Home / Complexity and uncertainty / Cynefin Framework

Complexity and uncertainty

Cynefin Framework

A sense-making model that matches your response to the kind of situation you're in

Also known as: Cynefin, The Snowden framework

Originated by Dave Snowden

THE IDEA

Different problems need different thinking

Not all situations are the same kind of problem, and they don’t all respond to the same kind of solution. This sounds obvious. In practice, most people and organisations have one mode - usually analysis and planning - and they apply it to everything regardless.

The Cynefin framework (pronounced “kuh-NEV-in,” a Welsh word meaning habitat or place) sorts situations into domains based on the relationship between cause and effect. In clear situations, cause and effect are obvious to everyone - sense, categorise, respond. Best practice works here. In complicated situations, cause and effect are discoverable but require expertise - sense, analyse, respond. Good practice works here. In complex situations, cause and effect are only visible in hindsight - probe, sense, respond. Emergent practice works here. In chaotic situations, there’s no discernible cause and effect - act, sense, respond. Novel practice is needed.

There’s a fifth space: confused, which is where you are when you don’t know which domain you’re in. The first job is always to get out of confused.

The framework’s power isn’t the categories themselves - it’s the discipline of matching your approach to the situation. An expert analysis is the right response to a complicated problem and a dangerous one in a complex problem. A rapid experiment is the right response to a complex problem and a waste of time in a clear one. Cynefin gives you permission to respond differently to different kinds of difficulty.

IN PRACTICE

Choosing the right gear

A factory production line stops. The cause is a known failure mode - a specific component overheats after a certain number of hours. This is a clear domain problem. The response is standard: replace the component, adjust the maintenance schedule. No analysis needed, no experiments. Follow the known procedure. Treating this as complex (running experiments to understand why the line stopped) would be absurd.

A patient presents with unusual symptoms that don’t match standard diagnostic categories. This is a complicated domain problem. The GP refers to a specialist, who runs tests, consults the literature, and eventually identifies a rare condition. The right response was expert analysis. Treating this as clear (following the standard symptom checklist) would have missed the diagnosis. Treating it as complex (trying treatments to see what happens) would have been reckless.

A company enters a new market in a country it’s never operated in. Customer behaviour is unfamiliar, regulations are different, competitors are unknown. This is a complex domain problem. No amount of analysis will produce a reliable plan because the system is too interconnected and adaptive. The right response is to run small experiments - a limited launch, a pilot partnership, a test campaign - and learn from what happens before committing. Treating this as complicated (hiring an expert to produce the definitive market entry strategy) would produce a beautiful plan disconnected from reality.

WORKING WITH THIS

Starting with “where am I?”

The first move in any difficult situation is to ask: what domain is this? Don’t start with “what should we do?” Start with “what kind of problem is this?” The answer shapes everything that follows.

Clear: apply best practice. Don’t overthink it. The risk here is complacency - treating everything as clear because you’ve seen something vaguely similar before.

Complicated: find expertise. Analyse. The risk is over-analysis - getting stuck in study when the situation has shifted from complicated to complex while you were researching.

Complex: run safe-to-fail experiments. Probe. Watch for patterns. Amplify what works, dampen what doesn’t. The risk is impatience - wanting an answer before the system has had time to reveal one.

Chaotic: act first. Establish some kind of order, then figure out where you are. The risk is freezing - waiting for information when the priority is stabilisation.

The biggest danger is boundary problems - when a situation shifts from one domain to another and you don’t notice. A complicated situation can tip into chaos if a key assumption breaks. A clear situation can become complex when human behaviour gets involved. Cynefin isn’t a one-time diagnosis. It’s an ongoing question: is this still the kind of problem I think it is?

THE INSIGHT

The line to remember

The most expensive mistake isn’t getting the wrong answer. It’s applying the right answer to the wrong kind of problem.

RECOGNITION

When this is in play

You’re in Cynefin territory when a proven approach stops working and nobody can explain why - the domain may have shifted. When a team is paralysed by analysis in a situation that needs experimentation. When someone demands certainty about something that’s fundamentally uncertain. When the response to every problem is the same regardless of its nature. When the experts disagree and somebody says “we just need more data” - that’s often a complicated-domain response to a complex-domain problem.

Connected concepts

Complexity vs Complication

Cynefin builds directly on the distinction between complicated and complex - and adds more domains

Emergence

In Cynefin's complex domain, outcomes emerge from interactions rather than being designed

Feedback loops

Each Cynefin domain requires different feedback strategies - tight loops in complex, analytical loops in complicated

Leverage Points

Cynefin helps you find leverage by first identifying what kind of system you're dealing with

Policy Resistance

Policy resistance often signals you've misidentified the domain - treating a complex problem with a complicated approach

Wicked Problems

Wicked problems are the signature challenge of Cynefin's complex domain - requiring probes, not plans

Complex Adaptive Systems

Cynefin's complex domain describes the territory where complex adaptive systems operate

Uncertainty vs Risk

Cynefin's complicated domain involves manageable risk. Its complex and chaotic domains involve genuine uncertainty

Safe-to-Fail Experiments

Safe-to-fail experiments are the primary intervention strategy for Cynefin's complex domain - probe, sense, respond

Sensemaking

Cynefin is fundamentally a sensemaking framework - its first job is helping you understand what kind of situation you're in

Sensemaking Under Pressure

Under pressure, people default to familiar domain responses - the risk is applying the wrong domain's approach

Probe-Sense-Respond

Probe-sense-respond is the prescribed strategy for Cynefin's complex domain

Sense-Analyse-Respond

Sense-analyse-respond is the prescribed strategy for Cynefin's complicated domain

decision-making sense-making complexity strategy