⚰️Pre-mortem

A meeting held before a project starts, in which everyone assumes it has already failed. The method is the psychologist Gary Klein's.

1. Describe the plan in ten plain lines.
2. Say: it is six months from now and this has failed. Not might fail. Has.
3. Everyone writes, alone and in silence, for five minutes: why did it fail?
4. Go round the room, one reason each, until the lists are empty.
5. Pick the few that are both likely and expensive. Each gets one line of then we would.

Why it works better than asking for risks: what could go wrong? invites loyalty. Why did it go wrong? invites explanation, and people are good at explaining.

Use when: the plan is firm enough to describe and loose enough to change.

Example: 10 March, the day before launch, with Priya Raman. Two pages. One line read we have to switch a feature off in the night. Twenty-eight hours later we did, in 📂 Exceptions Inbox general availability, and agreeing to it took four minutes.

When I am alone I let Pre-mortem prompt produce the obvious twenty and then write my own five. The principle underneath is older than the meeting: ⚓ Rehearse the loss. The private evening version is 🪷 Rehearsing what could go wrong.

When not to use it

• The night before, as a substitute for deciding. On 10 March it was nearly too late to change anything. It earned its hour only because the list included what we would do.
• When the real risk is already known and unwelcome. In January Nadia Rahimi said the load test was too gentle. No exercise was needed to find that. I needed to listen.
• Without the last column. A list of failures with no then we would is worrying with a pen.