Years ago I walked into a steering committee with the strongest recommendation I had ever built. Three regional operating models consolidated into one, the savings case, the risk position, the sequencing, 20 minutes of material with every objection answered before anyone could raise it.
The client’s operations director spoke for 4-5 minutes. He proposed the cautious option with a hardly tenth of my evidence, and the room agreed before even the coffee arrived.
I stayed behind to pack up, mostly so I didn’t have to share a lift with anyone. The programme sponsor stayed too.
“Your paper was better,” he said. “But Marcus spent the last two weeks having coffee with everyone in that room. You spent them with the data. He knew what the finance director needed to hear about headcount before she could nod. He knew the markets MD had promised his board there’d be no restructuring this year. What did you know about anyone at that table?”
I said something about the analysis speaking for itself.
“Nothing speaks for itself. Someone speaks for it, or nobody does.”
What I was calling politics
I had a name for what Marcus did, and the name was doing a lot of quiet work for me. Politics. The game around the work, played by people who couldn’t win on the work.
The label was very convenient because it turned a skill I didn’t have into a virtue I did. As long as pre-wiring a room counted as politics, my refusal to do it counted as integrity, and losing counted as being above the game.
"If you bring a finished answer to a room you haven't worked, you are the failure mode."
Here is the definition that survived that afternoon. A decision moves through people, and every person it touches carries stakes they have not declared. Reading those stakes is information work, the same as reading a contract or a cutover plan. Refuse it and you are not purer than the room. You are working with less information than everyone else in it.
Paul Nutt, a professor at Ohio State, spent two decades studying why organisational decisions fail. Half the decisions he tracked failed outright, and the failures piled up where somebody pushed a finished answer at people instead of working the decision through them. That finding usually gets read as advice for executives. Read it from our side of the table instead: if you bring a finished answer to a room you haven’t worked, you are the failure mode he was counting.
The question under the question
Every contested decision usually runs on two questions at once.
The stated one, which is on the agenda: which option is right.
And the operating one, which never appears anywhere: what will each person here have to explain, and to whom, if this goes each way.
The stated question is about the work. The operating one is about the people, and the operating one decides.
I told a shorter story on LinkedIn this week about being sent back for more analysis several times, by a room that turned out to be protecting a commitment nobody could mention. That is what the operating question looks like from the outside: it looks like the stated question, asked repeatedly.
You cannot hear the operating question directly, but it leaks.
A deferral that keeps finding new reasons.
A request for detail on ground the paper already covered.
The sponsor who agrees with you at his desk and goes quiet at the table.
None of those people are lying to you. They are answering a question you haven’t found yet.
"It looks like the stated question, asked repeatedly."
Reading a room before it decides
The method is unglamorous. Before any contested decision, go person by person through the room and ask what each of them has to explain if it goes ahead, and to whom. Their board, their boss, their own team, the colleague they overruled last quarter. Then ask what each of them has already promised that this decision touches. Prior commitments are where the unsayable lives, because reversing one in public costs standing, and people protect standing before they protect outcomes.
Then have the conversations the meeting can’t hold, one person at a time, before the meeting. Not to sell. To find out. Open with their problem, never with your decision: “What would need to be true about headcount for this to work from where you sit?” gets you the operating question.
“Can I walk you through the recommendation?” gets you a polite preview of the meeting.
The most senior people I’ve watched do this spend most of the conversation building the other person’s exit. If someone has to move off a position they’ve taken publicly, they need a story in which moving is sense rather than surrender. New information arrived. The context shifted. The phasing protects what they promised. Give someone a dignified road out and they will usually take it. Corner them with evidence and they will fight you with the full creativity of someone defending their standing.
"Corner someone with evidence and they will fight you with the full creativity of someone defending their standing."
The pre-read I run before any contested decision
One page, 6 questions, filled in before I book a single meeting. It has never once come back empty.
The decision as stated: the question on the agenda, in one line.
The explanations: for each person in the room, if this goes ahead, what do they have to explain, and to whom?
The prior commitments: what has anyone already promised, to a board, a budget round, their own team, that this decision touches?
The unsayable: what is true here that cannot be said in front of the others, and who is carrying it?
The conversations: who needs a private conversation before the meeting, in what order, and what does each need to be able to say in that room without losing standing?
The exits: for anyone who has to move, what is the story in which moving is sense rather than surrender?
Where this stops working
Three places, as far as I have learned.
The first is the line. Used to find the organisation’s best answer, this is reading. Used to engineer your answer past the organisation, it is manipulation, and people detect the difference with unnerving accuracy, usually only once. My test is simple: if the whole map were read aloud to the room, would I be comfortable? The moment the answer is no, I have crossed over.
The second is over-reading. Some rooms are genuinely undecided. If every deferral becomes a conspiracy, you will start negotiating with ghosts, and you will be exhausting to work with.
The third is the hardest. Sometimes the unsayable thing is about you. Your firm’s fees, your last delivery, the workstream your team is quietly failing on. A pre-read that never includes yourself is a comfort exercise. Some of the silence in your rooms is people protecting you.
The people deciding whether you become a transformation director are not re-marking your analysis. They are watching whether rooms move when you are in them, because a director is someone a partner can send into a client’s decision alone. Alone means nobody senior there to read the room for you, to have had the coffees for you, to build the exits for you. That is the actual test behind the title, and it is why some brilliant senior managers stall for years while less brilliant peers pass them: the stalled ones are still winning the stated question.
There is a quieter cost too. “I don’t do politics” functions as permission to skip the hardest skill in the job, and the people who grant themselves that permission usually discover the gate only after their third flat year.
A decision is settled before it is made. Execution prepares for the meeting. Leadership joins the settling.
Reply and tell me about the last decision that was settled before you walked in. The next issues get built from what comes back.
Roman