A premortem is a planning exercise where the team imagines a project has already failed, then each person writes down why. Gary Klein, who developed it, says it can take as little as 20 to 30 minutes: brief the plan, declare it a fiasco, give everyone two minutes to list reasons, then go round the room collecting one reason each until you have a list of showstoppers to fix.
Most kickoff meetings end with someone asking whether anyone sees any problems, and a room full of people saying no. The premortem was built to break that silence. Klein first described it for a wide audience in Harvard Business Review in 2007, under the title Performing a Project Premortem, and it's since spread to boardrooms, firefighting teams and the military. Here's how to run one well.
Why it works
Klein's explanation is about social pressure. In a normal kickoff there's a silent pressure not to raise doubts that might disturb the team's harmony. The premortem reverses that: once the imaginary crystal ball has shown a definite failure, people show how smart they are by the quality of the problems they raise. Dissent stops being disloyal and becomes the task.
It also changes how people think. Instead of asking what might go wrong, which invites vague answers, you tell them it did go wrong and ask them to explain it. Klein's point is that team members then draw on their own experience to imagine how the failure happened.
How to run a premortem, step by step
- Brief the plan. Run it at the end of a kickoff, once everyone understands the plan.
- Declare the failure. Tell the team you're looking into an imaginary, infallible crystal ball, and the plan has turned out to be a fiasco. Treat it as certain.
- Two minutes of silent writing. Each person writes down the reasons the plan failed, on their own.
- Round the room. Each person gives their top reason, then the next person gives a different one. Write them on a whiteboard. Klein suggests the team leader goes first, to set an example of candour.
- Keep going until the reasons run out. You now have a list of showstoppers.
- Two more minutes on fixes. Klein added this later to stop the session feeling discouraging: each person identifies actions they can take on some of the items.
- Rate the problems. A more recent refinement has the team assess each problem for likelihood, impact and ease of prevention, so you know where to act first.
Prompts to use in the room
The wording matters, so script it. These prompts follow Klein's crystal ball framing:
- "It's six months from now. This project has failed badly. Write down every reason why."
- "What went wrong that we could see coming today but didn't say?"
- "Which assumption in the plan turned out to be false?"
- "Which person, supplier or system let us down, and how?"
- "What did we run out of first: time, cash, people or attention?"
- "Now: what's one thing you personally could do this week to stop your top reason happening?"
The mistakes that ruin it
Klein and two colleagues, Paul Sonkin and Paul Johnson, have listed the common errors, and each one quietly turns the premortem back into an ordinary meeting:
| Mistake | Why it hurts | Do this instead |
|---|---|---|
| Asking "What could go wrong?" | Too soft; people defend the plan | State that the plan has failed, as a certainty |
| A slow, leisurely pace | Energy drops and people self-censor | Survey the room fast, one item per person per round |
| One senior person reads out their whole list | Everyone else goes quiet | Take turns, one reason each, around the room |
What the research says, and what it doesn't
Klein's HBR article leans on a 1989 study by Deborah Mitchell, Jay Russo and Nancy Pennington on prospective hindsight, imagining that an event has already happened. It's often quoted as showing that this improves the ability to correctly identify reasons for future outcomes by 30%.
Here's the catch a searcher wouldn't guess. Jason Collins points out in his evidence review that the 1989 study found prospective hindsight increased the number of reasons people generated by about 30%. It didn't assess the quality of those reasons. More ideas is useful, but it's not the same claim.
The direct test is smaller. Klein describes a study with Veinott and Wiggins in which 178 college students critiqued a university's plan for an H1N1 flu outbreak. Those in the premortem condition showed the greatest reduction in overconfidence, compared with simple critiques or listing pros and cons. Collins notes the academic literature overall is thin, so treat the premortem as a cheap, sensible habit rather than a proven cure.
Premortem vs post-mortem vs risk register
| Premortem | Post-mortem | Risk register | |
|---|---|---|---|
| When | Before work starts | After the project ends | Throughout the project |
| Question | Why did it fail? (imagined) | Why did it fail or succeed? (real) | What might happen, and how likely is it? |
| Main benefit | Surfaces hidden doubts early, when plans can still change | Lessons for next time | Ongoing tracking and ownership of risks |
| Weakness | Based on imagination, not data | Too late to help this project | Often filled in once and forgotten |
They work best together. Use the premortem to find the risks, record them in a register with owners, and review them after the project ends. For a broader look at assessing risk, see our risk management basics.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Variations worth knowing
Klein's training company, ShadowBox, teaches three variations: Racking and Stacking for prioritising failure modes, a Success Scan for goal-driven planning, and a Double-Barreled approach that weighs the risks of acting against the risks of doing nothing. The last one is useful for small business owners, where the bigger danger is often a good plan that never starts.
My advice
Run a premortem on anything that would hurt if it failed: a launch, a hire, a big contract, a move to new software. Keep it to half an hour, script the opening line so nobody softens it to "what could go wrong", and turn the top three reasons into tasks with names next to them before people leave the room. Then put those tasks straight into your project plan so they're tracked like any other work.
Frequently asked questions
What is a premortem?
It's a planning exercise where a team imagines a project has already failed and each person writes down why. Gary Klein developed it and described it in Harvard Business Review in 2007.
How long does a premortem take?
Klein says it can take as little as 20 to 30 minutes, including two minutes of silent writing and a round of the room to collect reasons.
What is the difference between a premortem and a post-mortem?
A premortem happens before work starts and imagines the failure, so the plan can still change. A post-mortem happens after the project ends and looks at what really happened.
Does the premortem have research behind it?
Some. A 1989 study found imagining an event had already happened increased the number of reasons people generated by about 30%, though it didn't measure their quality. A later study with 178 students found the premortem reduced overconfidence more than simple critiques or pros and cons.
What is the biggest mistake when running a premortem?
Asking "What could go wrong?" instead of declaring that the plan has failed. Klein also warns against a slow pace and letting one senior person dominate.
