What is a Pre-mortem?

Before any big commitment, a team can list the reasons it will succeed; those are in the deck. The reasons it will fail are usually available too, held privately by the people closest to the work. They are not missing. They are unwelcome, so the meeting never hears them. The pre-mortem is a ritual for extracting them while they are still cheap.

The exercise was described by research psychologist Gary Klein, and it inverts the post-mortem. Instead of examining a failure after it happens, the team assumes the failure before committing. The facilitator sets the scene: it is a year from now, the commitment was made, and it has failed, completely and undeniably. Each person, working alone, writes the story of why. The stories are read out, clustered, and the commitment is revised, protected or abandoned in the light of them.

Why imagined failure works

The mechanism Klein drew on is called prospective hindsight: imagining an event as certain and then explaining it. Asking what might go wrong invites probability estimates and polite hedging. Asking for the story of a failure that has definitely happened invites narrative, and people are far better at generating causes for a definite event than for a hypothetical one. The question changes what the mind goes looking for.

The second mechanism is social. By the time a big commitment reaches its final meeting it has momentum: a sponsor, a deck, a deadline. Doubt reads as disloyalty, so doubt stays private. The pre-mortem reverses the incentive. For one hour, everyone is required to be the author of a failure, so spotting the fatal flaw becomes a contribution instead of an obstruction, and the delivery lead’s quiet worry about the data migration finally gets written down. The independent writing matters as much as the premise: stories composed alone, before discussion, cannot anchor on the most senior story in the room.

A pre-mortem works because it moves the argument about failure to the only moment it is cheap: before the commitment is made.

Where it belongs in a services firm

The ritual earns its hour wherever a commitment is hard to unwind: the one-way doors of a people business. Three in particular.

• Big bids. Before the final price, assume the firm won the work and the work destroyed the margin and the relationship. The stories surface the difference between winning an engagement and surviving it.
• Fixed-price work. Assume the fixed price ran badly over. The stories will name the acceptance criteria that were never tightened, the client data quality that was assumed rather than verified, and the scope creep nobody priced.
• Key hires. Assume the new practice lead has left within the year. The stories describe the role as it was sold against the role as it exists, and the pipeline that was promised against the pipeline that is real.

One concrete example

Clearly illustrative, with no customer implied. A firm is about to sign a fixed-price replatforming engagement. The bid has passed every review, and the team is confident. One hour before final sign-off, the bid team and the named delivery leads run a pre-mortem: the engagement failed, write the story. The stories cluster around three causes. The client’s data quality was assumed, never verified. The two architects named in the plan are committed to another account for the first quarter. The acceptance criteria in the statement of work are vague enough to argue about for months. None of the three had appeared in any bid review, yet everyone recognised each of them instantly. Two become contract changes: a paid discovery phase to verify the data, and tightened acceptance criteria. One becomes a staffing decision made before signature instead of after failure. The price rises. The firm signs a smaller, truer commitment, or walks away knowing exactly why.

The pre-mortem in the ONX vocabulary

A pre-mortem generates hypotheses; the discipline of decision intelligence gives them somewhere to live. Each failure story points at an assumption with an evidence state: the claim that the client data is ready was merely stated, and the pre-mortem is what demands it be measured before the commitment leans on it. The stressed assumptions can then be run rather than debated: a Scenario Run prices the options with the failure conditions applied, so what if the data is not ready becomes a dated, costed answer instead of a mood. And because the stories are recorded against the decision itself, the eventual decision review can compare the failures the team imagined with the one that actually arrived. Over time the firm learns which of its fears predict, and which merely recur.