/cynefin

Cynefin Analysis

Analyze a situation through the Cynefin framework to understand what type of problem it is and choose the appropriate response.

Instructions

Determine which Cynefin domain the problem belongs to, then apply the appropriate response approach. Be willing to recognize that parts of a problem may span multiple domains.

Output Format

**Situation**: [What we're analyzing]

---

Initial Assessment

**What kind of problem is this?**

Characteristics Check

| Characteristic | Yes/No/Maybe | Evidence | |---------------|--------------|----------| | Cause and effect are clear and predictable | | | | Experts can analyze and determine the answer | | | | Outcomes can only be understood in hindsight | | | | The situation is unstable/in crisis | | | | We genuinely don't know what kind of problem this is | | |

---

Domain Classification

**Primary Domain**: [Simple/Complicated/Complex/Chaotic/Confused]

**Why this domain?** [Reasoning for the classification]

**Is this a pure domain or does it span multiple?** [Assessment of whether different aspects belong to different domains]

---

Domain-Appropriate Response

If SIMPLE/OBVIOUS

*Sense → Categorize → Respond*

**Best Practice to Apply**: [The established, proven approach]

**Implementation**: 1. [Step 1] 2. [Step 2] 3. [Step 3]

**Risk**: Don't overcomplicate this. Don't call in experts for simple problems.

---

If COMPLICATED

*Sense → Analyze → Respond*

**Experts Needed**: [Who has the expertise to analyze this?]

**Analysis Required**: [What do we need to investigate?]

**Good Practice Options**: | Option | Expert View | Trade-offs | |--------|-------------|------------| | [option] | [analysis] | [trade-offs] |

**Risk**: Multiple right answers exist. Analysis paralysis is possible.

---

If COMPLEX

*Probe → Sense → Respond*

**Safe-to-Fail Experiments**: | Probe | What We'd Learn | Amplify if... | Dampen if... | |-------|-----------------|---------------|---------------| | [experiment] | [insight] | [success indicators] | [failure indicators] |

**Emergence**: What patterns might emerge? We don't know yet—that's the point.

**Key Mindset**: Don't try to predict. Try things, observe, adapt.

**Risk**: Temptation to analyze instead of probe. "We need more data" delays action.

---

If CHAOTIC

*Act → Sense → Respond*

**Immediate Stabilization**: [What action do we take right now to stabilize?]

**Novel Practice**: [What approach might work? This is invention, not best practice]

**Move to Complex**: Once stabilized, this should become complex, not return to chaotic.

**Risk**: Don't try to analyze during chaos. Act first.

---

If CONFUSED

*Break down, then categorize*

**How to break this down**: | Sub-problem | Likely Domain | Why | |-------------|---------------|-----| | [sub-problem 1] | [domain] | [reasoning] | | [sub-problem 2] | [domain] | [reasoning] |

**First step**: [How to get out of confusion]

---

Boundary Warnings

**Are we in danger of slipping between domains?**

| Transition | Warning Signs | What to Do | |------------|--------------|------------| | Simple → Chaotic | Complacency, "we've always done it this way" | Challenge assumptions | | Complicated → Complex | Analysis isn't converging; experts disagree | Try probes instead | | Complex → Chaotic | Loss of control; things are degrading | Stabilize fast |

---

Recommended Approach

**Given this is a [DOMAIN] problem, we should:**

1. [Primary approach] 2. [Supporting action] 3. [Monitoring for domain shift]

**What NOT to do**: [Approaches that would be wrong for this domain]

Guidelines

• Most people default to Complicated (analysis) when Complex (probes) is needed
• "Best practice" is only valid in Simple domain
• In Complex, you can't predict—you can only experiment and adapt
• In Chaotic, act first, think later
• Confusion is normal—break the problem down

$ARGUMENTS