Skip to main content
McKinsey Quarterly

Bias busters: Premortems: Being smart at the start

Conducting a premortem can help teams anticipate project failure—and head it off at the pass.

Despite their best intentions, executives fall prey to cognitive and organizational biases that get in the way of good decision making. In this series, we highlight some of them and offer a few effective ways to address them.

Our topic this time?

Premortems: Being smart at the start

The dilemma

Your company just finished launching a parking app for a large US city—but lots went wrong along the way. Development and rollout were delayed because financing and system updates took longer than expected. Pilot tests revealed unanticipated flaws in the software and the physical infrastructure. The app hadn’t been configured for all computing platforms, for instance, and sensors embedded in parking areas in some parts of the city failed to communicate with central servers that fed the app. A postmortem session showed exactly where and when the project went off the rails. Why couldn’t the team have seen these things up front?

The research

There are lots of well-documented reasons why teams avoid considering potential problems at the outset of a project or initiative. Studies show that project leaders overwhelmingly tend to be overconfident. 1 The plans they’ve mapped out are reasonable, and every step is plausible—why worry? Additionally, the start of a project is typically the time of greatest harmony among team members. Bringing up problems can seem obstructionist and disloyal. In fact, research also shows that most individuals are afraid to speak out against the group and explicitly identify problems with a plan. 2 Even if a project leader asks for honest critiques, team members often hold back to protect political, organizational, or personal interests. Everyone desperately wants to believe in the plan they are getting ready to carry out.

The remedy

To ensure that projects get the scrutiny they need, teams should conduct a “premortem.” 3 This is an exercise in which, after a project team is briefed on a proposed plan, its members purposefully imagine that the plan has failed. The exercise prompts everyone to review the plan and anticipate potential threats and hurdles. The very structure of a premortem makes it safe to identify problems. Under this approach, the psychology is flipped, and blind support for ideas gives way to creative problem solving. In fact, we’ve seen team members compete to see who can raise the most worrisome issues, and those team members are admired for their foresight, not ostracized.

One technology company used this approach when designing a new advanced-analytics system for an aviation program. Before the project launch, the project leader (with support from the project sponsor) gathered the team in a conference room and asked them to peer into “an infallible crystal ball,” looking six months into the future. Bad news: the project was a flop.

The project leader asked each team member to take two minutes to write down thoughts on why the plan had failed. He then asked each person, in turn, to share one reason for the failure. (The project leader went first to model behaviors and assure everyone that the meeting was about honest disclosure.) All the answers were captured on a whiteboard.

After three rounds of disclosure, so multiple ideas could be recorded and everyone’s opinions could be heard, the potential pitfalls became apparent. The biggest issues were organizational and cultural ones—for instance, getting the resources and senior-level sign-offs needed to design, build, and roll out the advanced-analytics system quickly and countering key stakeholders’ resistance to having to learn a new system. Once the project team had identified the potential vulnerabilities, it conducted another reflection exercise—this time, discussing the things it could do to mitigate the issues listed on the whiteboard. The end results were a stronger plan and a more resilient team that was more aware of the challenges it was facing.

Research shows that premortems reduce teams’ overconfidence significantly more than other critiquing and risk-analysis methods do. 4 The process lets teams identify a wide range of potential stumbling blocks, many of which hadn’t been considered before. And it helps to forge a culture of candor: uncomfortable truths can be spoken without repercussion but instead with gratitude for courage and cleverness.

About the author(s)

Tim Koller is a partner in McKinsey’s New York office; Dan Lovallo, a senior adviser to McKinsey, is a professor of business strategy at the University of Sydney; and Gary Klein is a cognitive psychologist who developed the premortem method of risk assessment in 1992 and helped to establish the naturalistic decision-making movement in 1989.

Related Articles