How to Run a Steering Committee That Decides Things
The test of a steering committee is simple: count the decisions it took last quarter. If the answer is none, it is a reporting meeting with an expensive attendance list.
Steering committees drift into reporting because reporting is comfortable and deciding is not. The agenda fills with status, the papers arrive the night before, and the people with authority spend an hour receiving information they could have read. Meanwhile the decisions the engagement actually needs are taken in corridors by whoever is available.
Fixing it is mostly structural. The membership, the papers and the log determine what the meeting does far more than the chair's intentions.
Membership: authority, not representation
The only necessary members are the people who can commit resources, resolve a dispute between functions, and accept a deliverable on behalf of the organisation. Everyone else attends by invitation for the item that concerns them.
The commonest failure is representation: each affected function sends someone, the committee reaches fourteen people, and nobody in the room can commit anything without consulting elsewhere. A committee of five who can decide beats a committee of fourteen who can report. If a function genuinely needs a voice, give it a paper rather than a seat.
Cadence matched to the decision rate
Cadence should follow the rate at which decisions are actually needed, not a default monthly rhythm. A mobilisation phase may need fortnightly meetings; a long build phase may need one every six weeks with an escalation route in between.
The escalation route is the part that makes a slower cadence safe. If the only way to get a decision is to wait for the meeting, the meeting will be too frequent and still too slow. Agree in advance who can decide between meetings, up to what value, and how that decision is recorded so the committee sees it at the next sitting.
Papers that ask for something
Every paper should state, in its first three lines, what decision is sought and what happens if it is not taken. A paper with no recommendation is a briefing, and briefings should be circulated rather than tabled.
Circulate papers far enough ahead that reading them is possible, and open each item by asking whether anyone needs the summary rather than presenting it. That single change typically recovers half the meeting, and the recovered time goes to the discussion that actually needs the people in the room.
- The decision sought, stated as a question with a recommended answer.
- The options considered, including doing nothing, with the cost of each.
- The consequence of deferral, stated in dates and money.
- Who will implement the decision, and by when.
The decision log
A decision log is separate from minutes and more valuable. Minutes record what was discussed; the log records what was decided, by whom, on what date, on what basis, and what it changed. Six months later, when someone asks why the scope was narrowed, the log answers in one line.
It also protects the committee. A decision taken on the information available at the time is defensible even when it turns out badly, provided the basis was recorded. A decision with no recorded basis is indistinguishable from carelessness in hindsight.
Reviewing the committee itself
Once a quarter, ask three questions. How many decisions did we take, and could any of them have been taken elsewhere? How many items came back a second time because the first discussion did not conclude? Is anyone attending who has not spoken in three meetings?
The answers usually suggest a smaller membership, a slower cadence and a firmer expectation that items arrive with a recommendation. Very few governance problems are solved by adding a meeting.