Judgment Playbook

Pre-mortem

Imagine the plan has already failed, and find out why while you can still change it.

When you need itThe plan feels safe and you cannot say how it would fail6 minute readtaught in course chapter 9
Manage uncertaintyExecute and learn

What it is

Your team is about to approve a plan. The leader says: "It is a year from now, and the plan has failed. Write down why." Everyone writes alone, in silence, before anybody speaks. That is a pre-mortem. Teams usually look for the cause of a failure afterwards, in a post-mortem. In a pre-mortem they look before, while the plan can still change.

The exercise is short. You end with a list of specific causes of failure, the most likely first. You also agree on at least one change: to the plan, or to what the team will watch.

Where it comes from

The method comes from Gary Klein, a psychologist who studies how experts decide when time is short. He described it in Harvard Business Review in September 2007, in "Performing a Project Premortem". In January 2021 he wrote in Psychology Today that he had developed it three decades earlier, to cut problems in his company's projects.

The idea behind it is older. Researchers call it prospective hindsight: imagining a future event as if it has already happened. Klein's article traces it to a 1989 paper in the Journal of Behavioral Decision Making. Its authors were Deborah J. Mitchell of the Wharton School, J. Edward Russo of Cornell and Nancy Pennington of the University of Colorado. Klein says the trick made people 30% better at finding reasons for future outcomes. The paper's authors claimed less. They found that people who were sure of an outcome explained it differently from people who were not. Moving the event into the past mattered less.

Later Beth Veinott, Klein and Sterling Wiggins tested the exercise itself. They asked 178 people to review a plan for responding to a pandemic. Some ran a pre-mortem. Others simply criticised the plan, listed pros and cons, or listed only cons. The pre-mortem group lost the most confidence in the plan. Klein sees this as overconfidence falling, and that is the point of the exercise.

What it corrects

A plan has just been presented. The manager behind it clearly believes in it, and the meeting is nearly over. Someone in the room has a concrete doubt. Perhaps the plan depends on something they know is fragile, or a deadline clashes with another promise. They say nothing.

They stay quiet for two reasons. First, the leader asked "Any concerns?" To answer, people have to judge the plan, not picture how it fails. So they say "risky" instead of naming a cause. Second, speaking up costs something. The doubt sounds like criticism of a colleague's work. And if the project goes well, they will look foolish in public.

Working harder fixes neither problem. A careful person studies the plan more closely. But that plan is built on the very assumptions they ought to question. Klein changes the question instead. Once the failure is announced as a fact, nobody has to judge the plan. People only have to explain what happened, and experts do that quickly and in detail. Naming a cause is no longer an attack on a colleague. It is the job everyone in the room was given.

How it works

  1. Present the whole plan, so that everyone argues about the same plan.
  2. Announce, as a fact, that by a stated date the plan has failed. For example: the launch slipped nine months, and the product was withdrawn.
  3. Give everyone two or three minutes to write causes alone, in silence, including the ones they would normally keep to themselves.
  4. Go round the table, one cause per person, the leader first. Write every cause down before anyone argues about one.
  5. Merge the duplicates. Then rank the causes by how likely they are, not by how dramatic they sound.
  6. Turn each top cause into a change. Prevent it, fund a backup plan, or set a tripwire: a named warning sign that one person watches.

Worked example

Klein's 2007 article describes a project to give advanced planning algorithms to military officers who plan air campaigns. One team member had sat silent through a long kick-off meeting. In the pre-mortem he named a cause of failure: one of the algorithms was too much for the laptops used in the field. The software would run for hours, but the planners needed an answer in minutes. So the project would be impractical.

Then the developers admitted that they had already built a shortcut, a simpler version of the algorithm, but had not wanted to mention it. The team swapped it in, and the project succeeded. Nobody discovered anything new in that room. The team already knew about the laptops and the shortcut, and in the exercise people said so. Klein says the exercise takes twenty to thirty minutes.

In a Business Case Weekly case

The Rolls-Royce case ends with a decision about money in mid-2020. Revenue is down by about a third, the loss is heading for £4 billion, and 9,000 job cuts have been announced. The company must choose how to raise money. It can sell new shares to its shareholders at a deep discount (a rights issue). It can borrow more, ask the state for support, or sell parts of the business.

Do not aim the pre-mortem at the pandemic, which has already happened. Aim it at the mix of options you choose. Pick a date eighteen months ahead. Say that by then the company had to ask investors for money a second time, on worse terms. Then name the part of your own mix that caused it. Write down causes that only this company could have. How does it get paid? What will its promises to airlines cost in cash? And in each option, who is left holding the risk?

In your answer

  • "Say it is [date] and this decision has failed. The most likely cause is ..."
  • "That failure comes down to one assumption: ..."
  • "The first sign of it would be ..., and I would expect to see it by ..."
  • "So I would change the plan now by ..., and keep ... in reserve."

Common misuse

The usual mistake is the disaster list: a page of general catastrophes that gets written and filed away. People call it a pre-mortem, but it does not help. A cause that could hit any project tells you nothing about this one. You can spot a disaster list with two questions. Could the cause you wrote appear unchanged in a competitor's plan? If so, it is not specific enough yet. Did any line on the list change the plan, the budget or what you watch? If not, the meeting made people anxious and decided nothing.

References

Practise it on a case

More on thisWikipedia explainer

Red team

Next method

Hand the plan to a separate group whose only job is to break it, and let them succeed.

Open the next method