Something went wrong, and there is real pressure to identify who. A board wants to know. A community wants somebody answerable. The organization itself wants the matter closed.
Naming an individual satisfies all three immediately, which is exactly why it is dangerous. It ends the inquiry, and it ends it before the thing that actually caused the failure has been found.
The test
Would a competent person in the same role, with the same information, the same authority, the same training, and the same pressures have produced a materially different outcome?
If yes, this is genuinely about the individual, and individual accountability follows.
If no, removing the person changes nothing except the name of whoever is in position when it recurs. The system produced the outcome, the individual was the point at which it became visible, and the organization has now lost somebody and retained the defect.
This test is not a way of excusing people. Applied honestly it sometimes points squarely at an individual, and when it does the response should be proportionate and unhesitating.
Four things that are not individual failures
Each of these presents as somebody making a mistake and is not.
A decision made without the authority to make it. Somebody did the best available thing within limits they did not set.
Information that never arrived. A person cannot act on what they were not told, and the reason they were not told is nearly always structural.
A task nobody was assigned. When the notification does not go out and no role owned it, the failure is the assignment, not the person who happened not to do it.
Judgment degraded by conditions the organization created. Hour nineteen of a response with no rotation is a scheduling failure that presents as a personal one.
Where it is genuinely individual
The categories are narrower and clearer than the ones above.
Concealment. Somebody knew and did not say, and the delay is the harm.
A knowing breach of a standard the person understood and had the means to meet.
A pattern rather than an instance, particularly one previously addressed.
And failure to exercise authority the person actually held, which is the version most often mistaken for a system problem when it is not.
Where one of these applies, act, and act at a level the organization can see. Declining to hold a person accountable when the case is clear teaches everybody that standards apply only to systems, which is a lesson with a long tail.
Both errors are costly
Scapegoating removes somebody who was not the cause, leaves the defect in place, and teaches every remaining person that visibility is dangerous. They will respond by reducing what they report, which makes the next failure harder to see coming.
The opposite failure is less discussed and equally real. An organization that attributes everything to systems has made itself unable to address anybody, and the people who did their jobs properly watch somebody who did not carry no consequence. Systems language becomes a shelter, and it is recognised as one.
The distinguishing question is whether the analysis was conducted before the conclusion was reached. An organization that ran the test and landed on the system is being rigorous. One that reaches for systems language whenever an individual would be uncomfortable to name is doing something else.
The leader’s own account
Most system failures were somebody’s responsibility to design, and that somebody is frequently the person conducting the analysis.
The roles were unassigned because nobody assigned them. The authority limits were wrong because nobody reviewed them. The rotation did not happen because nobody scheduled it. Concluding that the system failed, without noting who owns the system, is a sophisticated version of avoiding accountability and everybody watching can see it.
Saying this out loud is the single act that makes the whole analysis credible. The notification did not go out because I never assigned it to anybody. That is on me and here is what I have changed.
External pressure for a name
Boards, funders, regulators, and communities frequently want an individual, and a leader will be tempted to supply one because it resolves the pressure quickly.
The defensible answer is to describe the analysis rather than to refuse. We looked at whether this was one person’s failure. It was not; here is what actually produced it and here is what has changed. Where it is one person’s failure, say that action has been taken without discussing a personnel matter publicly.
Supplying a name to relieve pressure is the most expensive thing available. It buys a fortnight, it costs a person who may not have deserved it, and the same failure recurs with a different name attached.
Edited by Patrick J. Wolf, PhD