Learn
How to run a pre-mortem before you commit
A pre-mortem asks a team to imagine that it is a year from now and the plan has failed, then to write down why. Stating the failure as a fact frees people to say what they already suspect. The exercise takes under an hour, and it finds problems while they are still cheap to fix.
Where the method comes from
The psychologist Gary Klein described the pre-mortem in the Harvard Business Review in 2007. He built it on research into what is called prospective hindsight: people asked to explain an outcome as though it had already happened produce more reasons, and more specific ones, than people asked whether it might happen.
The shift is small and the effect is large. Asking whether a plan could fail invites defense. Telling people that it did fail invites explanation.
Running one in five steps
Gather the people who will carry out the plan, not only those who designed it. Describe the plan briefly, then announce that it has failed badly. Give everyone a few minutes to write, alone and in silence, every reason they can think of.
Go around the room taking one reason from each person at a time until the lists are exhausted. Finally, choose the three reasons that are both likely and serious, and give each an owner and a change to the plan.
Writing alone first matters. It keeps the most senior voice from setting the list for everyone else.
Turning stories into tests
A reason such as the support team could not cope is a story. To act on it, turn it into a claim that can be checked: support can handle a stated number of tickets a week, and the launch will create no more than that. Now there is a figure to test, a person to test it, and a point at which the plan should change.
The common mistake
Running the exercise and filing the notes. A pre-mortem that ends as a list on a shared drive has cost an hour and changed nothing. The second mistake is accepting a mitigation that exists only on paper, such as a policy to monitor the risk, when the risk needs a budget, a person, or a date.
The pre-mortem in a Decision Room
Argumont gives the pre-mortem's work to Participants whose mandate it is. The Gray Wolf hunts the hidden weaknesses a hopeful team will not volunteer, the Red Teamer runs the scenarios that could break the plan outright, and the Grandfather asks what cannot be undone and what the cheapest way back would be. Each failure story becomes a claim the Inspector can mark, and the Verdict states which ones still threaten the leading option. The Human Decision remains with the team.
The launch that swamped support
A firm plans to launch a self-service product to 5,000 existing customers. In the pre-mortem, three people independently write that support was overwhelmed. The claim is tested against history: past launches drew tickets from about 12 percent of customers in their first two weeks, which here means 600 tickets in two weeks, or 300 a week. The support team can absorb 220 a week alongside its normal work, leaving a gap of 80 a week. The team staggers the launch over four weeks, bringing the load to about 150 a week, well inside capacity. One hour of imagined failure prevented a real one.
Questions
- How is a pre-mortem different from a risk assessment?
- A risk assessment asks what might go wrong, which invites cautious and general answers. A pre-mortem states that the plan has already failed, which draws out specific reasons people might otherwise keep to themselves.
- Who should attend a pre-mortem?
- The people who will carry out the plan, with at least one person from outside the team. Those closest to the work usually know the likeliest failures; the outsider notices what the team takes for granted.
Further reading
Bring Argumont a Decision worth challenging.
Start with a paid 60 day pilot on your own real Decisions.