premortem-skill

Premortem · Thinking Tool

"A premortem is the hypothetical opposite of a postmortem. A postmortem in a medical setting allows health professionals and the family to learn what caused a patient's death. A premortem in a business setting comes at the beginning of a project rather than the end, so that the project can be improved rather than autopsied." — Gary Klein

Activation Conditions and Trigger Words

• Direct invocation: "Use premortem to analyze...", "premortem this project", "assume it has already failed"
• Semantic trigger: the user is planning a new project or decision and wants to identify risks in advance
• Combined invocation: "First use premortem to identify risks, then use scenario planning to build responses"

Method Framework Overview

The core of premortem is to assume the project has already failed and work backward to the causes of failure. It identifies and eliminates risks in advance by releasing imagination through mental time travel, releasing genuine opinions through psychological safety, and triggering preventive action through visualization of failure scenarios.

Core Principles (3-7, each must be accompanied by ≥2 pieces of cross-domain evidence)

Principle 1: Prospective Hindsight Releases Imagination

One-sentence definition: When you stand in the future and assume you have already failed, you will see risks that "now" cannot see.

Cross-domain evidence：

1. Cognitive psychology: In experiments, Klein found that premortem can identify 30% more project risks (Source: Klein, Performing a Project Premortem, HBR, 2007)
2. Oil and gas industry: After British Petroleum implemented premortem on some projects, the rate of safety incidents was significantly reduced (Source: Klein, Streetlights and Shadows, 2009)

How to apply: In a formal meeting, say, "Today we assume this project is completely dead in 6 months. Everyone, take 10 minutes to write down: Why did it die?"

Limitations: If the team lacks experience in the relevant field, they may not be able to come up with meaningful risks.

Principle 2: Collective Wisdom Requires Psychological Safety

One-sentence definition: People are naturally unwilling to say in front of leaders, "I think this is going to crash." Premortem lowers the psychological barrier by "assuming it has already crashed."

Cross-domain evidence：

1. Organizational psychology: Google Project Aristotle found that psychological safety is the strongest predictor of high-performing teams (Source: Google re:Work, Project Aristotle, 2016)
2. Nuclear industry: The accident reporting system of the international nuclear power industry relies on anonymous reporting to reduce fear (Source: IAEA, Safety Culture in Nuclear Installations, 2001)

How to apply: The leader speaks last in the premortem. If the leader speaks first, others will "intelligently" agree.

Limitations: In an organization with a severe punitive culture, the form of premortem alone is not enough to establish genuine safety.

Principle 3: Temporal Distance Reduces Confirmation Bias

One-sentence definition: Looking at the future from now is always optimistic. Standing in the future and looking back, optimism bias naturally fades.

Cross-domain evidence：

1. Behavioral economics: Kahneman & Tversky's prospect theory shows that people are risk-averse in the domain of gains but risk-seeking in the domain of losses (Source: Kahneman, Thinking Fast and Slow, 2011)
2. Investment decisions: Behavioral finance has found that fund managers overestimate the probability of success while holding (present perspective), but are more accurate when reviewing and evaluating (hindsight perspective) (Source: Barberis & Thaler, A Survey of Behavioral Finance, 2003)

Application Method: Use "why did it die" instead of "what could go wrong"; the former forces answering from a future perspective.

Limitations: Imagining fictional future failure scenarios may trigger excessive pessimism and needs to be balanced.

Principle 4: Failure Visualization Triggers Preventive Action

One-Sentence Definition: When you specifically imagine the picture of failure, you automatically develop the urge to "how to prevent it from happening."

Cross-Domain Evidence：

1. Aviation Safety: NASA and the aviation industry widely use Failure Mode and Effects Analysis (FMEA) to convert imagined failure modes into preventive measures (Source: NASA, FMEA Handbook, 2008)
2. Medical Safety: Surgical safety checklists come from systematic imagination of surgical failure scenarios (Source: Gawande, The Checklist Manifesto, 2009)

Application Method: Do not list a bunch of vague "risks"; each risk must be specific to a trigger scenario: "If ______ happens, the project will die."

Limitations: If identified risks are not converted into specific action items, premortem is just an interesting thought experiment.

Operational Protocol (Agentic Protocol)

Step 1: Problem Classification

Applicable Signals：

• The project is about to launch, already planned but not executed
• Important decisions need to be examined from multiple angles
• The team is overconfident and blind optimism may exist

Not Applicable Signals：

• The project is already in its final execution stage (the value of premortem is beforehand)
• Purely personal everyday small decisions (what to eat for lunch does not need a premortem)
• The team lacks relevant experience in the field (cannot think of meaningful risks)

Step 2: Premortem-Style Analysis

Research Dimensions(each dimension must trace back to a core principle):

1. Assume Failure Has Already Occurred: Set the time point "N months later, this project died"

   • Source Principles: Principle 1 (Prospective Hindsight) + Principle 3 (Temporal Distance)
   • Operation Method: You cannot say "may fail"; you must say "has already died," describing it in a past tense

2. Reverse-Infer Causes of Death Categories: Internal causes (things the team could have done right but did not) + External causes (environmental changes) + Black swans (completely unexpected)

   • Source Principles: Principle 2 (Psychological Safety) + Principle 4 (Failure Visualization)
   • Operation Method: Each person writes independently for 10 minutes; leaders speak last

3. Precursor Signal Identification: If each cause of death were to occur in reality, what is the first signal we could observe?

   • Source Principle: Principle 4 (Failure Visualization → Preventive Action)
   • Operation Method: For each cause of death, note "if this starts happening, what would we see?"

4. Preventive Measure Design: For each cause of death, design 1 specific preventive action and 1 detection metric

   • Source Principle: Principle 4 (Trigger Preventive Action)
   • Operation Method: Format "Preventive action [who] does [what] at [when]. Detection metric [what data] reaches [how much] triggers warning"

Step 3: Premortem-Style Output

• Cause of death list (classified as internal/external/black swan)
• Precursor signals for each cause of death
• Preventive measures for the Top 5 highest-probability causes of death
• Regular review mechanism (recommended to review once a month)

Applicable/Not applicable judgment

Problem type	Applicability	Description
Risk assessment before project kickoff	High	Classic scenario
Stress testing before important strategic decisions	High	Helps reveal blind spots caused by optimism bias
Team investment decisions	High	Before investing, ask "how could this money go down the drain"
Purely execution-oriented tasks	Low	Cleaning a room does not require a premortem
Small personal everyday decisions	Low	Overuse leads to analysis paralysis

Typical case library

Success cases

1. Gary Klein's Premortem experiment(Cognitive psychology/1990s)

   • Problem: project managers severely underestimated risks before kickoff
   • Method application: divide the team into two groups: one using premortem and one using traditional risk analysis
   • Result: the premortem group identified 30% more risks and in greater detail
   • Source: Klein, Performing a Project Premortem, HBR (2007)

2. British Petroleum project safety forecasting(Energy/2000s)

   • Problem: deep-sea drilling projects have high safety risks
   • Method application: incorporate premortem into the standard process in some projects
   • Result: reduced the incidence of safety accidents
   • Source: Klein, Streetlights and Shadows (2009)

Failure cases

1. Using hindsight as prevention(Management/General)
   • Problem: holding a review meeting only after the project failed
   • Misuse method: treating a failed postmortem as "prevention" for the next time
   • Lesson: postmortem is necessary but insufficient to prevent unknown risks in new projects. Each new project needs its own premortem
   • Source: Edmondson, Strategies for Learning from Failure, HBR (2011)

Misuse detector (≥3 misuse patterns)

Misuse signal	Detection logic	Warning message	Recommended action
Method does not match the problem	The user wants to use premortem to analyze a failure that has already occurred	"Premortem is used for pre-event forecasting. You are already in the post-event review stage, so using postmortem directly is more appropriate."	Recommend postmortem
Skipping key steps	Conducted a premortem but did not record action items	"A premortem that is not converted into preventive action is just an interesting thought experiment. Every cause of death must have a corresponding preventive action."	Guide the addition of action items
Turning into a complaint session	During the premortem, members blame all failures on external factors and others	"The purpose of a premortem is to identify preventable risks, not to find scapegoats. Everyone needs to also write down 'mistakes we ourselves made.'"	Guide focus toward internally controllable factors

Honest boundaries (≥3 specific limitations)

1. Depends on participant experience: If the team is unfamiliar with the domain, they cannot think of meaningful risks. The quality of a Premortem is limited by the richness of the participants' domain knowledge
2. May create excessive anxiety: In teams with insufficient psychological safety, it may be misused to create panic
3. Ineffective against completely unknown risks: Premortem cannot predict true black swan events (events outside the participants' cognitive scope)
4. Results are not equivalent to predictions: The output of a Premortem is "what might happen" rather than "what will definitely happen"

Research sources (proportion of primary sources must be >50%)

#	Source	Type	Primary/Secondary	Weight
1	Klein, Performing a Project Premortem, HBR (2007)	Article	Primary	High
2	Klein, Streetlights and Shadows (2009)	Work	Primary	High
3	Kahneman, Thinking Fast and Slow (2011)	Work	Primary	High
4	Gawande, The Checklist Manifesto (2009)	Work	Primary	Medium
5	Edmondson, Strategies for Learning from Failure, HBR (2011)	Article	Primary	Medium
6	NASA, FMEA Handbook (2008)	Technical Manual	Primary	Medium
7	Google re:Work, Project Aristotle (2016)	Research Report	Primary	Medium

This Skill was generated by Forge Skill — Forging Thinking Tools Generated