The kickoff is booked for ninety minutes and there are fourteen people in the room. Slide four is headed “Steering Committee”. Underneath sit the names: every direct report of the sponsor, two partners from the advisory firm, a finance manager whose presence nobody explains. The slide holds the screen for about forty seconds, then the deck moves on to milestones.
Nobody asks what the committee is for.
The question would sound like an admission of ignorance, and this is the wrong room for that. Everyone present has sat on something called a steering committee before, so the term feels settled. It is not settled. It is inherited, and it arrives with an invitation list attached rather than a definition.
Ask those fourteen privately what they are meant to do at the second meeting. Some will say they are there to stay informed. Others will say they represent their function’s interests. One or two will say they are there to unblock things, without being able to name what they are permitted to unblock. Nobody says they are there to decide, because nobody has told them which decisions belong to the room.
They will not ask, either. They will work it out by behaving — and the way a senior manager demonstrates contribution in a meeting with no defined role is by having an opinion about whatever is on screen.
A steering committee is defined by what it can decide. Most are defined by who was invited.
Scope is the one thing a steering committee does not steer
The verb deserves to be taken literally. Steering implies a vessel that would otherwise travel somewhere you did not intend, and a hand able to change that. So the fair question, before any invitation goes out, is what is actually being steered.
It is not the scope. The scope was agreed, funded and baselined. If it is genuinely open to redirection every six weeks by a committee, the organisation has a problem that no governance design will fix.
What needs steering sits outside the project’s boundary. Priority, against everything else competing for the same twelve people. Resources actually released, as opposed to resources named in the business case. Tolerances — how far the work may drift before someone must intervene. The benefit commitment, and whether it is still worth the remaining spend. Continuation at all.
A project manager can report on every one of those. They can decide none of them. That gap is the entire reason a steering body exists. Everything else the meeting does could be done by email.
A sponsor and a steering committee solve different problems
PMI’s programme standard, in its fifth edition, keeps three roles carefully apart: the sponsor who owns the business case and secures resources, the programme manager who runs delivery within a mandate, and the programme office that supplies the governance infrastructure. Organisations compress all three routinely, then build a committee on top to hold whatever was squeezed out.
Before agreeing to chair one, name a decision the committee can take that the sponsor cannot take alone. If you cannot name one, you do not need a committee. You need thirty minutes in the sponsor’s diary at a predictable interval, which is cheaper and faster.
A committee earns its existence where the decision requires authority that no single executive holds — where three functions must each give something up and none of them reports to the others. That situation is real, common, and worth a room.
The failure mode runs the other way. Some committees exist so that sponsorship can be spread across seven people until no individual carries it. The programme then has seven supporters and no owner, and every escalation lands somewhere that must consult before it can answer. The delay is recorded later as a delivery problem. It is the same gap that turns a well-argued strategy into an unbudgeted intention.
A body that cannot stop the work is a reporting audience
The first test I apply is blunt.
Can this room halt, hold or re-sequence the work?
Where it cannot, it is not steering anything — it is an audience with a recurring calendar entry.
Four rules separate the two, and they have to hold in practice rather than in a terms-of-reference document nobody has opened since month one. Papers circulate beforehand and are never read aloud. Each item arrives as a proposal with an owner and at least two options, and leaves with a decision, a name and a date recorded against it. Decisions taken are not reopened at the following sitting. Escalations carry a deadline — in my own contracts the client side commits to answering within five working days.
That last rule matters more than it looks. An unbounded decision latency is functionally identical to no governance at all. Work does not pause while a room deliberates; it proceeds on assumptions, and those assumptions get made by whoever is closest to the next deadline.
Which decisions belong to the committee, which to the sponsor alone, and which are reserved to the owners and will never be delegated — all of that belongs in writing before the first meeting, not after the third escalation. I use a decision authority table for it: one row per decision area, a level per row, a limit per level. The professional standards are strong on the composition of governance bodies and almost silent on what happens when one declines to decide. That silence is where most transformation programmes actually live, and closing it is what a charter is for.
A deck sent afterwards as a PDF is not a report
Here is the complaint I hear most often from executives about their own programmes, usually delivered with slight embarrassment. Between meetings there is no way to find out where anything stands. You wait for the next steering committee, and a few days later somebody emails round the presentation as a PDF.
That is not a reporting inconvenience. It is an accurate description of the organisation’s reporting capability, and it costs in three ways.
Currency is the first. A deck assembled by hand over three days, from what stream leads say about their own streams, is out of date the moment it is shown — and it was written for the reaction in the room rather than for a decision.
Format is the second. A PDF cannot be queried, sorted, or traced back to the number underneath it. It records a performance rather than a state. The flatness is a feature of the artefact, not an oversight: you cannot check what you cannot open.
Rhythm is the third, and the most expensive. When the picture exists only on meeting day, the decision cadence becomes chained to the meeting cadence. A question that needed resolving in week two waits until week six, because week six is when the organisation can next see itself. The one hour of assembled authority is then spent transferring information, so decisions slide to the end of the agenda, where people are already standing up.
There is a fourth consequence, and it explains the seating plan. If attendance is the only reliable channel of information, every sensible director wants to be in the room. The committee does not swell because people are political. It swells because presence is the only way to know anything.
The remedy is not more reporting. It is a small set of facts that are continuously true, owned by named people, which the meeting refers to instead of recreating. The signal I trust here is simple: reporting volume rising while decision throughput stays flat.
A seat without a decision right turns into an opinion
The invitation list grows for reasons that are individually reasonable. Excluding one direct report while including the other five is a political act, and few sponsors have appetite for it on day one. Finance should probably see this. The advisory firm would like its partner in the room. Each addition costs nothing visible.
The cost arrives later and it lands on the people who were added. A senior manager given a seat and no decision right will not sit quietly for an hour every six weeks. They contribute in the only form the seat permits, which is commentary — so the meeting fills with well-informed views on matters that were not in question, offered by people demonstrating that their presence was warranted.
The second-order effect is worse. Participants who never received a brief will write themselves one. They start acting on their own reading of the programme inside their own function, and a transformation acquires six parallel interpretations of the same decision. Nobody misbehaved here. They were put in a room, given no instructions, and did the sensible thing.
Fourteen people can be informed. They cannot decide. Any group past roughly seven stops resolving things in the room and starts resolving them in the corridor afterwards, among the three who hold the authority — which is a functioning governance model, just not the one on slide four.
Every name on the list should survive one question
The question is: what happens if this person says no?
Where the answer is that a resource is not released, a system is not changed, a budget is not moved or a date is not accepted, you have a decision member. Where the answer is nothing in particular, you have something else — and being honest about which is a kindness to them as much as to the programme.
Four roles are enough. A chair, normally the sponsor, who closes items and owns the outcome. Decision members who can commit their function and live with the consequences, three to five of them and rarely more. The person who brings the decisions, usually the programme director, who prepares, presents and records, and who is not a member. Advisers pulled in for a single item and released afterwards.
Observer is not a role. If someone needs the information, send them the information — that is a distribution problem, and solving it with a chair at a governance table is an expensive way to run a mailing list. And if the honest answer is that they attend because there is no other way to find out anything, the fix belongs in the previous section rather than in the seating plan.
The advisory firm’s steering committee is a contract instrument
When a consulting firm arrives, a steering committee arrives with it, and the organisation accepts this as part of the method. It usually is part of a method. Theirs.
A committee established to run an engagement does a legitimate job. It secures acceptance of deliverables, keeps the scope contract alive, and distributes agreement across enough senior names that the work can proceed unchallenged. None of that is improper. It is simply not the same organ as a body that governs your business.
Two questions separate them.
Will this committee still be taking decisions six months after the final invoice?
And who chairs it, who writes the pack, and who records what was decided?
Where all three answers name the firm, you are attending your own governance as a guest. The awkward consequence is that the record of what your organisation decided walks out of the door when they do.
You are not obliged to accept the structure that came in the proposal. A perfectly workable arrangement is that your governance body decides, on your cadence, in your record, and the firm attends the items where its own work is under discussion. That choice sits closer to the distinction between buying advice and delegating a mandate than most buyers realise at signature.
Eight questions settle this faster than any governance review
None of these requires preparation, and I ask them in the first fortnight of every engagement. The answers usually arrive in under a minute, or not at all — which is itself the finding.
- What can this body decide that the programme director cannot?
- What can it decide that the sponsor cannot decide alone?
- Can it stop the work — halt, hold, re-sequence — or only comment on it?
- How long may an escalation wait before an answer is owed?
- When did it last take a decision that changed what happens next week?
- Whose budget or headcount actually moves when the room says yes?
- How would you learn tomorrow morning what state the programme is in, without asking a person?
- Who is in the room because they decide, and who because leaving them out would have caused offence?
A committee that cannot say no is not steering. It is watching, at intervals, in PDF.