Here’s a small mental exercise. Think of something you’re currently planning — a project at work, a trip, a career move, a purchase, a conversation you’re preparing for. Imagine it is now two years in the future, and the plan has failed catastrophically. Not just badly — catastrophically. The result was worse than almost anything you had expected at the start.
Now, before reading any further, spend about a minute listing all the reasons why it failed. What happened? What went wrong? Which of the things you’re now worrying about turned out to matter? Which of the things you’re not worrying about turned out to be the real problem?
This small exercise is called a pre-mortem, and it was developed by a research psychologist named Gary Klein. Klein, whose career was spent studying how experts make decisions under time pressure — firefighters, paramedics, military commanders — was struck by a particular pattern in organisational life. Bad plans almost always had warning signs before they went wrong. People inside the organisation often even knew those signs were there. But the warnings were rarely voiced clearly in advance, only named later, in the post-mortem, after the disaster.
Klein’s insight was simple. If the warnings existed before the disaster but were silenced by the social dynamics of a committed team, perhaps they could be surfaced by a small procedural trick. Instead of asking a team to evaluate their plan prospectively (which tends to produce polite approval and anchored thinking), ask them to imagine it has already failed, and then explain why.
The result, when people actually do this, is often startling. Reasons for potential failure that had been vague or unspoken suddenly become easy to articulate. Team members who would have hesitated to voice concerns in a “what do you think of this plan” conversation are willing to voice them as diagnoses of an imagined future. The pre-mortem gives people permission to say what they already suspected.
The endorsement from the biggest name in the field
Pre-mortems got their largest boost from an unlikely source — Daniel Kahneman, the Nobel laureate whose research on cognitive biases has shaped much of behavioural economics. In interviews and in his book Thinking, Fast and Slow, Kahneman has described the pre-mortem as one of the most practical debiasing tools he knows, and has endorsed it warmly.
What makes the technique work, Kahneman argues, is that it does something specific to the human mind. Most decision-making happens in what he calls System 1 — the fast, intuitive, overconfident mode of thinking that generates plans we then become attached to. Once we’ve generated a plan, we tend to defend it. Asking ourselves to imagine it failing in the abstract doesn’t break this defensiveness. But imagining the specific scene of its failure — a particular future moment, a specific scenario — triggers a different cognitive process. You’re no longer evaluating whether the plan will work; you’re backward-engineering why it didn’t. The mind switches from defence to detective.
Klein’s own later work suggests the technique is most effective when the imagined failure is made concrete. Not “imagine this plan has failed” — but “imagine it’s eighteen months from now. You’re walking out of a meeting where the senior leadership told you the project is being shut down. What did they say in that meeting? What specifically had gone wrong?” The specificity matters. Vague failure produces vague diagnoses. Concrete failure produces concrete ones.
The expert-intuition counterweight
Klein’s career has another important strand, though, that makes his enthusiasm for pre-mortems more interesting than it first appears. His broader research is on what he calls recognition-primed decision-making — the phenomenon where experienced experts make excellent decisions very quickly, without formal analysis, by pattern-matching against years of experience.
The firefighter who intuitively senses, without being able to articulate why, that the building they’re in is about to collapse. The paramedic who knows, at a glance, which of the wounded needs immediate attention. The chess grandmaster who sees the right move in seconds. These people aren’t ignoring information — they’re processing enormous amounts of it through pattern recognition that formal analysis cannot replicate in the available time.
The implication is subtle. For novices, the pre-mortem is almost unambiguously useful — it adds structure and catches errors that inexperienced minds wouldn’t have flagged. For experts, the picture is more complex. Formal pre-mortem exercises can sometimes override the faster, often-correct pattern recognition that expertise has painstakingly built. A surgeon who has done a procedure five hundred times, asked to do a pre-mortem for each case, may perform worse than one who trusts their trained judgement.
Klein himself has written about this tension. His best-known book, Sources of Power, makes the case for expert intuition. His work on pre-mortems supplements rather than contradicts it. The two frameworks are meant to apply to different situations: pre-mortems for novel, complex, stakes-heavy decisions where your pattern library is thin; trained intuition for the repeated, familiar decisions where your pattern library is deep.
Where pre-mortems can go wrong
A smaller body of research, coming from behavioural medicine and organisational psychology, has documented a less commonly-named downside of pre-mortem thinking. If you do it excessively, or for decisions that don’t need it, it can produce overcautious paralysis — the mirror image of the overconfidence it was designed to correct.
A team that does a pre-mortem on every decision, no matter how minor, starts generating reasons why everything might fail. The reasons aren’t wrong; there are always things that could go wrong. But the cumulative weight of identified risks can make good plans feel too dangerous to pursue. This is particularly a risk for anxious or risk-averse individuals, for whom the pre-mortem amplifies rather than corrects an existing cognitive bias.
Russo and Schoemaker, in their research on executive decision-making, make a related point. They distinguish between decision traps — cognitive errors that degrade decisions — and decision paranoia — the pathological expansion of debiasing techniques that degrade decisions in the opposite direction. Both are real. Both are costly. A team stuck in decision paranoia makes worse decisions than one making some occasional overconfident errors, because the paranoid team is usually making no decisions at all.
The practical principle from this research is that pre-mortems, like most debiasing tools, have a specific range of usefulness. They work best for:
Novel decisions, where your pattern library is thin and biases are more likely to mislead.
High-stakes decisions, where the cost of overconfidence is large.
Group decisions, where pluralistic ignorance and groupthink can suppress individual concerns.
They work less well or actually backfire for:
Routine decisions, which you’ll do quickly anyway and where the formal analysis adds no value.
Decisions where the expertise is already deep, where trained intuition outperforms deliberate analysis.
Decisions where the failure modes are too numerous, and generating a list mostly demoralises rather than clarifies.
How to actually do one
If you want to run a pre-mortem on a real decision, here is the version that research and practice both suggest works best.
Start with the plan clearly stated. Not “our plan” generally — but the specific plan, with specific timeframes, specific actions, specific expected outcomes.
Project yourself forward to a specific date — ideally the date by which the plan was supposed to have succeeded. Imagine that, at that date, the plan has failed. Not partially succeeded, not produced a mixed result — failed. Outright. In some clear and obvious way.
Now, from that imagined future, write out what happened. Be specific. Not “we ran into difficulties” but “the supplier delivered late, and by then we’d committed to the launch date, and there was no time to recover”. Not “the relationship didn’t work” but “we discovered, once we’d moved in together, that we had incompatible expectations about how weekends should be spent, and neither of us was willing to flex”. Specificity is where the insight lives.
Write out as many of these imagined explanations as you can. Klein’s research suggests that teams typically identify 30-40 per cent more potential problems in a pre-mortem than in a standard “what are the risks?” discussion. The pre-mortem unlocks concerns that prospective framing suppresses.
Then — and this is the critical step — actually use what you’ve found. Many teams run the pre-mortem and then file it. The point is to go back to the plan and revise it. Which of the imagined failures can you actually prevent? Which can you detect early enough to respond? Which are unavoidable — and if so, are you still comfortable with the plan’s expected value?
The question that remains
The most important thing the pre-mortem teaches you isn’t specific to any decision. It’s the habit of treating your current plans with the same critical eye you’ll eventually have about them in hindsight. Your future self, looking back from whatever happens, will see your current plans more clearly than you do now. The pre-mortem is an attempt to borrow some of that clarity in advance, for the small cost of a few minutes of imagination.
The question to carry, for any plan important enough to be worth doing at all:
If this plan failed two years from now, what would be the most embarrassingly obvious reason — the one you wish, in retrospect, you had taken seriously — and what’s stopping you from taking it seriously today?
Key research referenced: Gary Klein’s work on pre-mortems (Harvard Business Review, 2007) and on recognition-primed decision-making (Sources of Power, 1998); Daniel Kahneman, Thinking, Fast and Slow (2011); Russo and Schoemaker on decision traps (Winning Decisions, 2002).