Lesson 5 of 7
The pre-mortem
Before buying, assume the investment has already failed and write down why. Each cause becomes something to monitor.
By Javier Sanz
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
In one minute
- A pre-mortem imagines the investment has already failed, three years out, and asks why.
- Assuming failure surfaces doubts that optimism during research tends to hide.
- Each imagined cause becomes a specific condition to monitor, tied to a filed figure.
Before buying, imagine that three years have passed and the investment has clearly failed. Write down every reason you can think of. The exercise works because it assumes failure as a fact, which removes the reluctance most people feel to raise a doubt before a decision is made.
The pre-mortem, in order
- 1Imagine the investment has already failed
- 2List every plausible cause, without editing
- 3Rank the causes by damage and likelihoodKeep the top two or three
- 4Turn the top causes into conditions to monitor
Why assuming failure works
Klein's premortem method asks a team to imagine, before a decision is made, that a plan has already failed, then to spend a few minutes writing down why. His argument was that this simple reframing surfaces objections that a straightforward "any concerns?" does not. A team member raising a doubt about a decision not yet made can look like disloyalty, while explaining an assumed failure does not. Kahneman later endorsed the method for exactly this reason: it gives legitimacy to the doubts that overconfidence and social pressure otherwise suppress.
Do it alone, in writing, before the group talks
Klein's format has each person write down every reason for the imagined failure alone, on paper, before anyone speaks. Voicing an idea first tends to anchor everyone else's list on it. Writing first, then sharing, gets a wider set of causes than opening straight into discussion.
From an imagined cause to a testable condition
A cause of failure is only useful once it points at something you can actually check later. "Competition could get worse" is a worry. "Revenue growth falls below 5 percent in a fiscal year" is a condition. The pre-mortem's real output is not the list of fears. It is the short set of figures, drawn from what the company files, that would move first if any of those fears came true.
Do it before the purchase, not after
Run it before capital moves, while you can still decide not to buy or to size smaller. On a holding you already own, run it at each review, knowing ownership pulls you toward defending it.
Check yourself
What does a pre-mortem ask you to assume?
- That the company's management is untrustworthy
- That the investment has already failed, three years from now
- That the investment will probably succeed
Show the answer
Answer: B.
A pre-mortem starts from failure as a settled fact and works backward to the causes, which is what makes it different from an ordinary risk list.Why does assuming failure surface more doubts than simply asking 'any concerns?'
- Because it requires more time to complete
- Because it makes the exercise feel more urgent
- Explaining an assumed failure carries less social cost than raising a doubt about an undecided plan
Show the answer
Answer: C.
Klein's argument was that voicing a doubt before a decision can look like disloyalty, while explaining a failure already assumed to have happened does not.What should each cause identified in a pre-mortem turn into?
- A condition tied to a filed figure
- A reason to avoid the investment entirely
- A general note of caution
Show the answer
Answer: A.
A cause of failure is only useful once it points at a figure you can actually watch. "Growth could slow" becomes a specific yearly threshold to monitor.Sources
- Klein, 'Performing a Project Premortem', Harvard Business Review, 2007.
- Kahneman, Thinking, Fast and Slow, 2011, on the premortem as a check on overconfidence and social pressure.
What this leaves out
- The worked example uses round, invented causes and figures. A real pre-mortem must be run before an actual purchase, not reconstructed afterward.
- A pre-mortem only surfaces failure modes someone in the exercise can imagine. It will miss causes outside that experience.
- Nothing here is advice about any security.