“I went with the recommendation” is not a defence. It is a description of a decision you made, and the decision was about a person rather than about the substance.
This is the thing that gets lost when deference is discussed as either humility or abdication. It is neither. Choosing to accept somebody else’s judgment is itself an exercise of judgment, it can be done well or badly, and the accountability does not transfer along with the decision.
You are deferring to the wrong attribute
The common failure is not deferring too much or too little. It is deferring on the basis of something that has no bearing on the question.
Rooms defer to seniority, which measures time rather than repetitions of this particular judgment. They defer to confidence, which measures temperament. They defer to credentials, which measure what somebody studied twenty years ago. They defer to whoever has talked the longest, which measures stamina, and to whoever is most senior in the room, which measures the org chart. Each of these is a proxy, and each is uncorrelated with the only attribute that actually matters.
The attribute that matters is closed repetitions of this specific question. How many times has this person made a judgment of this exact type and then found out whether they were right?
That criterion produces uncomfortable results. The right person is regularly junior. A thirty-one-year-old project engineer who has run eleven change-order negotiations in eighteen months has more usable judgment about the twelfth than a vice president with thirty years of general construction experience and four such negotiations, all of them long ago. Everything in the culture of an organization pushes against acting on this. Almost nothing in the evidence supports the alternative.
Three conditions worth checking
Before accepting somebody’s judgment on something consequential, three things need to hold, and they take about a minute to run through.
Their repetitions in this specific question exceed yours, and by enough to matter. Their domain is one where the repetitions could have taught them anything, meaning outcomes actually came back and came back clearly. And their interests either align with yours or are visible enough that you can price the divergence.
The third condition is the one people skip out of politeness. A contractor’s judgment about whether the schedule is achievable is affected by what a longer schedule costs him. This does not make him dishonest and it does not disqualify his view. It means the view has a known lean, and a leader who has not identified the lean is not deferring to judgment. They are absorbing a position.
Defer out loud
There is a way of deferring that costs a leader authority and a way that adds to it, and the difference is entirely in whether the deference is stated.
Silent acquiescence looks like absence. The room proposes, the leader nods, the thing proceeds, and nobody can tell whether a decision occurred at all. Over time this reads as a leader who has stopped participating, and it invites people to route around the position.
Naming it does the opposite. Saying plainly that you are going with a particular person’s read, because they have done twenty of these and you have done two, and that you will own the outcome, accomplishes three things at once. It keeps the accountability where it belongs. It teaches everyone present what the actual criterion is, which is worth more than any statement about values. And it makes the deference legible as a choice rather than as a gap.
The person deferred to also behaves differently once it has been said out loud, because their judgment has been publicly relied upon and they now carry something.
Deferring to a room
A particular version of this deserves separate treatment, because it is common and it feels responsible.
Deferring to the consensus of a group is usually not deference to collective judgment. It is deference to whoever spoke first and most confidently, filtered through everyone else’s reluctance to be the lone objector. A room can be unanimously wrong with great conviction, and unanimity in a group that reports to the same person is close to uninformative, since the mechanism producing it may simply be that nobody wants to be the one who disagreed.
If the room’s view is going to be relied upon, it has to be gathered before the discussion rather than after. Independent written estimates, collected first, tell you something real. The same people polled after forty minutes of conversation tell you what happened in the conversation.
When refusing to defer is the failure
The mirror error is rarer but considerably more expensive when it occurs.
Somebody in the room has done this thing many times and found out. They say the ground is wrong, or the number will not hold, or the person should not be hired. The leader overrides, and the reason, examined honestly, is that overriding is what the position permits and accepting would have been a visible concession.
This is not a strong exercise of authority. It is negligence with a signature on it, and it is worth naming that plainly because it is almost never described in those terms afterward. When the outcome arrives, the account given is that reasonable people disagreed. What actually happened is that the best available judgment was present in the room and was set aside for reasons that had nothing to do with the question.
Edited by Patrick J. Wolf, PhD