A pre-mortem is a meeting where a team imagines the project has already failed and works backwards to explain why. The technique was written up by Gary Klein in the Harvard Business Review in 2007, and it exists because people are much better at explaining a failure than at predicting one. Present the failure as a fact and you get more honest answers.
It works on personal goals for exactly the same reason, and it takes twenty minutes.
The setup is one sentence. It is six months from now and this did not happen. What went wrong?
Answer that before you start, while you are still enthusiastic enough to be generous with the truth and not yet invested enough to be defensive about it.
Why this beats thinking about risks
Ask what might go wrong and you get categories. Ask what did go wrong and you get scenes.
The forward-looking version produces a short list of respectable answers. Time. Money. Discipline. None of those generates a next action.
The past tense changes the quality of the answer:
- "Work went mad in October and I never got the Tuesday slot back."
- "I got three chapters in, read them, and they were bad enough that I stopped."
- "My gym partner moved and I never went alone."
Those are useful because they are specific enough to plan against. The past tense is doing the work: it gives you permission to describe a failure without it being an admission about your character.
Worth knowing: the same trick makes the second question in a weekly review work. "What got in the way" is answerable in the abstract; "which day went and why" is not.
Write five, not two
Two is the number you produce by stopping the moment the exercise starts to feel uncomfortable.
Give yourself five. Three through five are where the real ones live, and the prompts below are the ones that reliably surface them.
- What went wrong that had nothing to do with me?
- What did I get bored of, and roughly when?
- What did I discover I did not actually want, once I was doing it?
- Who or what did I turn out to be depending on?
- What did I tell myself at the time that sounded reasonable?
That last question is the sharpest one in the set. The sentence you use to let yourself off is usually predictable and usually the same one every time. "I will pick it back up after the holidays." "It makes more sense to start properly in January."
Writing it down in advance takes most of its power away, because you recognize it when it arrives.
The third question is worth taking seriously rather than dismissing. Discovering the goal is not one you want is a result, and it is close kin to why you procrastinate on the thing you actually want.
Sort them into two piles
Not every predicted failure deserves a plan.
Pile one: things you can remove now. These are the good ones. A goal that depends on a gym partner can be re-specified not to. A plan that dies when the Tuesday slot goes can have a named fallback slot. A first draft you expect to be disappointing can be pre-labeled as a bad draft, in writing, today, so that discovering it is bad in month two is confirmation rather than a verdict.
Pile two: things you can only prepare for. Illness, a work crisis, a genuinely full month. You cannot design these out. What you can do is decide the minimum version in advance: the smallest thing that counts as keeping the goal alive during a bad month. Fifteen minutes on a named day.
Deciding that now rather than on the bad day is the entire point, because bad days do not produce good decisions. Finding an hour that survives a full month is the same argument applied to the calendar.
Turn each one into a line you can act on
The output of a pre-mortem is three or four if-then lines sitting next to your plan. Not a list of fears.
- If Tuesday morning goes, it happens Saturday before ten.
- If the draft is bad at three chapters, I keep going to five before deciding anything.
- If I have not opened it in two weeks, the next session is fifteen minutes and no more.
- If I am waiting on someone else for more than a week, I do the next unblocked step instead.
Short, specific, and written where the plan is rather than in a separate document. Four lines is plenty. This is not a risk register.
The reason to write them as if-then rather than as intentions is Peter Gollwitzer's work on implementation intentions: a commitment with a trigger and a time attached behaves differently from the same commitment held as a general resolution.
Go further: the fourth stage of a GROW session asks what could stop the action and what the plan is if it happens. That is a one-line pre-mortem, and this is the twenty-minute version of the same question.
Where the predictions usually cluster
Across most personal goals the five answers land in the same four places, which is useful if you are stuck for a third one.
- The slot. Something took the hour and it never came back. The most common single failure by a wide margin.
- The size. The next step needed a clear afternoon, so it stayed at the top of the plan looking reasonable for six weeks.
- The evidence. You were measuring something that only moves slowly, so weeks of real work looked like nothing.
- The verdict. You looked at the work, compared it to the version in your head, and the gap did not seem worth closing.
If your five are all about motivation, you have written the polite version. Go back and name a day, an hour and a specific thing that took it.
Do it again at the halfway point
The pre-mortem you write at the start is guessing. The one you write in month three is informed, and it is the more valuable of the two.
By then you know which of your five predictions was right, which never came up, and which failure you did not see at all.
Twenty minutes to redo it with that information is the cheapest thing available, and almost nobody does it, because by month three the goal has either stalled and you do not want to look at it, or it is going well and it does not seem to need the attention.
Put the date in the calendar now, while the first one is fresh.
The limitation, stated plainly
A pre-mortem is a document, and documents do not enforce themselves.
The failure mode of this whole technique is a beautifully written set of if-then lines sitting in a file you have not opened since the day you wrote them, while the exact failure you predicted proceeds on schedule.
Two fixes, both unromantic. Put the lines where the goal lives rather than in a separate notebook, so reading the plan means reading them. And have something raise them at the point the condition triggers, which is what a file cannot do.
If you are doing this alone, the nearest approximation is a calendar entry at week three and week eight that says "read the pre-mortem". It is not as good, and week three is not an arbitrary choice.
The exercise, in five lines
- One sentence: it is six months from now and this did not happen
- Five answers in the past tense, not two
- Sorted into removable now and preparable only
- Three or four if-then lines, stored next to the plan
- A calendar entry at the halfway point to redo it
Common questions
How is a pre-mortem different from just listing risks?
The tense, and it changes the answers more than it sounds like it should. A risk list produces categories you already knew about. A pre-mortem produces scenes with days and causes in them, because you are explaining rather than forecasting.
Twenty minutes seems short. Should it take longer?
Twenty minutes is enough to get five answers and four lines, and longer sessions tend to produce a risk register instead, which is a document nobody reads. If you have time to spare, spend it on the halfway repeat rather than on making the first one comprehensive.
Does this work for goals with no deadline?
Yes, and it matters more there, because a goal with no deadline fails by fading rather than by missing a date. Set an artificial horizon for the exercise. Six months is a reasonable default and the specific number does not much matter.
Can I do this with someone else?
It is better with someone else, for the reason the technique was invented: a second person asks what you left out, and the fifth answer is usually the one you would not have written alone. Fifteen minutes over coffee is enough. A pre-mortem is also one piece of a larger pattern: what a coach actually does differently at the moment a goal gets set, start to finish.
What if my predictions all come true?
Then the technique worked and the plan was wrong, which is a cheap way to find out. The useful follow-up is whether the if-then lines fired: a prediction that came true with a prepared response is a good outcome, and one that came true with a line you never read is the limitation above.
The takeaway
The value is not in predicting accurately. It is in having three or four decisions already made before the day you need them, because the day you need them is the day you make bad decisions.
Twenty minutes, five answers in the past tense, four lines beside the plan, and a calendar entry to do it again in three months.
One thing to do today. Write the one sentence, answer it five times, and turn only the first answer into an if-then line. One line beside the plan is worth more than five in a document.
