A leader can describe the standard they hold, and the organization has already established a different figure.
The operative standard is the worst behaviour that has been allowed to pass without consequence. Not the average, not the aspiration, and not what gets praised. The floor, wherever the last unaddressed instance happened to fall.
Why the floor rather than the average
People are not computing a central tendency. They are identifying a boundary, because the boundary is what determines their exposure.
The question anybody asks about a standard is not what is typical here. It is how far can this go before something happens. That question is answered by the most extreme case that produced no consequence, and one such case answers it completely.
This is why a single tolerated instance outweighs a great deal of consistent enforcement. Ninety-nine corrections establish that the rule is usually applied. One uncorrected violation establishes where it stops applying, and the second piece of information is the one people need.
Tolerance is mostly passive
Almost nobody decides to permit something. The permission is granted by not acting, which does not feel like an act.
A leader notices something, judges it too small to raise, is busy that week, or reasons that mentioning it would be disproportionate. Each of those is a defensible judgment about one occasion, and none of them registers as having established anything.
From the outside it is not ambiguous at all. Somebody did the thing, the leader was present or was told, and nothing followed. The organization does not know about the leader’s reasoning; it knows about the outcome, and it draws the only available inference.
Which produces the uncomfortable formulation: a leader is setting the standard continuously whether or not they are attending to it, and the moments that set it are precisely the ones that felt too minor to be worth the friction.
The four occasions that establish it
Most of the operative standard is set by a small number of recurring situations.
The first violation of a new expectation. The response, or its absence, converts the expectation into a rule or a suggestion, permanently.
The productive person’s exception. Whether the standard applies to somebody the organization cannot easily replace. Everybody is watching this one specifically.
The senior person’s conduct. A standard the leadership is visibly exempt from is not a standard. It is a rule for other people, and it is classified correctly within about a week.
Behaviour under pressure. What is permitted during the difficult quarter, when the deadline is real and the crew is short. Anything conceded then is understood to be conditional, and conditions recur.
The ratchet
Tolerance moves in one direction more easily than the other, and this is what makes small concessions expensive.
Lowering a standard requires no announcement. It happens by omission, once, and takes effect immediately.
Raising it requires stating that something previously permitted is now not, which means telling people who were behaving acceptably last month that they are not this month. That conversation is fair only if it is explicit, and it costs standing to have.
Which is why the correct response to a first small violation is not proportional to its size. It is proportional to the cost of the alternative, which is a standard that has to be reconstructed later against people who reasonably believed the old one applied.
What actually costs the friction
The reason leaders tolerate is not indifference, and naming it accurately is what makes the behaviour changeable.
Raising a small thing feels petty. It risks a reputation for being difficult over trivia. It requires a conversation nobody enjoys about something that does not obviously warrant one. And the person is otherwise doing fine, so the whole exchange feels disproportionate.
All of that is immediate, personal, and certain. The benefit — a standard that holds for years — is diffuse, invisible, and will never be attributed to the four-minute conversation that produced it.
That asymmetry appears throughout this Library and it resolves the same way here as everywhere else: by deciding in advance what will be answered, so the decision is not being made in a moment engineered to produce the wrong one.
The test
Name the worst instance of each stated standard that has occurred in the last year and gone unaddressed.
If you cannot recall one, ask a supervisor, who will be able to name it immediately.
That instance is the standard. Everything above it is what the organization aspires to, and everybody working there has known the difference since about a week after it happened.
Edited by Patrick J. Wolf, PhD