A pre-mortem does not ask what could go wrong; it declares that the plan has failed and asks why. Asking what could go wrong keeps failure hypothetical, so every doubt reads as an argument against a plan the group already backs. Declaring failure makes it the premise, so explaining it becomes the task.
The decision researcher Gary Klein described the method in Harvard Business Review in 2007. The name reverses the post-mortem, the review held after something has failed: you run a pre-mortem before you commit, while the plan can still change. The format below adapts Klein’s procedure into a 20-minute meeting for a small team. Use it to surface risks and decide what to do about them. A longer risk list does not, by itself, prove a better decision.
What changes when you declare “it has failed”?
Klein’s diagnosis is blunt: “Projects fail at a spectacular rate,” partly because “too many people are reluctant to speak up about their reservations during the all-important planning phase.” In a pre-mortem, everyone explains the same imagined failure, so a doubt is an answer, not an objection. The wording alone cannot make people feel safe enough to speak, though; how the decision owner reacts to the first uncomfortable reason still matters.
Keep the distinction clear as you facilitate. Failure is certain inside the exercise. The causes people propose remain possibilities to examine.
What one pre-mortem surfaced
Klein’s article describes a pre-mortem on a project to give military air-campaign planners advanced computer algorithms. A team member who had said nothing during the kickoff meeting pointed out that one algorithm would not fit easily on the laptops used in the field: the software would take hours to run when users needed quick answers.
It turned out that the algorithm developers already had a shortcut, which they had been reluctant to mention. It went into the plan, and Klein reports that the project was highly successful.
The example shows the intended sequence: a quiet reservation becomes explicit, and the plan changes. As one reported case, it does not show how much the pre-mortem contributed to the outcome.
What does the research actually show?
Klein cites a 1989 study by Deborah Mitchell, J. Edward Russo and Nancy Pennington, and writes that prospective hindsight, imagining that an event has already happened, “increases the ability to correctly identify reasons for future outcomes by 30%.”
The study’s abstract reports something narrower. Explanations of outcomes framed as certain “tended to be longer, to contain a higher proportion of episodic reasons, and to be expressed in past tense.” Those are differences in the explanations, not a reported measure of accuracy. Jason Collins’s course notes on the pre-mortem state that the study “did not assess the quality of the reasons.”
This page draws on that abstract and Collins’s commentary; the full paper was not reviewed. Neither supports reading the popular “30% better” claim as a demonstrated gain in decision accuracy.
Chapter 6 of The Calibrated Mind keeps the same boundary. It treats the pre-mortem as strong for surfacing risks but cites no randomized trial isolating its effect on decision accuracy. Whether the exercise improves the final decision remains open on the evidence reviewed here.
The 20-minute script
Choose a real plan that everyone attending already knows. Share it before the meeting, so the first two minutes are a recap, not a pitch.
Assign three roles:
- Decision owner: accountable for the plan and the follow-up. Listens without defending the plan during the read-out.
- Facilitator: keeps time and holds discussion until every reason is collected. Ideally someone other than the decision owner.
- Scribe: records the reasons where everyone can see them. In a small team, the facilitator can do this too.
| Minutes | Step | What happens |
|---|---|---|
| 0-2 | Frame | The decision owner recaps the plan in one sentence, with a date. No new proposals. |
| 2-3 | Declare failure | The facilitator says: “It is one year from today. This plan failed badly. Not might have. Did.” |
| 3-8 | Write alone | In silence, everyone writes down every reason they can think of for the failure, including reservations they would normally keep to themselves. |
| 8-14 | Read out | One reason per person per turn, starting with the decision owner. No debate, no defence. At minute 14, collect any reasons not yet read out. |
| 14-17 | Group | Merge overlaps. Choose three risks that are plausible and costly. A serious cause raised by only one person can make the three. |
| 17-20 | Assign | Give each risk an owner and a response: prevent, monitor or accept. For a monitored risk, name the warning sign and the threshold that reopens the plan. |
The first four steps follow Klein, whose read-out starts with the project manager and continues until every reason is recorded. The timings, roles and grouping are this page’s practical design, not a tested protocol, and the prevent / monitor / accept rule comes from The Calibrated Mind. If time runs out, keep the unread reasons for the decision owner’s review rather than treating the spoken list as complete.
Copy-paste agenda
PRE-MORTEM (20 min)
Before the meeting: share the plan with everyone attending.
Plan: __________________________ by (date): __________
Decision owner: ________ Facilitator: ________ Scribe: ________
00-02 Frame: owner recaps the plan in one sentence, with a date
02-03 Declare: "It is one year from today. This plan failed badly. Not might have. Did."
03-08 Write alone, in silence: every reason it failed
08-14 Read out: one reason per person per turn, owner first, no debate
At minute 14, collect any reasons not yet read out.
14-17 Group overlaps; choose 3 plausible, costly risks
17-20 Assign each risk an owner and a response: prevent / monitor / accept
For monitor: warning sign + threshold that reopens the plan
For each of the 3 risks:
Risk: __________________________
Owner: _________________________
Response: prevent / monitor / accept
Action, or reason for accepting: __________________________
If monitoring, warning sign and threshold: ________________
Follow-up: owner reviews the full list; check responses at the first milestone.
Plan changed? Y / N. If not, why not? __________________________Turn the list into decisions
After the session, the decision owner reviews the full list for ways to strengthen the plan, as in Klein’s procedure. The three priorities focus the immediate work; they are not permission to discard everything else. A concern that came up several times deserves attention, but repetition is not a measure of likelihood, so examine serious causes that only one person noticed, too.
An unchanged plan is not automatically a failed exercise. Record why the risks are already covered or worth accepting. The useful test is whether each serious concern received an explicit response.
Keep the risks in the project plan and, for a decision that is hard to undo, in a five-line decision journal. Check the responses at the first milestone.
How does a pre-mortem go wrong?
It becomes a complaint session. Require each reason to explain how this plan failed, in the past tense, with a cause. “Management never backs us” names no cause. “The launch slipped because legal review took six weeks and nobody booked it” does. Ask for the missing cause before moving on.
It becomes a vote. An imagined failure is not evidence that you should abandon the plan. Review the causes and what you could do about them first, then decide whether to proceed.
One voice takes over. Keep the writing silent and independent. During the read-out, take one reason per person per turn and stop anyone from defending the plan.
It becomes routine overhead. Reserve the session for consequential plans that are expensive to reverse. For a cheap, reversible choice, trying it may be the more useful test.
The solo version (10 minutes)
Working alone? Write a paragraph in the past tense: “It is a year from now and this failed. Here is what happened.” Underline three causes and give each a response: prevent, monitor or accept.
You lose other people’s knowledge and objections, so treat the result as one person’s account of how the plan could fail. Use it to decide what to check before you commit.
Where this sits in CALIBRATE
CALIBRATE is the nine-step method in The Calibrated Mind for matching confidence to evidence. You do not need the book to run this session.
The pre-mortem belongs in step 4, Investigate alternatives: test a belief against the strongest competing explanation. Here the belief is that the plan will work, and the competing explanation is a specific story of how it failed.
Consider a hypothetical five-person software company planning to move every customer to a new billing system in the third quarter. Its failure story is specific: “It failed because customers on old annual contracts could not be moved automatically, support was flooded in week one, and nobody had been assigned to fix contracts by hand.”
That story gives the team a risk (old contracts), an owner (the finance lead) and something to monitor (manual billing tickets). More than ten in the first week is an illustrative threshold for reopening the plan, not a research-backed benchmark.
The CALIBRATE framework page sets out all nine steps. Calibration techniques for founders covers the habits around it.
Try this today (20 minutes)
Pick a plan you expect to commit to in the next month. Share it with two or three people who know the work, and paste the agenda above into the invitation.
After 20 minutes, check the output: three risks, an owner and a response for each, and a warning sign and threshold for anything you will monitor. Record what changed, or why nothing needed to. That is the test you can run today. Whether it led to a better decision is a separate question, and one meeting cannot answer it.
Cite this:How to run a pre-mortem in 20 minutes.Len P. van der Hof. https://lenvanderhof.com/en/blog/how-to-run-a-pre-mortem/ ·