Everything you have transmitted over four years can be confirmed or overwritten in about ninety minutes.
A crisis is the highest-bandwidth teaching moment available to a man, and almost nobody thinks of it that way while it is happening.
What people are actually watching
Not whether you solve it. They mostly cannot evaluate that.
Four things, and they can evaluate all four.
Your first thirty seconds. Whether you get sharp, go quiet, or start moving. That is the input everybody remembers and it happens before you have decided anything.
Who you blame in the first hour. Nothing is more diagnostic. A man who identifies a person in the first hour has taught the room what happens here when something goes wrong, and the lesson lands permanently.
And who you protect. This is the one that gets remembered for a decade.
Whether you absorbed it or distributed it. Whether the man who made the error was named to the customer. Whether you took it in the room and handled him privately.
The first thirty seconds
You will not think your way through this in the moment, which is why it has to be decided in advance.
Three things, in order.
Do not react. Ask.
What happened? Then be quiet and let the whole answer arrive.
A man who reacts before he has the facts has spent his credibility on a version of events that is usually incomplete, and he will have to walk it back, and everybody watched him do both.
Then say what you are going to do first.
I am going to call them. Then we will figure out the rest.
One action, out loud. It converts a room full of people with nothing to do into a room with a direction, and it costs four seconds.
And say the timeline.
Give me an hour and I will tell you where we are.
Uncertainty with a date attached is survivable. Uncertainty with no date produces speculation, and speculation is always worse than the facts.
The person who caused it
The single most instructive thing you will do.
In the room, in front of everybody: you own it.
This is on us. I will handle it.
Not Marcus made an error. Even when Marcus made an error. Especially then.
Two things that produces.
Everybody in that room now knows that a mistake here does not become a public execution, which means they will bring you the next one immediately rather than concealing it for three days.
And Marcus knows he was protected, and men do not forget that, and it buys more than any amount of loyalty language.
The exception is a genuine integrity violation. A safety issue somebody hid, or a dishonesty. That is a different category and it does not get absorbed, and the difference is obvious in the moment.
After
The part men skip and it is where the learning gets fixed.
Do the review within a week, when it is cold enough to think and warm enough to remember.
Three questions, and the framing determines whether you get anything.
What actually happened, in sequence?
Where was the first point we could have caught it?
What changes?
And write down what changes, because a crisis review that produces insight and no artifact produces the same crisis in eighteen months.
The review, worked
What the three questions actually produce, and the difference between a review that works and one that produces nothing.
The crisis: a job was installed to the wrong specification and the customer found it before anybody else did.
The version that produces nothing.
Why did this happen?
That question produces a name inside ninety seconds. Somebody says Marcus measured it wrong, everybody nods, and the review is over.
And nothing changes, because the finding was a person, and persons are already trying, and the next crisis will have a different name attached.
The version that works.
What actually happened, in sequence?
The spec changed on a Tuesday. The change was communicated by phone. Marcus was on a roof and did not write it down. The old spec was on the job sheet and nobody updated it.
Where was the first point we could have caught it?
Not the measurement. The phone call. A spec change communicated verbally with no written follow-up.
And notice that the first catchable point was four days before the error and two people upstream of the man whose name came up in the other version.
What changes?
Spec changes go in writing or they do not count. One line, added to the job sheet process.
Twenty minutes and one line of documentation, and the same crisis does not recur.
The thing that determines which version you get
The first review sets the pattern permanently.
If the first one produced a name, the second one will produce a defense, because everybody in the room learned what these are for.
And you will never get a clean sequence again, because men who expect to be blamed give incomplete accounts, and the incompleteness is not dishonesty, it is self-protection and it is rational.
Which means the first review is the highest-leverage thirty minutes here, and it is worth stating the rule out loud before it starts.
I am not looking for who. I am looking for where the first catchable point was.
The crisis that reveals the man
One honest paragraph.
Some men are better in a crisis than in ordinary time.
Calm, clear, decisive, and everybody wants them in the room. And they are difficult to work for on a Tuesday, because ordinary time bores them and they are subtly waiting for the next one.
If that is you, know it. A man who is at his best in an emergency will unconsciously permit conditions that produce emergencies, and everybody around him lives inside that, and nobody will ever say it to him.
Ask one man. Am I better in a crisis than the rest of the time, and is that a problem?
Drawn from Carry the Weight