StrategyAdvanced7 min read

Cynefin Framework

Cynefin (pronounced 'kuh-NEV-in') is Dave Snowden's sense-making framework that sorts problems into five domains so you stop using the wrong playbook for the wrong situation. Clear (Obvious): cause-and-effect is known — sense, categorize, respond using best practice. Complicated: cause-and-effect requires expertise — sense, analyze, respond using good practice. Complex: cause-and-effect is only knowable in retrospect — probe, sense, respond with safe-to-fail experiments. Chaotic: no cause-and-effect at all — act first to stabilize, then sense, then respond. Disorder: you don't know which domain you're in (the most dangerous state). The framework's core claim: most strategy failures come from treating Complex problems as if they were Complicated — applying expert analysis to systems that can only be understood by acting on them.

Also known asCynefinSnowden FrameworkSense-Making FrameworkDomains of Decision

Challenge a friend Browse library

The Trap

The trap is reaching for 'best practice' when you're in a Complex domain. Best practices only work in Clear domains where the same input always produces the same output. Trying to plan a digital transformation by copying what McKinsey says worked at three other companies is exactly this mistake: the system is complex, the variables are entangled, and copying someone else's solution will fail because their context isn't yours. Equally bad: treating a Chaotic situation (a live outage, a PR crisis) as if it's Complicated — convening a strategy committee while the building burns. In Chaos, you act first.

What to Do

Before any major decision, classify the domain in one minute: (1) Do experts agree on the answer? Clear. (2) Do experts disagree but converge after analysis? Complicated. (3) Do experts disagree and the answer only emerges from trying? Complex. (4) Is the system actively breaking and there's no time to think? Chaotic. Then match your method to the domain — best practice for Clear, expert analysis for Complicated, parallel safe-to-fail probes for Complex, decisive stabilizing action for Chaotic. Stop pretending you're in a different domain than you actually are.

In Practice

Dave Snowden developed Cynefin while at IBM Global Services in 1999, originally to make sense of why knowledge management projects kept failing. His insight: companies were applying ordered-domain methods (process maps, expert systems) to problems that lived in the unordered Complex domain. The framework was formalized in a 2007 Harvard Business Review article ('A Leader's Framework for Decision Making') co-authored with Mary Boone, which became one of the most-cited HBR pieces of the 2000s. Snowden's lasting POV: 'You can't manage a complex system the way you manage a complicated one — and most management training teaches the wrong one.'

Pro Tips

01
The boundary between Clear and Chaotic is a cliff, not a slope. Things classified as 'obvious' (a routine deploy, a standard contract review) can fall into chaos with no warning when an edge case hits — and recovery is brutal because nobody was paying attention.
02
When in doubt, assume Complex. The default management instinct is to assume Complicated (hire a consultant, run an analysis). Defaulting to Complex forces you to run cheap experiments instead of expensive plans.
03
Disorder is where most exec meetings live. Half the room thinks the problem is Complicated, half thinks it's Complex, nobody names the disagreement — so the meeting produces a plan that satisfies no domain.

Myth vs Reality

Myth

“Cynefin is a 2x2 matrix you can put in a deck.”

Reality

Cynefin explicitly rejects the 2x2 framing. Snowden has said publicly that turning it into a quadrant chart 'destroys the entire point' — the boundaries are fuzzy, the central Disorder domain has no axis position, and the Clear/Chaotic boundary is a cliff. If your version of Cynefin fits in a 2x2, you're using something else.

Myth

“Complex problems can eventually be solved with enough analysis.”

Reality

Complex systems are emergent — the answer doesn't pre-exist your interaction with the system. More analysis on a Complex problem produces more theory, not more truth. The only way to learn is to probe (run small experiments), sense the response, and amplify what works.

Try it

Run the numbers.

Pressure-test the concept against your own knowledge — answer the challenge or try the live scenario.

🧪

Knowledge Check

Your CEO asks you to design a culture-change program for a 5,000-person organization after a merger. Which Cynefin domain are you in, and what's the right approach?

Decision scenario

The 'Best Practice' Trap

You're CEO of a 200-person B2B SaaS company. Sales growth has stalled at 12% YoY for three quarters. Your board hires a top-tier consulting firm that returns with a 90-page deck recommending the 'sales playbook used by three comparable companies that grew 40%+': hire a CRO, segment the sales team into hunters and farmers, switch to MEDDPICC, implement Gong. The deck is beautiful. The board loves it.

Revenue Growth

12% YoY (stalled)

Sales Headcount

Recommended Investment

$2.4M (consultants + tools + new CRO)

Cynefin Domain

Likely Complex (entangled product/sales/market)

Decision 1

Before authorizing the $2.4M, you ask: what domain is this problem in? The 'three comparable companies' that grew 40% had different products, different ICPs, different competitive environments. Their playbook worked because of their context — context you don't share. This has the fingerprints of a Complex problem being treated as Complicated.

Approve the full $2.4M plan — the consultants are experts, the comparables are real, and the board wants action.Reveal

Six months later: CRO hired, team resegmented, MEDDPICC training delivered, Gong installed. Growth stays at 11%. The new CRO blames the product. The product team blames sales. The consultants say 'execution issues.' You've spent $2.4M and a year of organizational disruption to learn that the imported playbook didn't fit your context — exactly the failure mode Cynefin predicts when you treat Complex as Complicated.

Cash Spent: $0 → $2.4MRevenue Growth (12 months later): 12% → 11%Sales Team Morale: Neutral → Low (resegmentation churn)

Reframe as Complex. Allocate $400K to run six parallel safe-to-fail probes for one quarter (e.g., a vertical-specific SDR pod, a product-led trial flow for SMB, a partner channel pilot, a usage-based pricing test for one segment, a CS-led expansion motion, a free-tools content engine). Kill the failures fast, double down on the winners.Reveal

Within 90 days, two probes show clear lift (vertical SDR pod and product-led SMB trial), three are flat, one is actively negative. You kill the four losers, scale the two winners, and run a second round of probes informed by what you learned. By month 9 you've grown 22% with $600K invested — not because the probes were genius but because the Complex domain rewards probe-sense-respond, not import-best-practice.

Cash Spent: $0 → $600KRevenue Growth (9 months later): 12% → 22%Organizational Learning: Zero → High (real data on what works)

Related concepts