Essay · leadership
Accountability is a claim you have to be willing to withdraw.
A number goes red and the room asks who owns it. That question wants a name. Accountability wants a standard, a record the person has already seen, and the nerve to take the claim back when the evidence turns.
A number goes red in the monthly review. Before anybody asks what moved in the queue, somebody asks who owns it.
The question sounds like rigor. What it wants is a name, and a name is available in the time it takes to scroll one slide. Producing one requires knowing nothing about what happened.
The expensive version of that question goes by the same word.
Three things wearing one word
Was the person where they were supposed to be, did the number hit, or is anybody better off for their having been here?
The first two are cheap. Presence has a schedule row and a timestamp behind it. Activity has a count. Both can be pulled by somebody who never watched the work, which is most of why they get used.
The third has no report behind it. It gets answered by somebody who can name what changed for a customer or a teammate, and point at it. An operation can run for years on the first two and call the result a performance system.
Why the fast answer keeps winning
Blame is quick. It needs no prior conversation. It survives a room where nobody has read the detail. It ends the inquiry, which is the part that feels like resolution.
The first artifact is the performance conversation that names a number. A person is told their handle time is high, or their quality score slipped. They leave knowing the verdict and not knowing what to do differently on the next call.
The second is the review that surfaces something for the first time. The team leader has been carrying it for weeks, in a notebook nobody else has seen, waiting for the moment when the form makes saying it easier. That team leader is not the villain here. They learned that the informal version costs them a hard afternoon and the formal one costs them a signature, and nobody priced the difference.
The third is the escalation nobody tested. Before anything about willingness belongs on the table there are four ordinary explanations, each of them something the organization failed to supply: a skill nobody taught, a tool broken for two weeks, a policy that says the opposite of the coaching, and an expectation nobody stated out loud. Check those and a good number of escalations lose their subject.
Step back a level and the same reflex runs the building. Where the hierarchy is steep and the performance system points only downward, raising a problem is how a person gets named in a meeting. Problems stop being raised. What the floor knows dies at the first level of supervision, and mistakes stay hidden until they are expensive. The organization concludes it has an accountability problem. It has the reverse.
What the record has to survive
The standard has to name a behavior. Not a number, and not more ownership. A person has to be able to hear it on Thursday and do something different on Friday.
Nothing appears in a formal conversation that the person has not already heard informally. Before the review, the manager strikes from their own list anything they cannot attach to a conversation that happened: a date, the behavior they watched, the ask, the follow-up. We talked about your numbers fails all four. In the room, the person can say they are hearing an item for the first time, and it comes out of the record. The gap is the manager's. The price is saying the thing in the week it happened, to someone's face, with nothing formal standing behind you. That discomfort is the work.
The case has to be withdrawable. An escalation is a claim about a person, and claims can be wrong. When one of those four turns up halfway through, the case for that issue goes back to the start: fix the gap, confirm the fix held, then watch a fair stretch of ordinary work before anyone escalates again. The earlier steps stay in the file as history. They stop counting as evidence about what the person was willing to do.
An operations director reading this has the objection ready, and it is a fair one. Run it and nobody is responsible for anything. Every escalation resets on a newly discovered excuse, the writing lands on managers already underwater, and the checked-out person learns the process faster than anyone and rides it for a year. The reset is where that fear is earned. A reset triggered by anything a person says would be the failure they describe, but there are four triggers here and each names something the organization owed and did not deliver. A bad quarter is not one.
The sequence still ends. Testing an explanation changes what has to be true before a conversation reaches the point where the fit itself is on the table.
That writing is the manager's job arriving in the week it happened rather than all at once in a review nobody can defend. Skipping it does not save the conversation. It defers it, and the deferral gets paid twice: by the people who resign having never been told, and by the cases that fall apart the first time anyone reads the record.
None of this covers conduct. Theft, harassment, falsifying a record: those never enter a coaching sequence, and the people who have to be in the room are there from the first conversation. Putting coaching first is a commitment about the coachable cases, not a claim that every case is one.
One test tells you which word an organization is using. Watch what happens when the evidence turns against a claim it has already made. If nothing in the record can ever be withdrawn, the word is doing different work than it says.
Want to talk about this? Email me.