A post-mortem looks at what went wrong after a project ends. A pre-mortem does it before you start. The team imagines that it is a year from now and the project has failed, then works out why. It sounds gloomy, but it is one of the most useful conversations a team can have.
The psychologist Gary Klein described the technique in the Harvard Business Review in 2007. Imagining that failure has already happened makes it easier for people to raise doubts they might otherwise keep to themselves, especially when a plan has a lot of enthusiasm or a senior sponsor behind it. Use it once a plan is agreed in outline, but before you commit time and money.
Method
Remind the group of the plan, set the scene of failure, collect reasons individually, share and group them, then agree what to change.
Step 1: Remind everyone of the plan
Spend five minutes summarising the project or decision: what it aims to achieve, how and by when. Make sure everyone has the same picture before you start.
Step 2: Set the scene
Say something like: “Imagine it’s a year from now. We went ahead with this plan exactly as it is, and it has been a disaster. Not a small setback, a real failure. What happened?”
Step 3: Write reasons alone
Give everyone five to ten minutes to write down every reason they can think of for the failure, one per sticky note, in silence. Encourage them to include the uncomfortable ones, such as “we didn’t really have the skills” or “nobody actually wanted this”.
Step 4: Share in rounds
Go round the group, with each person reading out one reason at a time and adding it to the wall, until all the notes are up. Ask the person leading the project to go last, so their reasons don’t shape everyone else’s.
Step 5: Group and prioritise
Cluster similar reasons together and give each cluster a name. Then ask everyone to place three dot stickers on the risks they think are most likely or would do the most damage.
Step 6: Plan what to change
For the top three to five risks, agree what you will do: change the plan, add a check, find more information or accept the risk knowingly. Give each action an owner and a date, and record them where the team will see them again.
Outcome
A stronger plan, with the biggest risks identified and addressed before they happen. The team also gets permission to voice doubts in a constructive way, which builds trust and makes it more likely that people will speak up when something starts to go wrong later.
Tips for Facilitators
- Make it safe to be negative: explain that the exercise only works if people are honest, and thank people for the difficult points, not just the easy ones.
- Get the sponsor on board: brief the project lead or senior sponsor beforehand, so they welcome criticism of the plan rather than defending it.
- Keep it about the plan, not people: reasons should describe what happened, such as “the training came too late”, not who was to blame.
- Look beyond the obvious: prompt people to think about people, money, time, technology, the wider world and the people the project is meant to help.
- Don’t stop at the list: a pre-mortem is only useful if it changes something. Leave at least a third of the time for actions.
Variations
- Pre-parade: run the same exercise imagining wild success, and ask what made it go so well, to find what to protect and build on.
- Quick check: for a small decision, give people three minutes to write their top reason for failure, then discuss the top three.
- Prioritise the actions: plot the actions on an Impact/Effort Matrix to decide which to do first.
- Follow it up: revisit the risks and actions at each review, for example in a Decisions and Actions Review.
Accessibility and Online Sessions
Let people write their reasons on paper, type them or say them aloud to a partner who writes them down, and read every note out when it goes on the wall. Some people find imagining failure stressful, especially if their job depends on the project, so keep the tone practical and remind everyone that it’s a thinking exercise. Online, use a shared whiteboard, give people time to add their notes privately before revealing them all at once, and use the whiteboard’s voting tool in place of dot stickers.
