How To Fight Groupthink

Groups of decision makers tend to engage in groupthink, an overemphasis on harmony and consensus. This can get in the way of examining all the options objectively, leading to weaker—and sometimes disastrous—decisions. Consider the failed Bay of Pigs invasion of Cuba during U.S. president John F. Kennedy’s administration. Arthur Schlesinger Jr., one of Kennedy’s advisers, wrote this about his participation in the debate leading up to the humiliating defeat of U.S.-backed Cuban exiles trying to overthrow the regime of Cuban leader Fidel Castro: “In the months after the Bay of Pigs I bitterly reproached myself for having kept so silent in the Cabinet Room…I can only explain my failure to do more than raise a few timid questions by reporting that one’s impulse to blow the whistle on this nonsense was simply undone by the circumstance.”                                                                    

A variation on this failing occurs when participants don’t speak up because they feel the subject under discussion does not fall into their area of responsibility or expertise. At one global agriculture company, the members of the executive committee tended to speak up during strategy conversations only if their area of the business was being discussed. The tacit assumption was that colleagues wouldn’t intrude on other colleagues’ areas of responsibility—an assumption that deprived the committee of their insights.

The weight of evidence strongly supports that decisions are better when there is rigorous debate. One research effort found that for big-bet decisions, high-quality debate led to decisions that were 2.3 times more likely to be successful.  Extensive study has explored the importance of vigorous debate in improving decision making.            

Ideally, a company dedicated to pursuing long-term strategic success should have a culture of dissent, where rigorous debate is the norm. But most companies need to take more active steps to stimulate debate. The key ingredient is to depersonalize debate and make it socially acceptable to be a contrarian. Here are some useful techniques:                                                                        

• Assigning a devil’s advocate. At a strategy discussion, assign someone the task of taking an opposing point of view. Make sure this contrarian’s contribution is more than just offering opinions. The focus should be on calling attention to potential alternate scenarios or highlighting missing information important to the debate.
• Bringing a diverse group to the discussion. More than 150 years ago, John Stuart Mill wrote in On Liberty, “The only way in which a human being can make some approach to knowing the whole of a subject is by hearing what can be said about it by persons of every variety of opinion.” More recent research has proven his point. Diversity means drawing on the opinions of people from different disciplines, roles, genders and races in important discussions. Bring in more junior people with special expertise, create an environment where it is safe for them to speak up and ask them for ideas.
• Encouraging debate with secret ballots. Use a secret ballot at the beginning of the debate, not the end. Once a proposal has been presented and before it is debated, ask participants to vote on the idea in secret. The request could be for a yes-or-no vote on a project or for a ranking of investment priorities. When the results are revealed, assuming participants discover at least one other person shares their views, the knowledge will likely make them more comfortable expressing their opinion.
• Setting up a red-team/blue-team activity for large investments.Arrange two teams to prepare arguments for opposing outcomes. While undertaking the preparatory work and analysis for this approach is expensive, it can make a difference for particularly large decisions with high uncertainty.

Confirmation Bias and Excessive Optimism

Confirmation bias and overoptimism are two distinct biases. However, the same set of techniques applies to both, so we discuss them together.
                                                                                                                               
Confirmation bias is the tendency to look for evidence that supports your hypothesis or to interpret ambiguous data in a way that achieves the same result. For business decisions, this often takes the form of “I have a hunch that investing in x would create value. Therefore, let’s look for some supporting facts that will back up our hunch.” The universal foundation of the scientific approach to addressing a hypothesis is the opposite: You should look for disconfirming evidence.                                                                                            

Overoptimism is the tendency to assume that everything will go right with a project, even though past projects tell us that such smooth outcomes are rare. A classic example is the construction of the famous Sydney Opera House, whose schedule and budget were both overly optimistic. The project was completed 10 years late and cost 14 times the original budget.                                                

Some of the techniques used to overcome groupthink, such as the use of opposing red and blue teams, can help here. The simplest approaches are to avoid developing hypotheses too early in the process and to actively look for contrary evidence. Other potential correctives for confirmation bias and overoptimism include the following two methods:                               

• Conducting a pre-mortem.A “pre-mortem” is an exercise in which, after a project team has developed a proposal, its members purposely imagine that the plan has failed. The very structure of a pre-mortem makes it safe to identify problems.
• Taking the outside view. This is an exercise that involves building a statistical view of a project based on a reference class of similar projects. For instance, a group at a private-equity company was asked to build a forecast for an ongoing investment from the bottom up—tracing its path from beginning to end and noting the key steps, actions and milestones required to meet proposed targets. The group was then asked to compare that ongoing investment with categories of similar investments, looking at factors such as relative quality of the investment and average return for an investment category. Using this outside view, the group saw that its median expected rate of return was more than double that of the most similar investments.                                                                                         

Inertia (Stability Bias)                                                                                   

Inertia, or stability bias, is the natural tendency of organizations to resist change. One study found that, among the companies it studied, spending allocations across business units were correlated by an average of more than 90 percent from year to year. In other words, the allocation of spending to business units essentially never changed. The same study showed that companies that reallocated more resources—the top third of the sample—earned, on average, 30 percent higher total shareholder returns (TSR) annually than companies in the bottom third of the sample.                                               

The solution to inertia bias is relatively straightforward. Rank initiatives across the entire enterprise. In addition, ensure that the budget you are building is rooted in the current strategic plan, not last year’s budget. The essential idea is to ignore as much as possible the influences of past allocations or budgets. In practice, you may not be able to shift resources as quickly or as much as you should. But trying to ignore the past is a starting point and will help you minimize inertia. 

Excerpted with permission from the publisher, Wiley, from Valuation: Measuring and Managing the Value of Companies by Tim Koller, Marc Goedhart, David Wessels. Copyright © 2025 by McKinsey & Company. All rights reserved. This book is available wherever books and eBooks are sold.