Pre Mortem

Use this skill to run Pre Mortem as a practical consulting method, not as a generic framework explanation.

Method Notes

• Use prospective hindsight: assume the plan failed, then explain why.
• Convert failure causes into prevention actions and early warnings.

Required Inputs

Collect or infer these inputs before execution:

• planned action
• success criteria
• timeline
• stakeholders
• known risks

If an input is missing, do not block automatically. Mark it as missing, state the assumption used, and add a validation action.

When Not To Use

Do not use after a failure has already occurred; use root-cause analysis instead. A pre-mortem requires a specific plan and time horizon, not a generic list of things that might go wrong.

Adjacent Methods

• risk-matrix: prioritize known risk events after they are identified.
• scenario-planning: compare plausible external futures, not failure causes of one plan.

Step-by-Step Execution

Output Template

### 1. Plan And Failure Horizon
Plan:
Launch / decision date:
Assumed failure date:
Success definition:

### 2. Failure Narrative
The plan failed because:
Observed consequences:
Earliest warning:

### 3. Failure Modes
| Failure mode | Cause | Preventability | Evidence | Priority |
|---|---|---|---|---|
|  |  |  |  |  |

### 4. Prevention Plan
| Prevention / contingency | Owner | Timing | Trigger |
|---|---|---|---|
|  |  |  |  |

### Evidence And Next Decision
- Confirmed facts:
- Assumptions:
- Missing evidence:
- Next action, owner, and timing:
- Expected signal and decision threshold:

Quality Gate

• Produce the method-specific outputs for Plan And Failure Horizon, Failure Narrative, Failure Modes; do not substitute a generic framework summary.
• Tie every material score, causal claim, or stakeholder judgment to evidence or label it as an assumption.
• Include at least one disconfirming check, sensitivity, alternative explanation, or failure condition appropriate to the method.
• Convert the result into a named action or decision with an owner or stakeholder, timing, and observable signal.
• State missing inputs and the smallest validation action instead of inventing precision.
• Keep wording professional and plain enough that a smart non-specialist can use the output directly.