Most round tables produce a summary nobody acts on, because the format rewards agreement. A session that opens by asking each participant for their strongest objection to a written proposition produces a different output: a list of the specific conditions under which the plan fails.
What format actually produces decisions?
One that starts from a written proposition rather than an agenda. An agenda invites discussion of topics; a proposition invites agreement or objection, and only the second produces something you can act on.
The proposition should be circulated far enough ahead that participants arrive having formed a view, which is the difference between a session that tests a plan and one that explains it — the same reason an outside operator is briefed before the offsite rather than during it.
- 01
Circulate a proposition
Send a written claim, not an agenda, far enough ahead for people to form a view.
- 02
Cap the room
Eight at most, so speaking time supports an argument rather than a statement.
- 03
Objections first
Go round the table asking for the strongest objection before asking for views.
- 04
Close on conditions
Record the conditions under which the plan fails, with names against each.
How many people should attend?
Eight at most. Above that, speaking time per participant falls below the point where anyone develops an argument, and the session becomes a sequence of opening statements delivered to a room that has stopped listening.
The same arithmetic applies to expert panels, and for the same reason: the value is in the exchange rather than in the coverage.
What should the chair do?
Protect disagreement. The default social pressure in a senior room runs toward consensus, and a chair who manages the discussion smoothly will produce a smooth discussion with no output. A contradiction left standing is worth more than an average.
Practically that means keeping an objection alive long enough for it to be tested rather than absorbed, and being willing to return to it after the room has moved on.
A chair also has to be willing to be the least popular person in the room for a few minutes at a time, which is why the role works better when it is held by someone who does not have to work with the participants afterward.
| Panel-style | Structured objection round | |
|---|---|---|
| Opens with | An agenda and introductions | A written proposition and a round of objections |
| Social default | Toward consensus | Toward disagreement, which is the point |
| Output | A narrative summary | Conditions under which the plan fails, with names |
How do you surface disagreement?
Ask for objections before you ask for views. Once someone has stated a position, defending it becomes the game; asking for the objection first makes disagreement the expected contribution rather than a social cost.
Going round the table in a fixed order for that first pass also stops the most senior voice from setting the frame.
What is the output document?
A list of conditions under which the proposition fails, with names against each. That is actionable: each condition can be tested, and the person who raised it has a stake in whether it holds — which is what a planning cycle can carry forward.
A narrative summary of the discussion is a record rather than an output, and nobody reads it twice.
The document also has to be circulated quickly enough to be believed. A list of conditions that arrives a fortnight later reads as a record of a meeting; one that arrives the next morning, while everyone still recognizes their own objection, reads as something they agreed to. Speed of return does more for adoption than polish does.