Premortem Analysis

Definition

A premortem is a method developed by psychologist Gary Klein to help decision-makers anticipate problems before they occur. The process assumes the plan has already failed, then asks participants: why did it fail? What could have caused this? The goal is to identify key vulnerabilities and break false confidence without the full deliberation of formal risk analysis. It is "pre" (before execution) and "mortem" (as if dead)—imagining failure to prevent it.

In the Book

The handbook emphasizes that premortems work because people often feel overconfident once they've committed to a plan, especially if they lack experience with similar decisions. The premortem breaks that false confidence and punctures groupthink by shifting from "why will this work?" to "why will this fail?" The method is simple and fast (20–30 minutes). Step 1: all members familiarize themselves with the base plan. Step 2: imagine it failed; ask why. Step 3: individuals write down possible reasons for failure—this individual-first step ensures each person's experience and judgment surfaces. Step 4: consolidate lists in round-robin fashion, following four rules: more ideas are better; ideas can build on others; wacky ideas are welcome (they spark serious thought); and ideas must not be evaluated, criticized, or judged by body language. Step 5: use the failure list to revisit the plan and determine what must be mitigated. The ideal timing is just before war-gaming in the planning cycle, when the group has settled on a course of action but hasn't locked into execution mode.

Why It Matters

Premortem analysis is a low-cost way to surface what the group already suspects but hasn't voiced—the vulnerabilities each member perceives but doesn't want to raise. It uses mental simulation (imagining vs. analyzing) to bypass defensive reasoning. By framing the session as "we've failed; what went wrong?" rather than "what's wrong with this plan?", it makes critique about the plan's fate, not the planner's judgment. The technique works in any domain where plans are consequential: military operations, product launches, organizational change, medical procedures. It asks the question most plans never ask: if this goes badly, what's the most likely reason?