- Commit for 6 months (review then)
- "Disagree and commit" - those who disagree still execute
The Forcing Function
"We will decide by [Friday]. Whoever feels strongest makes the call and is accountable for making it work. We all commit to supporting it for 6 months before reassessing."
Type 2 Permission
Say this to the team: "This is a two-way door. We can change it later. But we can't debate forever. Let's pick one, run it for 6 months, measure, and adjust. The worst outcome is paralysis."
Checklists & Templates
Quick Classification Checklist
## Is This Type 1 or Type 2?
□ Can we undo this in <30 days?
□ If wrong, will we lose <10% of something important?
□ Is this a common decision (we'll make many like it)?
□ Can we experiment/test before committing?
□ Are the consequences contained?
**Mostly YES → Type 2 (Move fast)**
**Mostly NO → Type 1 (Move carefully)**
### Default Rule
"When in doubt, it's Type 2. Most decisions are."
Team Decision Matrix Template
## Team Decision-Making Framework
### Type 2 Decisions (Individual/Small Group)
- Feature prioritization
- Bug fixes
- Process improvements
- Hiring (non-leadership)
- Tool selection
- Meeting schedules
- Internal communications
**Process:** Inform, decide, execute
**Timeline:** Hours to days
**Approval:** None needed
### Type 1 Decisions (Leadership/Broader Input)
- Strategic direction
- Major investments (>$X)
- Leadership hiring
- Pricing strategy
- Market entry/exit
- Partnerships
- Shutting down products
**Process:** Analyze, consult, deliberate, decide
**Timeline:** Weeks
**Approval:** [Define levels]
### Escalation Criteria
Escalate Type 2 to Type 1 if:
- Cost exceeds $[X]
- Affects >N customers
- Creates legal/compliance risk
- Changes company strategy
- Irreversible commitment
Skill Boundaries
What This Skill Does Well
- Structuring persuasive content
- Applying copywriting frameworks
- Creating draft variations
- Analyzing competitor approaches
What This Skill Cannot Do
- Guarantee conversion rates
- Replace brand voice development
- Know your specific audience
- Make final approval decisions
References
- Bezos, Jeff. "Amazon Shareholder Letters" (2015, 2016) - Type 1/Type 2 framework
- Blank, Steve. "The Four Steps to the Epiphany" - Speed in startups
- Ries, Eric. "The Lean Startup" - Reversible experiments
- Farnam Street. "Mental Models" - Decision frameworks
- Amazon. "Leadership Principles" - Bias for action
Related Skills
- second-order-thinking - Consider consequences
- regret-minimization - Long-term decision view
- first-principles - Challenge assumptions
- pre-mortem - Anticipate failures
- eisenhower-matrix - Prioritization
Skill Metadata
- Mode: cyborg
name: reversible-decisions
category: thinking
subcategory: decision-making
version: 1.0
author: MKTG Skills
source_expert: Jeff Bezos
source_work: Amazon Shareholder Letters
difficulty: beginner
estimated_value: $2,000 management consulting session
tags: [decisions, Bezos, Amazon, speed, reversibility, management, delegation]
created: 2026-01-25
updated: 2026-01-25
Dépôt GitHub
Compétences associées
executing-plans
DesignUtilisez la compétence executing-plans lorsque vous disposez d'un plan de mise en œuvre complet à exécuter par lots contrôlés avec des points de contrôle de revue. Elle charge et examine le plan de manière critique, puis exécute les tâches par petits lots (3 tâches par défaut) tout en rapportant la progression entre chaque lot pour une revue par l'architecte. Cela garantit une mise en œuvre systématique avec des points de contrôle de qualité intégrés.
requesting-code-review
DesignCette compétence délègue un sous-agent réviseur de code pour analyser les modifications apportées au code par rapport aux exigences avant de poursuivre. Elle doit être utilisée après avoir terminé des tâches, implémenté des fonctionnalités majeures, ou avant une fusion vers la branche principale. La revue aide à détecter précocement les problèmes en comparant l'implémentation actuelle avec le plan initial.
connect-mcp-server
DesignCette compétence fournit un guide complet permettant aux développeurs de connecter des serveurs MCP à Claude Code via les transports HTTP, stdio ou SSE. Elle couvre l'installation, la configuration, l'authentification et la sécurité pour intégrer des services externes tels que GitHub, Notion et des API personnalisées. Utilisez-la lors de la configuration d'intégrations MCP, de la configuration d'outils externes ou du travail avec le Protocole de Contexte de Modèle de Claude.
web-cli-teleport
DesignCette compétence aide les développeurs à choisir entre les interfaces Web et CLI de Claude Code en fonction de l'analyse des tâches, puis permet une téléportation transparente des sessions entre ces environnements. Elle optimise le flux de travail en gérant l'état et le contexte de la session lors du passage entre le web, la CLI ou le mobile. Utilisez-la pour des projets complexes nécessitant différents outils à diverses étapes.
