All three are legitimate. Each is correct in conditions the others are not, and a leader who only has one of them will be wrong roughly two thirds of the time.
The failure this entry is actually about is not choosing badly among them. It is running one mode while signalling another, which happens constantly and does more damage than any of the three used honestly.
The three
Command. One person decides and announces. No input is sought, or input is sought and the person deciding owes nobody an account of how it was weighed. Fast, clear, and it spends standing every time it is used.
Consultation. Views are gathered genuinely, and one named person then decides. The people consulted have a right to be heard and no right to prevail. This is the mode most decisions should use and the one most often faked.
Consensus. The group works until everyone can live with the outcome. Slow, expensive in hours, and it produces something the others cannot: an entire group that owns the result and will not relitigate it privately.
When each is right
Command belongs to genuine time pressure, to situations where the information sits with one person, and to decisions nobody else should have to carry. A leader who commands routinely is either in a crisis or has confused decisiveness with the appearance of it. A leader who never commands has left their organization without a mechanism for the moments that require one.
Consultation is the default for most consequential decisions, because it gathers the information that only exists in other heads while keeping the accountability in one place. It also produces a decision the organization can live with, since being heard, even when overruled, is most of what people need in order to accept an outcome.
Consensus earns its cost in a narrow band: decisions that require sustained voluntary effort from people who could quietly decline to give it. A change to how a professional team works, a partnership among peers, a volunteer board setting its own direction. Where implementation depends on the willing participation of the people in the room, their ownership of the decision is worth the hours it takes to build.
The impersonation
Here is the thing that does the actual damage.
A leader has already decided. They convene the meeting anyway, because deciding alone would look imperious, and they run what appears to be a consultation. People prepare. They bring analysis, they argue, they invest in the discussion. The decision is then announced, and it is the one the leader arrived with.
The outcome may well be correct. What has been destroyed is the value of every future meeting of that kind. Everyone present now knows the format is theatre, and the next time genuine input is wanted, they will supply the performance of participation rather than the substance, because the substance was not used last time and they have other work.
The equivalent failure with consensus is worse in its own way: a group is told to work it out together, spends six weeks doing so, and the leader then overrules the result because it was not what they wanted. That group will never again negotiate in good faith, and it will be described afterward as having become difficult.
Say the mode at the start
The whole correction is one sentence spoken before the discussion begins.
I am going to decide this and I want your input first. Or: we are going to work this until we all agree, and I will not override the outcome. Or: this is decided, I am telling you why, and the meeting is not about whether.
Nobody resents any of the three when told which one they are in. What people resent, correctly, is having spent two hours arguing under one set of assumptions and discovering afterward that a different set applied. The sentence costs nothing and removes almost the entire problem, and the reason leaders skip it is that command stated plainly sounds harsher than command disguised as consultation. It is not harsher. It is merely honest, and the disguised version was always visible anyway.
A note on the drift toward consensus
Organizations that pride themselves on collaboration tend to slide into consensus as a default, and the slide is rarely examined because it feels like a virtue.
Consensus applied to everything has two costs. It is enormously slow, which is visible and complained about. And it hands an effective veto to whoever is most willing to withhold agreement, which is invisible and matters more. In a group operating by consensus, the most obstinate person in the room has more power than the most senior one, and the outcomes will bend toward whatever they can be persuaded to tolerate rather than toward what the group actually thinks is best.
Which is why consensus needs a stated boundary and a stated deadline: this decision, worked this way, until this date, after which I decide. That preserves what consensus is good for without granting anyone a permanent hold on the organization’s ability to move.
Edited by Patrick J. Wolf, PhD