I have sat in a lot of leadership meetings where a hard problem was raised and, within about four minutes, someone had the answer. Usually a reorganization. Sometimes a new process, a new hire or a new tool.
The speed is the tell. We read a situation through the way we prefer to act, and most of us, most of the time, reach for simplification. There is an urgent desire to fix.
Simplifying is right when the problem is simple. The damage comes from simplifying something complex without walking anyone through the reasoning.
Call a complex problem simple and you have already chosen the answer.
When a leader calls something simple, a solution comes with it. Sometimes it is written into the problem statement. Sometimes it hides in a KPI that rules out every other answer. Either way, the people about to do the work lose the part that was theirs: working out what the problem was.
The tool I reach for is Dave Snowden's Cynefin framework (say kuh-NEV-in), from 1999. Snowden calls it a sensemaking framework, not a categorization one, and the difference matters.
With most two-by-twos the framework comes first and the data is sorted into it. That is fast, and close to useless while things are changing. With Cynefin the data comes first, and the framework helps you see what you are looking at.
If you have run card sorts, it is the difference between a closed sort and an open one. A closed sort forces a good share of the cards into piles they do not belong in.
Clear. Cause and effect are obvious, stable and repeatable. Sense, categorize, respond, and use best practice. Do not reinvent the wheel. The stance that fights it treats every problem as a failure of process.
Complicated. Cause and effect exist, but finding them takes expertise, and there is more than one right answer. Sense, analyze, respond. You can reason your way to a decision with refined judgment. The stance that fights it asks for more time to analyze, every time.
Complex. Cause and effect are only clear looking back. Complexity is not how many parts a system has; it is how the parts relate. Probe, sense, respond. Run experiments that are safe to fail instead of designing something fail-safe, then amplify what works and dampen what does not. Good engineering teams already work this way.
Chaotic. There is no time to find cause and effect. Act, sense, respond: establish order, find where stability exists, then move the situation toward complex. Enter chaos on purpose and it is innovation. Arrive by accident and it is a crisis.
The edge between clear and chaotic
is not a border. It is a cliff.
Complacency walks you off it: the belief that everything is simple and last year's answer will hold.
Every other move between domains is gradual. That one is a fall, and the climb back is expensive.
The more common failure runs the other way, and slowly. A leader files a complex problem as clear, hands down the best practice, and the team builds it. Nothing breaks. The problem just does not move, and nobody can say why.
An uninformed, oversimplified answer to a complex problem is one of the most destructive things a senior leader can do to a team of builders. Not because the answer is wrong. Because of what it teaches them about whether to bother next time.
The framework cuts both ways. It will also tell you when you, or a colleague, are turning something obvious into something novel because novel is more interesting.
Before the answer, agree on the question.
It takes ten minutes. Ask everyone in the room to place the problem on the four domains, on their own, then compare.
Where people disagree is the useful part. An engineer who calls it complex and a CFO who calls it clear are not arguing about the answer yet. They are arguing about what kind of problem they have.
First published on Medium in January 2021, as part of a longer piece. Edited in 2026; the argument is unchanged.
Revisited, October 2026
The clearest case I have lived since: at Unbounce, churn was filed as a support number. Filed there, it could only get support answers. It was the ceiling on the company, and it moved once we renamed it. That story is here.
Five years on, it is still most of the work I do. The facts are rarely wrong. The heading usually is.