On working a complex problem — patient observation, and the discipline of not arriving with the answer.
Most consulting arrives with the answer. It comes through the door carrying a framework, a benchmark, a shape borrowed from somewhere else, and the engagement becomes the work of pressing that borrowed shape onto the situation until something gives. I have come to distrust this almost entirely — not because frameworks are worthless, they are often useful and I use them, but because a genuinely complex problem, the kind that involves people, rarely yields to a solution imported from outside it. The answer to a stuck problem is usually already present: in the situation itself, or in the way the people closest to it describe what is wrong. The work is not to invent it. The work is to find it, and to make it usable.
That belief shapes how I begin, which is by doing very little.
The first discipline is to observe, and to listen, for longer than is comfortable. Most problems are misdiagnosed at the moment they are framed — the presenting complaint is rarely the real one, and the story an organisation tells about itself is not how it behaves. So the early work is patient and largely quiet: watching how decisions actually get made rather than how the structure says they should, noticing who speaks and who goes silent, hearing what is said and, more revealingly, what is carefully not said. Active listening is not waiting for your turn. It is attending closely enough that the other person says the thing they did not quite mean to — the thing underneath — which is usually where the problem really lives. The hardest part is resisting the reflex to arrive with the answer, because the answer that arrives early is almost always the answer to the wrong problem.
Out of that comes the first real task: to separate the problem as presented from the problem as it actually is, and then to separate the two kinds of difficulty inside it — the complicated parts, which a competent person can simply work through, and the complex, human core, where the difficulty truly sits. Most stuck situations are a tractable technical problem wrapped around an untended human one, and almost all of the wasted effort goes into solving the wrapper. Naming the core correctly is often most of the value; a surprising amount of what follows is straightforward once the right problem is on the table.
Then the work is to shape — to turn what has surfaced into a solution, and, just as importantly, into the specific decisions it will demand of the people who have to live with it. This is where importing an answer fails and extracting one succeeds, because a solution the client recognises as already theirs will survive contact with the organisation, and one that arrives as a stranger’s will not, however elegant it is. Much of shaping is making legible what was already half-known in the room — giving people language for something they could feel but not articulate — so that the solution belongs to them rather than to me. Ownership is not a nicety here. In a complex problem it is the difference between something that holds and something quietly undone the moment attention moves elsewhere.
And then it has to be done, which is the part that most cleanly separates advice from consequence. I do not hand off at the point of execution, because in complex work the seam between deciding and doing is exactly where things fail: the plan that was sound in the room meets the people it did not account for, and needs adjusting by someone who understands why it was the plan in the first place. So the method scales to the problem rather than to a fixed model — sometimes a single conversation with the person who can act, sometimes standing alongside them and doing the detailed work myself, sometimes defining and fielding a team. What stays constant is that the thinking and the doing remain joined, held together by the same understanding of the people that the observation produced at the start.
The truest test comes late, not early — not at the blank page, but after commitment, when a course is set, money and reputation are on the line, and something is going wrong. This is the moment most engagements are not built for and most people handle badly, because correcting a committed course is not a technical act but a human one. It has to be done deftly: surfacing the problem without assigning blame, changing direction without detonating the confidence and the relationships the original decision was built on, so that the cost of the correction never exceeds the cost of the mistake it fixes. This is the rarest part of the craft, and it is only possible because of everything upstream — because you observed carefully enough, and earned enough trust and enough shared understanding, to be allowed to change the plan when the plan needs changing.
None of this is a process, and it does not run in a straight line; observation continues right through delivery, and a correction sends you back to listening. But the spine is constant, and it follows directly from what a complex problem actually is. If the complexity is human, then the way through it is human too: watch, listen, find the answer that is already there, and stay close enough to the people to see it through — and to put it right when it wanders. Everything else is technique.
Further reading: Complicated Is Not Complex · The Wrong Tool for the Right Job