The decision was made the moment someone leaned back and said, remember what happened with the Pune rollout.
Nobody in the room needed the sentence finished. The Pune rollout was the thing that had gone wrong, spectacularly, eighteen months earlier, and it had since acquired the status of scripture. There had been a deck before that sentence. There were numbers in the deck: forty-odd implementations, the overwhelming majority of them uneventful, a few that wobbled and recovered, one that burned. The numbers were on the screen while the sentence was spoken. And the moment it was spoken, the numbers stopped mattering. The room reorganized itself around the one that burned. The new policy, the one that would govern the next forty rollouts, was written that afternoon to prevent a fire that had happened once.
You have sat in this meeting. You may have been the person who said the sentence.
The case with a face beats the case with a sample size.
Here is what actually happened in that room, underneath what everyone believed was happening. A group of capable people thought they were weighing evidence. They were weighing one piece of evidence and ignoring thirty-nine, and they could not feel the difference, because the one piece had something the thirty-nine did not. It had a story. A name, a timeline, a villain or at least a scapegoat, a moment where it went sideways that people could still picture. The thirty-nine successful rollouts had none of that. Success is quiet. It files no incident report, holds no post-mortem, generates no story anyone repeats at the next planning meeting. The evidence for success exists, but it exists as a number, and a number has no face.
The mind does not count. It recalls. When a leader asks how risky something is, they do not retrieve a frequency, they retrieve an instance, and the instance that surfaces is whichever one is easiest to summon. Ease of recall is not frequency. It is vividness, recency, and pain. The rollout that burned is easy to summon precisely because it was rare enough to be shocking, which means the organization is now treating the rarity of an event as evidence of its likelihood. The more unusual the disaster, the more memorable it becomes, and the more memorable it becomes, the more it distorts the very estimate it is being used to inform. The exception does not just survive the meeting. It runs it.
And notice who wins these rooms. Not the person who is right. The person who tells the better story. The analyst with the cohort data loses to the colleague with the anecdote, almost every time, because the anecdote is built out of the material human attention is tuned to and the cohort data is built out of the material human attention slides off. We flatter this by calling it experience. Often it is just the loudest memory in the room, wearing the clothes of judgment.
The cost is not that one bad decision gets made. The cost is more expensive and much harder to see. The organization builds its entire apparatus, its approval gates, its extra sign-offs, its mandatory reviews, to prevent the recurrence of its most memorable failure, and then applies that apparatus to every ordinary case that will never resemble it. Forty teams now carry the weight of a process designed for one. The friction is permanent. The fire was a one-time event. You pay for the disaster continuously, long after the disaster itself is over, as a standing tax on everything that was working fine.
The story you remember is not data. It is a single data point wearing the costume of a pattern.
Now the reframe, and it should sit badly, because it absolves no one by pretending the room was lazy. The room was not lazy. The people in it did not skip the analysis. That is the uncomfortable part. They ran the analysis, faithfully, on a sample of one, and the sample was selected for them not by any method but by which event had hurt the most to live through. This is the trap. It does not feel like ignoring evidence. It feels like having been there, like hard-won wisdom, like the scar that teaches. Every instinct tells the leader they are being appropriately cautious. The instinct is the malfunction. The decision was not under-analyzed. It was analyzed with enormous care on the wrong number, and care applied to the wrong number produces more confidence, not less, which is exactly what makes it dangerous.
What it takes to decide by the forty, not the one.
Rebuilding this is not a matter of telling people to be more rational, which has never once worked in the history of organizations. It is a matter of changing what the room sees before anyone gets to tell the story. The frequency has to be in the room first, and it has to arrive as concrete as the anecdote, because an abstract base rate will always lose to a vivid instance in a fair fight, so the fight cannot be left fair. Someone has to be able to ask, out loud and without it being career-limiting, how often does this actually happen, and the answer has to be visible before the Pune rollout is invoked rather than after. The question that disarms the anecdote is not whether it was real. It was real. The question is whether it was representative, and that is a question most organizations never build a place for, because the person holding the painful memory experiences the question as having their pain minimized.
This is the pattern SSUNDAR's simulations tend to pull into the open, and leaders are usually startled by it. Put a capable executive under a running clock, stack a crisis that rhymes with something from their own history, and watch how fast they reach for the instance instead of the rate. They generalize confidently from a single vivid case because the pressure of the moment makes the available memory feel like the relevant one, and they do not notice they have done it. What the simulation reveals is not whether a leader is intelligent. It is whether they can tell the difference, with the clock running, between a pattern and a story that merely feels like one. Almost everyone believes they can. The data on that belief is not kind.
The most expensive decisions in most organizations are not the ones made carelessly. They are the ones made carefully, in answer to the thing that was easiest to remember.
The story you cannot forget is almost never the one you should decide by.