Skip to content

Pre-mortem Analysis

Why It Matters

Humans are overconfident in their judgments and plans. Prospective analysis ("what could go wrong?") generates superficial concerns. Retrospective analysis ("it failed - why?") generates concrete, specific failure modes. The pre-mortem exploits hindsight bias - the tendency to see past events as inevitable - by placing you mentally in the future where failure has already occurred.

The Key Move

Gather your team. Announce: "It is 18 months from now. Our analysis was completely wrong [or our plan failed catastrophically]. Take 5 minutes to write down why." Each person identifies a plausible cause of failure. The exercise surfaces concerns people were reluctant to raise (social permission to criticize) and generates specific failure modes (not generic "it might not work"). Then systematically evaluate: which failures are most likely? Which would be most consequential? What could prevent them?

Classic Application

Intelligence assessment of Iraq's alleged mobile bioweapons labs. A pre-mortem would have asked: "We concluded these are bioweapons labs, but we're wrong - why?" Plausible answers: "The source was fabricating," "The equipment has dual use," "Iraqi deception designed to be discovered." These were the actual causes of the analytical failure, but weren't seriously considered because the evidence seemed compelling.

Surprising Application

Software deployment planning. Before major releases, engineering teams conduct pre-mortems: "The deployment failed catastrophically - what happened?" Engineers surface specific concerns about database migrations, API compatibility, traffic spikes, and edge cases that standard risk analysis missed. One team discovered a critical race condition this way that conventional testing hadn't revealed.

Failure Modes

Generating only generic failures ("bad luck," "unexpected events") rather than specific mechanisms. Conducting it too late after commitment is made (pre-mortem must precede decision lock-in). Not acting on identified failure modes (it's diagnostic, not decorative). Using it to assign blame rather than surface risks (creates defensiveness). Assuming listing failures prevents them (you must design mitigations).

Go Deeper

Klein, "Performing a Project Premortem" (Harvard Business Review, 2007) - Klein invented the technique. Kahneman, "Thinking, Fast and Slow" Chapter 24. Pherson and Heuer, "Structured Analytic Techniques for Intelligence Analysis," Chapter 8.