It is eighteen months from now. This failed.
Not what could go wrong — what did. Assuming the failure is what gets people to say the thing they have been privately worried about.
Say it out loud first: “It is eighteen months from now. The plan failed. It is not in doubt and nobody is being blamed — we are only working out why.” Then everyone writes their reasons privately for five minutes before anyone speaks, because the first thing said out loud sets the range for everything after it. Write each one in the past tense.
Keep going past the obvious ones. The first three are always what is already being discussed; the useful ones tend to arrive around the sixth or seventh.
A structured way to run a conversation. The value is in what gets written down, not in the arithmetic.
Assume it failed, then ask why
The pre-mortem is Gary Klein’s technique and its power is psychological rather than analytical. Every team has been asked what could go wrong, and every team produces the same careful list — the risks that are safe to name, phrased so that naming them does not sound like a lack of confidence in the plan. The information that would have been useful stays in people’s heads.
Changing the tense changes the room. It is eighteen months from now, the plan failed, it is not in doubt, and nobody is being blamed. There is nothing left to defend, so the answers stop being diplomatic. People say the thing they have been privately worried about since the second week, and they say it in specific terms because they are describing an event rather than hedging about a possibility.
Two details make the difference between running this properly and running a meeting. Everyone writes before anyone speaks, because the first reason said out loud anchors the whole discussion and quietly persuades the person with the worse worry to revise it downward. And every reason is written in the past tense, because “we ran out of cash in month nine” produces a different conversation from “cash could become tight”.
Then the ranking, where most risk exercises fall apart. Scored individually every item looks important, so the list grows until nothing on it gets done, and the things nobody can influence crowd out the things somebody could fix by Friday. Three priorities, separated from a watch list, with the risks nobody would see coming marked as the expensive ones — that is a usable output. Eleven colour-coded rows in a spreadsheet is not.
What this tool does not do
It ranks what you put into it and can do nothing about what you left out - which is the whole risk of the exercise, because the failure that ends a business is usually the one nobody wrote down. The scores are subjective and the arithmetic that combines them is a convention rather than a measurement; the ordering is the useful output, not the numbers behind it. It cannot tell you whether the plan is good, only what the people in the room are worried about.
- Finding risks you did not think of - a facilitator, an outsider or someone who has done this before is worth more here than any tool
- Judging whether your scores are realistic, which is the thing most likely to be wrong
- Quantifying anything in money, probability or expected value
- Turning the priorities into a plan with owners and dates, which is the work that follows
- Insurance, contractual protection or anything about transferring risk rather than managing it
- Telling you when to run it again - once at the start is the common mistake; the useful cadence is at every point the plan materially changes
Frequently asked questions
How is this different from listing the risks?
The tense. Asking what could go wrong invites people to defend the plan, and the answers come back hedged. Asserting that it has already failed and asking why removes the need to defend anything — the failure is a given, nobody is being blamed, and the only job left is explaining it. People will say things in that framing they would never volunteer in the other one, and those are usually the things they were already quietly worried about.
Why should everyone write privately first?
Because the first reason said out loud sets the range for every reason after it. If someone opens with a moderate operational concern, the room calibrates to that and the person worried about something far worse quietly revises down. Five minutes of silent writing before anyone speaks costs nothing and is the difference between a real pre-mortem and a meeting.
Why does the tool only let three things be priorities?
Because a pre-mortem producing eleven priorities has produced none. Scored on their own, every risk looks important — that is why it got written down. The discipline is entirely in the ordering and in the willingness to leave the rest below the line until the top few are actually dealt with. Nothing is deleted; it waits.
What is detectability and why does it matter so much?
How early you would know. It is the axis almost nobody scores and it is usually what separates a manageable problem from a fatal one. A severe, likely risk you would spot within a week is something you handle. The identical risk, invisible until it has already happened, is the one that ends companies. The useful response to the second kind is often not to address the risk at all but to buy visibility of it — a weekly number on a dashboard is cheaper than the thing it warns you about.
What if everything on my list is outside my control?
That usually means the exercise has not gone deep enough. "The market turned" is rarely the real cause — the real one is "we had no cash buffer when the market turned", and that second version is a task. Where something genuinely is outside your control, it belongs in a watch list rather than a plan, with two things written next to it: what the earliest visible signal would be, and what you would do on the day you saw it.