A premortem begins by imagining that a proposed plan has failed, then asking what plausible conditions contributed. Collect explanations independently before discussing them. Select the concerns that can change the plan and assign prevention or observation steps. The exercise exposes assumptions; it does not predict failure or prove a risk will occur.
Set a failure story that people can question
Choose a specific plan and a sensible time horizon. “The company failed” is too broad for a useful discussion about next month’s onboarding experiment. A more workable starting point is that customers could not complete the intended handover and the team abandoned the trial. Define the imagined result clearly enough that colleagues can consider realistic causes rather than invent unrelated disasters.
Explain that this is a planning exercise and that concerns are welcome without requiring a defence of the current proposal. Invite contributions from people who understand the operational edges, including support or delivery colleagues outside the project’s core group. A plan can look coherent from the meeting room while depending on a routine process its designers have never actually observed.
Allow a short period of independent writing before the group conversation. A confident early speaker can otherwise set the entire discussion’s direction. Ask participants to describe a condition and its consequence, such as unavailable customer data preventing migration. Avoid naming a colleague as the cause. The relevant question concerns the system, dependency or decision that could make the imagined outcome plausible.
Turn concerns into decisions rather than a long risk list
Group related explanations and ask which ones would materially change the plan if true. Missing authority, unrealistic demand assumptions and an overloaded support team require different responses. Some concerns can be checked before work begins; others need an observation during the trial. Record the uncertainty plainly instead of assigning a precise likelihood merely because a spreadsheet includes a probability column.
For each selected concern, decide on a proportionate response. A sample data check may answer a compatibility question, while a handover rehearsal may reveal missing ownership. Agree who will perform the action, when the result is needed and how the plan changes if the concern remains unresolved. Keep essential safeguards in place; a premortem is not permission to experiment carelessly with people’s commitments.
Use the exercise to improve the plan
Finish by revising the working plan, not only saving the meeting notes. Add a dependency, narrow the scope or change a review trigger where the discussion justifies it. Leave other concerns on a brief watch list if immediate action would cost more than the available evidence warrants. Explain why they are being watched so silence is not mistaken for a forgotten issue.
Revisit selected concerns after the work has produced evidence. An imagined failure that never materialises does not prove the precaution was unnecessary, and a problem the group missed does not make the entire exercise worthless. Ask whether the conversation changed a useful decision and whether a previously overlooked perspective should participate next time. Preserve candid discussion by examining causes without turning the record into a blame document.
Fictional example: a trial depends on an absent approver
A fictional scenario to explain the method, not a customer result or a market benchmark.
A fictional document service plans a two-week trial with a regional association. During a premortem, participants imagine that the trial ended with no usable submissions. One person writes that members never received clear consent information; another suspects that the association’s approver was unavailable. A third worries that the file format was incompatible with older devices.
The team discovers that the planned launch overlaps the approver’s leave. It arranges an authorised substitute, checks the consent explanation and tests one representative older device before inviting participants. It does not build a replacement upload system on the strength of a hypothetical concern alone. These invented events illustrate different responses to different uncertainties. They do not claim that the trial succeeded or quantify a reduction in failure risk.
TAKE THIS INTO YOUR NEXT CONVERSATION
A practical next-step checklist.
- Define the imagined failure and time horizon.
- Collect explanations independently before group discussion.
- Choose actionable conditions and proportionate checks.
- Update the plan and assign a review owner.
Common questions.
Does a premortem require anonymous submissions?
Anonymity can help in some teams, but it does not replace respectful discussion. Choose a format that lets people raise concerns without personal consequences.
Should every concern receive a mitigation task?
No. Prioritise conditions that can change a meaningful decision. Watch or explicitly accept lesser uncertainties with an owner and rationale.
Further reading & source context.
These primary references provide background. They are not evidence for the fictional example above and do not endorse AccelerateLink.
Found an error or a missing qualification? Send an editorial correction.
Find my next move.
Use a practical tool to turn this idea into a clearer starting point.
Find my next move