Skip to main content

Cynefin Framework

TL;DR

Cynefin: Match your response to your domain. Clear problems have known solutions β€” apply best practice. Complicated problems need analysis β€” bring in experts. Complex problems have no known solutions β€” run probes and learn. Chaotic problems need immediate action to stabilise. The most common error: treating Complex like Complicated (hiring experts to "solve" what needs to be explored) or Clear like Complex (overanalysing obvious problems).

What Is the Cynefin Framework?​

Developed by Dave Snowden at IBM's Institute for Knowledge Management in 1999, Cynefin (a Welsh word meaning habitat or the place of multiple belongings) is a decision-making framework that classifies situations by their relationship between cause and effect. The framework argues that different types of problems require fundamentally different responses β€” and that misclassification is a primary source of managerial failure.

The five domains:

Clear (formerly Simple): Cause and effect are obvious to everyone. Best practices exist and are known. Sense β†’ Categorise β†’ Respond. Examples: processing a standard customer refund, following a fire evacuation procedure.

Complicated: Cause and effect are not immediately obvious but can be determined by analysis. Multiple right answers exist; expertise is required to identify them. Sense β†’ Analyse β†’ Respond. Examples: diagnosing a mechanical fault, designing a bridge.

Complex: Cause and effect can only be understood retrospectively; they cannot be predicted in advance. Multiple interacting agents produce emergent outcomes that no expert can predict. Probe β†’ Sense β†’ Respond. Examples: launching a new product into an uncertain market, managing organisational culture change.

Chaotic: There is no cause-and-effect relationship discernible. Immediate action is required to establish order before any analysis is possible. Act β†’ Sense β†’ Respond. Examples: a crisis, a sudden market collapse, a major security breach in progress.

Disorder (the centre): The domain in which most organisations actually spend their time β€” not knowing which domain applies and defaulting to whatever response they're most comfortable with.

How It Works​

Step 1: Assess cause-and-effect relationship
  β€” Is the cause-effect relationship obvious? β†’ Clear
  β€” Discoverable through analysis? β†’ Complicated
  β€” Only apparent in hindsight? β†’ Complex
  β€” Not discernible at all right now? β†’ Chaotic

Step 2: Select the appropriate response pattern
  β€” Clear: Sense β†’ Categorise β†’ Respond (apply best practice)
  β€” Complicated: Sense β†’ Analyse β†’ Respond (use expertise)
  β€” Complex: Probe β†’ Sense β†’ Respond (run experiments)
  β€” Chaotic: Act β†’ Sense β†’ Respond (stabilise first)

Step 3: Watch for domain shifts
  β€” Clear can become Chaotic if over-simplified
  β€” Chaotic should move toward Complex as order is restored
  β€” Complex can produce local pockets of Complicated clarity

Step 4: Avoid domain misclassification
  β€” Expert bias: treating Complex like Complicated
  β€” Action bias: treating Complicated like Chaotic

Three Real-World Examples​

Startup Product Development​

A startup building a new consumer app is operating in the Complex domain: they cannot predict in advance which features users will value, how the market will respond to pricing, or which acquisition channels will work. The appropriate response is Probe β†’ Sense β†’ Respond: run small experiments (MVPs, landing page tests, user interviews), observe results, and iterate based on what's learned.

The common error is treating this as Complicated: hiring experts to "design the right product" before building it. In complex environments, expertise at predicting user behaviour is limited β€” no expert can tell you in advance whether your specific feature set will resonate with your specific target market. Only probing (shipping, measuring) reveals the answer.

Aviation Maintenance​

Aircraft maintenance operates in the Complicated domain: cause-and-effect relationships are not immediately obvious to untrained observers, but are completely discoverable through engineering analysis by qualified engineers. Checklists, FAA maintenance manuals, and certified mechanics provide the expertise needed to Sense β†’ Analyse β†’ Respond correctly.

The error would be treating this as Complex (running experiments to discover whether maintenance procedures work) or Clear (assuming mechanics can improvise). Both create serious safety risks.

Organisational Crisis Response​

A major data breach is discovered on a Friday evening (Chaotic domain). The immediate appropriate response is Act β†’ Sense β†’ Respond: contain the breach (isolate affected systems), communicate to relevant stakeholders, and stabilise the situation before conducting any analysis. Running RCAs and comprehensive reviews before stabilisation delays the essential containment action.

As the immediate crisis is contained (over hours to days), the situation moves from Chaotic toward Complex: understanding the full scope requires investigation, teams are adapting, and cause-and-effect relationships become clearer. The response approach shifts accordingly.

When to Use It​

βœ… Cynefin is most valuable for:

  • Leadership teams facing unfamiliar problem types
  • Diagnosing why past approaches failed (wrong domain classification)
  • Designing the right process for each type of work (not one-size-fits-all)
  • Managing transitions between domains (crisis β†’ stable β†’ growth)

❌ Less useful for:

  • Very clear, operational decisions that don't require a framework
  • Individual personal decisions (designed for organisational/leadership contexts)
  • As a bureaucratic classification step β€” it's a thinking tool, not a form to fill
Pairs well withWhy
Complex Adaptive SystemsCynefin's Complex domain is the realm of CAS dynamics
Minimum Viable TestMVT is the operational approach for Cynefin's Complex domain (Probe β†’ Sense β†’ Respond)
Scientific MethodScientific method applies in the Complicated domain; MVT probes apply in Complex
Pre-mortemPre-mortems help assess which domain a planned initiative is actually in

Common Misuses and Limitations​

The "expert trap" β€” treating Complex as Complicated. The most expensive error for organisations. Complex systems are genuinely unpredictable β€” no amount of expertise allows prediction before action. Organisations that hire McKinsey to "design" a culture change, "solve" an innovation challenge, or "determine" the winning product are often applying Complicated thinking to Complex problems. The consultants can provide frameworks and analysis; they cannot predict the emergent outcome of a complex social system.

Over-classifying as Complex. The opposite error: treating routine problems as Complex to avoid accountability for outcomes. "We ran experiments but the market is too complex to understand" can be an intellectually honest statement or a cover for poor execution.

Ignoring the cliff at the Clear/Chaotic border. Snowden specifically warns that Clear problems that are "over-simplified" β€” handled so routinely that no one questions the approach β€” can suddenly fall into Chaos when conditions change unexpectedly. The complacency born of long stability is a specific failure mode of overly Clear thinking.

Using it as a rigid classification. Domain boundaries are fuzzy; situations can be partly Complicated and partly Complex. The framework is a sense-making heuristic, not a taxonomic system.

ModelRelationship
Complex Adaptive SystemsCAS is the theoretical basis for Cynefin's Complex domain
Minimum Viable TestMVT operationalises Probe β†’ Sense β†’ Respond for the Complex domain
Scenario PlanningScenario planning is appropriate for Complex and Complicated strategic problems
Black Box ThinkingBlack Box Thinking applies most powerfully in the Complicated and Complex domains

Frequently Asked Questions​

What's the difference between Complicated and Complex?

In Complicated systems, the relationship between cause and effect is discoverable with sufficient expertise. An expert can study the system and determine the right answer. In Complex systems, no expert can predict outcomes in advance β€” the system's behaviour emerges from the interaction of many agents and can only be understood retrospectively. A machine is Complicated (an engineer can determine how it works); an ecosystem is Complex (interactions produce emergent behaviour that surprises even ecologists). Markets, organisations, and societies are typically Complex; engineering and medical diagnosis are typically Complicated.

How do you recognise when a situation has shifted domains?

Key signals: (1) Complicated β†’ Complex: the experts disagree, past solutions no longer work, small interventions produce disproportionate or unexpected effects. (2) Complex β†’ Complicated: patterns become clearer, cause-and-effect relationships become more predictable, best practices begin to emerge. (3) Clear β†’ Chaotic: what worked before suddenly doesn't, the situation deteriorates faster than expected, standard procedures become ineffective. Domain shifts often happen during organisational growth, market disruption, or crisis events.

How does Cynefin relate to Agile methodology?

Agile methodologies (Scrum, Kanban, XP) are explicitly designed for the Complex domain β€” short cycles, continuous feedback, adaptation over following a plan. Waterfall project management is designed for the Complicated domain β€” thorough upfront analysis, planned execution. The reason Agile succeeded in software development is that software development (especially new products) is genuinely Complex, and Agile's Probe β†’ Sense β†’ Respond rhythm matches the domain. Applying Agile to genuinely Complicated problems (like building a bridge) would be reckless; applying Waterfall to Complex problems (like a new mobile app) systematically fails.

Further Reading​

  • Snowden, D. & Boone, M. (2007). "A Leader's Framework for Decision Making." Harvard Business Review
  • Kurtz, C. & Snowden, D. (2003). "The New Dynamics of Strategy." IBM Systems Journal
  • Snowden, D. (2000). "Cynefin: A Sense of Time and Space" β€” the original framework paper

Apply with AI​

πŸš€ Classify your problem with Cynefin using MindMax β†’

This page is part of the MindMax Mental Models Knowledge Base.