Cynefin Framework Questionnaire

To use Cynefin as a diagnostic, use this questionnaire. Unlike the others here, it does not score a single scale: it asks a series of questions whose answers place a specific situation in a domain, and then prescribes the response appropriate to that domain.

Run it on one named situation at a time. A department is not in a domain; a particular problem is.

See cynefin-framework.

Evidence: Weak empirically, high conceptual utility. Domain assignment is a judgment and two competent people can disagree. The disagreement is usually the productive part.

Step 0: Name the situation

State one specific situation, problem, or decision.

Situation: ____________________________________________________

Who is answering: _____________________________________________

Have several people answer independently. Divergence is a finding.

Step 1: Domain placement

Score each statement 1 (strongly disagree) to 5 (strongly agree).

Clear indicators

1. The cause-and-effect relationship here is obvious to any competent observer. [ ]
2. There is an established best practice, and everyone would name the same one. [ ]
3. If we simply categorized the situation correctly, the right action would follow automatically. [ ]

Complicated indicators

4. There is a right answer, and an expert could find it with enough analysis and data. [ ]
5. Several defensible approaches exist, and choosing between them requires expertise rather than experimentation. [ ]
6. We could write a specification for the solution before building it. [ ]

Complex indicators

7. The situation responds to our interventions, so what worked last time may not work now. [ ]
8. We can explain why things happened afterwards, but could not have predicted them beforehand. [ ]
9. Competent, well-informed people genuinely disagree about what will happen if we act. [ ]
10. Any business case we write will be a guess dressed as a forecast. [ ]

Chaotic indicators

11. The situation is actively deteriorating and no stable pattern is discernible. [ ]
12. Doing nothing while we analyze is the worst available option. [ ]

Confused indicators

13. Different people in the room are confidently applying incompatible methods. [ ]
14. We are arguing about how to proceed more than about what to do. [ ]

Step 2: Read the placement

• Highest scores on 1–3 → Clear. Sense, categorize, respond. Apply best practice. Watch for complacency, and for the cliff into Chaotic when conditions change and nobody notices.

• Highest scores on 4–6 → Complicated. Sense, analyze, respond. Use experts, apply good practice, accept there are several valid answers. Watch for expert overconfidence and entrained thinking.

• Highest scores on 7–10 → Complex. Probe, sense, respond. Run parallel safe-to-fail experiments; amplify what works, dampen what does not. Watch for the demand for a guaranteed outcome, which forces the problem to be treated as Complicated.

• Highest scores on 11–12 → Chaotic. Act, sense, respond. Stabilize first, then move the situation into Complex. Watch for anyone enjoying the authority that crisis grants.

• High scores on 13–14 regardless of the rest → Confused. Break the situation into parts and place each part separately. This is the most common real state, and naming it is the intervention.

Step 3: The mismatch check

The most valuable question in the set.

• Question: Which domain does our current management approach assume? (Detailed multi-year plan and business case = Complicated. Best-practice rollout = Clear. Portfolio of experiments = Complex. Command and stabilize = Chaotic.)

  • Answer:

• Question: Does that match the domain we placed the situation in?

  • Rating (1–5, where 5 = matches well): [ ]

A low score here is the finding. The characteristic organizational error is managing a Complex problem as though it were Complicated: a detailed plan, a business case promising a specific number, and a consultant selling a known answer. Culture change, leadership behavior, and most organizational change sit in Complex and are almost always managed as Complicated.

Step 4: If Complex — design safe-to-fail experiments

For each proposed experiment, all five fields must be filled in before starting.

• Question: Are these experiments genuinely parallel and genuinely different from each other, or are they three versions of the same idea?

  • Rating (1–5): [ ]

• Question: Is failure of any one of them survivable and non-newsworthy?

  • Rating (1–5): [ ]

Step 5: Analysis

• Domains are situational and they move. Re-place the situation when conditions change. A stable Clear process becomes Chaotic the moment its assumptions break, and that transition is the dangerous one because nobody is watching for it.

• Mixed programs need mixed management. A merger's legal and financial integration is Complicated; its cultural integration is Complex. Running both under one plan with one set of milestones guarantees the cultural half is judged by the wrong criteria and declared complete when it is not.

• Resist the demand to convert Complex into Complicated for governance convenience. The demand is understandable — boards want forecasts — and meeting it produces a plan that cannot adapt. The honest response is a portfolio with a stated learning cadence and a defined spend, not a fabricated forecast.

• Do not use this as a permanent classification. Cynefin is a sense-making aid for a conversation, not a taxonomy for a filing system.