How to run a project pre-mortem meeting
A pre-mortem is a short meeting at the start of a project where the team imagines the project has already failed and writes down why. Run it after the plan is agreed and before work starts. Leave with a ranked risk list, an owner and due date for each mitigation, and a date to revisit it.
What a pre-mortem is
Psychologist Gary Klein described the method in Harvard Business Review in 2007. Team members assume the project they are planning has just failed, then generate plausible reasons for that failure.
The point is to make it safe to speak up. Klein writes that one reason projects fail is that people keep their reservations to themselves during planning. A pre-mortem asks for exactly those reservations, while the plan can still change.
How it differs from a postmortem and a kickoff
A postmortem looks back at an incident that really happened. A pre-mortem looks forward at a failure that hasn't happened yet, while the plan can still change.
A kickoff agrees the project's purpose, scope and owners. The pre-mortem tests that plan. Klein runs it as part of the kickoff, once the plan has been discussed. On a large project, you can give it its own slot.
When to hold it and who to invite
Hold it after the plan exists but before the team commits to it. Atlassian's Team Playbook runs it before a project starts, and suggests a shorter version during sprint planning.
- Size: Atlassian suggests 3 to 11 people.
- Who: the people doing the work, plus anyone who knows where similar projects went wrong.
- Time: Klein says it can take as little as 20 to 30 minutes. Atlassian's version takes 60.
- Materials: a whiteboard and sticky notes, or a shared document for remote teams.
A 60-minute pre-mortem agenda
This agenda combines Klein's steps with the voting and discussion steps from Atlassian's play. Shorten the discussion block if you only have half an hour.
| Time | Step | What happens |
|---|---|---|
| 5 min | Set the scene | Recap the plan. Then ask people to imagine looking into a crystal ball: the project has failed. Not slipped, but a fiasco. |
| 5 min | Write silently | Everyone writes down the reasons it failed, alone. Klein gives this two minutes. |
| 15 min | Go around the room | Each person reads their top reason, one per turn, with no repeats. The leader goes first. Capture every reason where all can see it. |
| 5 min | Group and vote | Merge duplicates. Each person gets three votes for the biggest threats. |
| 25 min | Plan the top risks | Take the top three. For each, judge likelihood, impact and ease of prevention, and agree at least one action. |
| 5 min | Recap actions | Read back every action with its owner and due date. |
Facilitate it well
Klein names three common mistakes. Each one quietly turns the pre-mortem back into an ordinary risk review.
- Asking "What can go wrong?" instead of stating that the project has failed. Klein calls that the non-challenging question.
- Letting the pace drag. Move quickly around the room to keep energy up.
- Letting one person, often the most senior, read out their whole list. Take turns, one reason each.
What to leave with and how to follow up
The meeting should produce a short, owned plan, not just a list of worries. Atlassian notes that giving each action an owner and due date makes it more likely to get done.
- A risk list: every reason raised, with the top risks marked.
- For each top risk: likelihood, impact and how easy it is to prevent.
- Mitigations: at least one action per top risk, each with one owner and a due date.
- Follow-up: store the list where the team can find it, add actions to your project tracker and revisit the list as the project runs.
A hypothetical worked example
For example, a seven-person team plans to move customers to a new billing system over one quarter. The project lead opens: "It's the end of the quarter, and the migration has failed."
After two minutes of writing, the round robin surfaces nine reasons. The three with the most votes are: invoices went out wrong in the first month, the finance team wasn't trained in time, and the old system's data needed far more cleanup than planned.
| Risk | Mitigation | Owner and due date |
|---|---|---|
| Wrong invoices in month one | Run both systems in parallel for one billing cycle and compare totals | Engineering lead, before cutover |
| Finance not trained | Two hands-on sessions using test accounts | Finance manager, two weeks before cutover |
| Data cleanup underestimated | Sample 200 customer records now and size the cleanup | Data analyst, end of week two |
Keep the risk plan from getting lost
After a meeting in Bridge, the structured notes include a summary, decisions and follow-ups. Action items assigned in the meeting also appear in the Action items list on Home.
You can create a custom recap template, for example with sections for risks, owners and mitigations. Then file the meeting in a shared space, so people with access can reopen the list when the team revisits the plan.




