RESULT OBTAINED
Canonical method
Pre-mortem and red team: attacking the retained option before committing
What must be established
A useful analysis keeps its controls visible.
- 01
Set the option tested, the horizon and what failure means for this decision before the workshop.
- 02
Link every cause of failure to an assumption of the choice and to an observable early signal.
- 03
Give the challenge to people who did not design the choice, and record the objections left unresolved.
Decision reference
Make the next choice more explicit.
DECISION SUPPORTED
What would make the retained option fail, and how can it be tested before resources are committed?
DELIVERABLE TO KEEP
A dated register of causes of failure, each linked to an assumption, a signal and a decision.
NEXT ELEMENT TO EXAMINE
Strategic assumptionExplore →Write the failure before looking for its causes
A pre-mortem opens with a sentence in the past tense: at the horizon set, the retained option has failed. That inversion changes what the group looks for. As long as failure remains a possibility, everyone defends the choice they have just approved; once failure is taken as given, the discussion turns to its causes and speech is freed. Failure still has to be defined before the workshop, otherwise each participant reasons about their own: adoption that is too slow, insufficient margin, withdrawal from the market or damage to reputation share neither the same causes nor the same remedies. The statement therefore sets the option tested, the horizon, the measure that establishes failure and the scope concerned. It also separates failure of the choice, which comes from a false assumption, from failure of execution, which comes from a badly run plan. The latter exist, but they belong to the plan rather than to the decision: mixing them means reopening a choice for reasons that do not concern it.
- Option tested
- Dated horizon
- Measure of failure
- Choice or execution
Trace the cause back to the assumption behind the choice
A cause of failure only becomes useful once attached to what made it possible. Every stated cause is therefore linked to an assumption of the choice: a supposed customer preference, a supposed available capability, a supposedly sustainable price, a supposedly short lead time. That tracing separates the causes that attack the reasoning from those describing a hazard unrelated to it. It also makes the exercise verifiable later: to every assumption targeted corresponds an early signal, observable before failure has been consummated, and a threshold set before looking at the data. Without that threshold, interpretation follows the result one hopes for. The register keeps the cause, the mechanism through which it produces failure, the assumption targeted, the signal, the threshold and the intended countermeasure. A cause without an assumption or a signal remains a worry: on its own it does not justify abandoning the option, but it does point to a gap in the evidence.
- Cause and mechanism
- Assumption targeted
- Early signal
- Threshold set beforehand
Organise contradiction without turning it into a ritual
The red team extends the pre-mortem by giving the challenge to people who did not design the choice. Its value lies in that independence: whoever built the option defends its assumptions without meaning to. The mandate must be written, otherwise contradiction turns into a review of form: what the team may attack, what it must produce, to whom it reports, and what becomes of the objections left unresolved. That last point separates a useful exercise from a ritual. An objection received without an answer must stay visible in the decision dossier, with its status, rather than disappear into minutes. The exercise ends in a written and dated decision: maintain the option, amend it by narrowing the scope, sequencing it or keeping an exit option, or abandon it. Without that conclusion, the pre-mortem produces a list of worries that the commitment of resources will render moot.
- Written mandate
- Independent team
- Unresolved objection
- Dated decision
DECISION ASSET
Pre-mortem register
Every supposed cause of failure is linked to an assumption of the choice, to an early signal and to a written decision.
Failure
- Input
- Option, horizon, measure
- Control
- What counts as failure?
- Output
- Dated statement
Cause
- Input
- Mechanism and sequence
- Control
- How does failure arrive?
- Output
- Linked cause
Assumption
- Input
- Supposed evidence of the choice
- Control
- Which assumption falls?
- Output
- Early signal
Decision
- Input
- Unresolved objections
- Control
- Maintain, amend, abandon?
- Output
- Dated decision
- Gary Klein, Performing a Project Premortem, Harvard Business Review, 2007 (opens in a new tab)Founding article of the method: assume failure has happened, then look for its causes.
- Ministry of Defence, Red Teaming Handbook, third edition, 2021 (opens in a new tab)Public handbook of organised challenge, in which the pre-mortem is one technique.
MOVE FROM ANALYSIS TO YOUR DECISION
Produce a testable decision mandate
MOVE FROM ANALYSIS TO YOUR DECISION
Produce a testable decision mandate
RETURN TO THE DECISION SYSTEM
See the canonical definition of marketing strategy→