Problem solving Explore

Cynefin Framework

The Cynefin framework sorts a situation into Clear, Complicated, Complex, or Chaotic, and changes how you respond.

The questionIn this situation, right now, what is the relationship between cause and effect?

Pronounced kuh-NEV-in. Welsh for habitat.

Model · Dave Snowden, Late 1990s

Clear Sense, categorize, respond
Complicated Sense, analyze, respond
Complex Probe, sense, respond
Chaotic Act, sense, respond
Confusion Not sure yet. Split the question into smaller pieces.

Kinds of situation, not levels. One engagement usually contains more than one.

When to use it

A team is arguing about method before it has agreed what kind of situation it is in.
Someone wants a standard response in a place where cause and effect are not stable.
You are choosing between a standard procedure, expert analysis, a small experiment, and immediate action.

When to avoid it

The domain is already settled and you need a tool inside it, such as a process map, a role chart, or a pricing model.
The room wants a label for a person or a team, rather than a way to decide about a situation.
Placing the decision would take longer than making it.

Overview

Cynefin is a way of choosing how to respond to a situation. First ask how cause and effect relate in it. Then pick the kind of response that fits: apply a standard, call in experts, run a small experiment, or act to stabilize things. The common mistake is using one of those responses everywhere.

It sorts situations, not people. The same executive can face a Clear payroll question at nine and a Complex integration question at two. Domains also shift. A Complex question can become Complicated as you learn more, and a Clear process can break when something it relied on changes.

The domains

Clear

Cause and effect are obvious to anyone who looks. Sense the situation, categorize it, and apply the standard. This is where best practice lives. The risk is complacency. The standard keeps running after the situation has changed. Earlier versions called this domain Simple, and then Obvious.

Complicated

Cause and effect exist but take expertise to see. Sense, analyze, respond. More than one sound answer is usually available, so this is where good practice lives, not a single best one. The risk is experts stalling each other, or an analysis that never ends because nobody wants to be the one who chose.

Complex

Cause and effect can only be seen in hindsight. Run small experiments you can afford to stop, called safe-to-fail probes. Sense what happened, then amplify what works and dampen what does not. The practice is emergent. You find it by trying, not by analyzing. The risks are demanding a forecast the situation cannot give, and copying a standard from somewhere the constraints were different.

Chaotic

Constraints have broken down, and searching for a cause will not stabilize anything. Act to create some order, sense where things stand, then respond. The practice is novel, because nobody has a playbook for this exact situation. The risk is staying in command mode after the crisis has passed.

Confusion

You do not know which domain you are in. Older writing called this Disorder. People in this state fall back on habit. The expert analyzes, the operator issues a rule, the innovator launches a pilot, and each reads the whole situation through the domain they prefer. Split the situation into pieces and place each one. Most real engagements are several domains at once.

What it is not

A two-by-two. Plotting it on agreement and certainty gives you the Stacey Matrix, which is a different framework. Cynefin is about how cause and effect behave.
A maturity scale. Clear is not a higher grade than Complex, and a situation can move in either direction.
A type for people, teams, or cultures. The same people face different domains on the same day.

The steps

Name the decision

Bound it in time and scope. "The company is complex" cannot be placed. "How these two sales forces cover the top twenty accounts this quarter" can.
Ask how cause and effect behave

Would a newcomer see the cause? Only an expert? Only in hindsight? Or does nothing hold still long enough to tell? Those answers point to Clear, Complicated, Complex, and Chaotic. If nobody can answer, you are in Confusion. Split the question into smaller pieces.
Place the pieces

Put each piece in one domain. Do not label the whole program. Write down each placement and the evidence that would change it.
Match the response

Clear, apply the standard. Complicated, have people who have seen this before analyze it, then choose among the sound options. Complex, run probes that are safe to fail, and decide in advance what you will amplify and what you will drop. Chaotic, act to restore enough order to see what is happening.
Read the shift

Set a date to place the work again. A probe that produces a stable pattern may mean a piece is moving toward Complicated. A Clear process that starts throwing exceptions may be leaving Clear. A placement describes the situation today. It is not a permanent label.

In practice

A leadership team wants the acquired company run on the acquirer's playbook. Payroll, benefits administration, and the chart of accounts are Clear, and a standard is the right tool. The acquired product architecture is Complicated. It needs the few people who know that system, and more than one design would work. How the two sales forces behave over the next two quarters is Complex. A mandated target culture will not settle it. Run two or three small changes to account coverage, incentives, and who speaks to the key accounts, and agree in advance what would make you stop each one.

Most of the argument in the room is about the playbook itself. The problem is that a Clear-domain tool is being applied to a Complex-domain question.

What you leave with

A decision stated narrowly enough to place, in time and scope.
Each piece of that decision placed in a domain.
A response for each placement: a standard, an analysis that ends in a choice, a probe with a stop rule, or an action that restores order.
The evidence that would move a piece to another domain, and a date to check again.

Failure points

Best practice everywhere

A standard that works in a Clear domain is imposed on a Complex one. When people resist, it gets called a change-management problem. The diagnosis was wrong.

Analysis as a place to hide

More data and more experts help in the Complicated domain. Applied to a Complex situation, they produce a better model of the past and no decision.

Chaos habits in ordinary work

Acting first because it looks decisive, when a short, serious analysis would have prevented the damage.

One label for the enterprise

"We are a Complex organization" is a slogan. Nothing changes, because nobody placed a specific decision.

Probes that are not safe to fail

A probe has to be small enough to stop. If it is public, expensive, or tied to a commitment, nobody will stop it, and it is a bet.

The failures above are usually Framing: the wrong problem was chosen before this framework was applied.

Origin

Dave Snowden developed Cynefin in the late 1990s at IBM Global Services, where he worked on knowledge management and complexity. The November 2007 Harvard Business Review article he wrote with Mary E. Boone, "A Leader's Framework for Decision Making," was a cover story and brought the framework to a general management audience. It uses the older names Simple and Disorder.

Since then Simple has become Obvious and then Clear, and Disorder is usually called Confusion. The four responses have not changed. The terms still move, so check the Cynefin Co site before you teach it.