The review of the failed decision ran two hours, and it was merciless, and everyone in the room knew the verdict before the first slide loaded. The project had missed by a wide margin. The numbers were on the screen in a color nobody uses for good news. One by one the room walked back through the choice that had started it all, and with the outcome sitting right there in front of them, every warning sign that had been invisible eighteen months earlier was suddenly, damningly obvious. How did we not see this. The person who had made the call sat very still and said little, because there was nothing left to say that the result had not already said louder. They were, for every practical purpose, finished, and they knew it.
Two floors up, in a meeting nobody was dreading, a different leader was being congratulated for a bet that had paid off handsomely. It had been made on roughly the same quality of information, under roughly the same honest fog, as the one being dismantled downstairs. It simply landed the other way. Two decisions. Two leaders. One promoted in spirit, one retired in all but title. The only variable that truly separated them was the one neither of them fully controlled.
The organization owns one instrument, and it reads the wrong thing.
An organization will tell you, with complete sincerity, that it holds people accountable for their decisions. It does not. It holds them accountable for their outcomes, which is a different thing wearing the same suit. The two line up often enough that the swap goes unnoticed. Good judgment does produce good results more reliably than bad judgment, across a long enough run of calls. But any single decision, the kind that actually earns a review, is a tangle of judgment and luck, and the organization owns no instrument that can pull the two apart. So it measures the only strand it can see. The outcome is legible, countable, and already settled. The quality of the reasoning at the moment of the call is invisible, unmeasured, and gone. Handed one readable signal and one unreadable one, the review does what every system does. It grades what it can read, and it calls the reading judgment.
Hindsight dresses this up as rigor. Once you know how a thing turned out, the decision stops looking like a decision and starts looking like an answer that was always sitting there to be found. The uncertainty that genuinely surrounded the call, the real fork where careful people could have gone either way, collapses the instant the result is known. The post-mortem then runs on a set of facts the decider never had. Everyone in the room can see the iceberg in perfect detail. Nobody asks what the water looked like from the bridge at the time, in the dark, at speed, with the radar returning the same ambiguous smear it returns before every accident and before every uneventful crossing.
So the lesson lands, and it is the wrong lesson, and the whole building learns it quickly. The lesson is not decide well. The lesson is do not be the name on a bad result. Those are not the same instruction, and the distance between them is exactly where organizational nerve goes to die. A capable person who has just watched a colleague taken apart for an honest call that broke the wrong way runs the arithmetic before lunch. The reliable way to survive a system that scores outcomes is to never own an uncertain one.
Which is where all the familiar machinery comes from. Spread the call across enough people that no single name sits on it. Ask for another round of analysis, which buys time and quietly transfers the weight onto the data. Send it to the function whose entire job is to say no. Let the deadline make the choice so that no person has to. None of these moves improves a single decision. Every one of them is a rational answer to a scoreboard that punishes unlucky judgment far more dependably than it punishes poor judgment, because unlucky judgment at least leaves a body to stand over.
You are grading luck and holding ceremonies about it.
Here is the part that should unsettle anyone who chairs these reviews. The organization believes it is measuring the quality of its people's judgment. It is measuring, in large part, the quality of their luck, and then conducting solemn ceremonies around the measurement. And the better your decision-makers actually are, the more unjustly the system treats them, because real skill applied to genuine uncertainty still loses a respectable fraction of the time. Your scoreboard cannot tell a good decider who got unlucky from a poor one who got caught. It reads them as the same person. It elevates the coin that came up heads, discards the coin that came up tails, and files the difference under merit.
A sound decision can produce a terrible outcome, and a reckless one can produce a triumph. An organization that cannot tell these apart is not managing judgment. It is running a lottery with a blame ceremony attached.
Grade the call before you know the score.
Rebuilding this does not mean forgiving bad outcomes or pretending results are irrelevant. Results matter. It means the organization needs some way to judge a decision on the information that existed at the moment it was made, before the outcome arrives to contaminate the verdict. That is genuinely hard to build, because it cuts against the grain of hindsight and against the deep comfort of a clean number. It means reconstructing the fork as it actually stood: what was known, what was unknowable, what the honest spread of outcomes was, and whether the reasoning that chose among them was sound given all of it. A decision that was well made and broke badly should walk out of that review intact. A decision that was recklessly made and happened to win should not. Almost nobody runs reviews this way, because reading the scoreboard is easier, and calling the reading wisdom is easier still.
It is also why the most revealing thing you can watch is not a leader's track record but a leader deciding, live, before any result exists to flatter them or convict them. Put someone inside a crisis that is still unfolding, where the information is partial and the clock is already running and no outcome has been written yet, and you finally see the only thing that was ever truly theirs: how they reason when the answer is not yet knowable. A quarterly result can never show you this, because by the time the result arrives the reasoning has already been quietly rewritten to match it. We built the simulation to catch a decision in the one state where it still tells the truth, the state before anyone knows how it ends. It does not score whether the leader turned out to be right. It surfaces how they decided, which is the only part the next crisis will reuse.
The scoreboard cannot tell your best decision-maker from your luckiest one. The next hard call will.