Decisions
1 h

Premortem

Assume the decision has already failed, then write the causes — which surfaces objections that nobody would raise as a prediction.

Time cost
1 h
Output
A risk register with tripwires and owners.
Steps
6
Run it — Premortem generator →

Use when

  • You have essentially decided and want to pressure-test before committing.
  • A group has converged and nobody is dissenting, which is usually not agreement.
  • The plan has no slack and everyone is confident.

Do not use when

  • You want to kill a decision you have already rejected. The exercise will supply ammunition and look neutral doing it.
  • The decision is reversible and small. The failure is cheaper than the meeting.

Inputs required

  • A decision specific enough to fail
  • A horizon
  • People who will say what they think

Procedure

  1. 01

    State the decision and the horizon

    Specific enough that failure would be recognisable. “It is eighteen months from now” beats “eventually”.

  2. 02

    Assert the failure

    Not “what could go wrong” — “it failed, comprehensively”. The grammatical certainty is what unlocks the objections, because nobody has to be the pessimist.

  3. 03

    Write causes alone first

    Five minutes, silent, individually. Group generation anchors on whoever speaks first and loses most of the range.

  4. 04

    Pool and sort

    Read them all out before discussing any. Then sort into three: detectable early, preventable now, and neither.

  5. 05

    Assign a tripwire and an owner

    For every detectable cause: what signal, what threshold, who is watching, and what happens when it trips. A cause without an owner is a note.

  6. 06

    Act on the preventable

    Change the plan now for anything in the preventable column. If the plan does not change, the premortem was a discussion.

Characteristic failure mode

Producing a list and filing it. The output that matters is a changed plan and a set of owned tripwires; a document of risks nobody watches is the standard result and it is worthless.

Worked example

A household plans a major renovation over nine months.

  1. 01It is nine months from now and the project failed. Why?
  2. 02Fourteen causes written individually, pooled.
  3. 03Detectable early: supplier lead times slipping, the budget’s contingency being spent before month three.
  4. 04Preventable now: no written scope for the electrical work.

Result

Two tripwires with owners and dates, and one plan change made before starting. The contingency tripwire fires in month two, which is the point.

Where to go next

Also cited by
Second-order thinking · WRAP · Reference-class forecasting · Red-teaming · Nominal group technique · After-action review