Project Premortem
A project premortem imagines that an initiative has already failed and works backward to expose plausible causes while preventive action is still possible.
Download Markdown (.md)Soroush’s Product Wiki · https://wiki.soroush.uk/topic/project-premortem/
01What it means
A project premortem is a structured risk exercise that asks participants to assume a proposed initiative ended in failure and independently explain what caused the outcome. This prospective-hindsight framing gives concerns permission to surface before launch, when plans can still change, rather than after an incident when hindsight and blame distort the discussion.[1][2]
- 01Set future checkpointDefine the initiative and when success would be judged
- 02Assume failureState that the initiative ended badly
- 03Write causes silentlyIndependent operational, customer, market, and ethical risks
- 04Cluster + assessPlausibility, severity, detectability, and influence
- 05Plan treatmentPrevention, contingency triggers, owner, and review date
- 06RevisitTrack signals and update the live plan
Imagined causes are not forecasts. Diverse expertise, external evidence, named owners, and follow-through matter more than producing a long dramatic risk list.
02Why it matters
Optimism, hierarchy, social pressure, and commitment to a favored plan can suppress weak signals. By making failure the premise rather than a criticism, a premortem broadens the set of risks a team considers and makes dissent easier to express. The output can strengthen plans, clarify leading indicators, and reveal assumptions that deserve testing.
03When to use it
Run a premortem before committing to a consequential launch, migration, strategy, partnership, or delivery plan, especially when uncertainty is high or participants hold different information. Repeat it when conditions change materially. It should supplement—not replace—technical analysis, security review, quantitative forecasting, or specialist risk assessment.
04How it works
- Describe the decision or plan and the future point at which success will be evaluated.
- State that the initiative has failed badly, then ask each participant to write plausible causes independently and silently.
- Collect causes without debate, inviting operational, customer, market, organizational, ethical, and dependency failures.
- Cluster and assess the causes by plausibility, severity, detectability, and ability to influence them.
- Assign preventive actions, contingency triggers, owners, and review dates for the most material risks.
05A practical example
Before moving every customer to a new billing system, a team imagines the migration caused revenue loss and account cancellations. Individual notes reveal an undocumented tax exception, support capacity limits, and a rollback process that has never been tested. The team stages the migration, rehearses rollback, and adds monitoring for invoice anomalies.
06Where it breaks
Imagined failure is not a forecast. Participants can anchor on vivid disasters, reproduce shared blind spots, or create a long risk list without changing the plan. A premortem also cannot uncover facts absent from the room, so diverse expertise, evidence, and accountable follow-through matter more than the workshop itself.
Common mistakes
- Brainstorming aloud from the start, allowing senior voices to anchor everyone else.
- Listing generic risks without converting important ones into controls, triggers, or owners.
- Using the exercise to perform skepticism while protecting a decision that leadership has made irreversible.
07Sources
- Performing a Project PremortemHarvard Business Review · Original practitioner articlePerforming a Project Premortem
- Back to the Future: Temporal Perspective in the Explanation of EventsJournal of Behavioral Decision Making · Research paperBack to the Future: Temporal Perspective in the Explanation of Events