Alignment · Vocabulary
What Is Groupthink?
A leadership team that agrees quickly on something hard has either done the work or skipped it, and from the inside the two feel identical. Groupthink is the name for the second one.
Definition
Groupthink is the mode in which a cohesive group's drive for unanimity overrides its realistic appraisal of alternatives - the point at which preserving the group's agreement displaces evaluating the group's options as the real work of the meeting.
The term is precise, and it is worth keeping it that way. Groupthink is not a synonym for a bad group decision, for conformity in general, or for any room that happens to agree. It names a specific failure mode with a specific precondition: cohesion. The groups it describes are not weak, disengaged, or badly staffed. They are close, experienced, mutually loyal, and trying to get the answer right - which is exactly why the failure is hard to see from inside. Cohesion is an asset that becomes a liability at the moment the group starts protecting it. Nobody in the room decides to value harmony over accuracy. The group simply stops looking, and no one is assigned to notice.
The vocabulary comes from Irving Janis, who named the pattern while studying policy fiascoes: decisions made by capable advisory groups that went wrong in ways their own members could not afterward explain. What he produced was not a mood or a metaphor but a taxonomy of observable symptoms, which is what makes it useful to an executive rather than merely interesting. A leader cannot audit a group's state of mind. A leader can notice that nobody has argued the other side, that the risks discussed are all the ones everybody already knew, and that the meeting ended earlier than the decision deserved. Modern research works the same territory from the other direction: Amy Edmondson's psychological safety describes the climate in which a doubt can be said out loud, and psychological safety is now the more common lens on why candor does or does not appear. Janis's contribution is the symptom list, and the symptom list is what you can see on a Tuesday.
Four Symptoms Worth Knowing by Name
Janis catalogued more than four. Collective rationalization discounts the warning that would force a reconsideration; direct pressure on dissenters makes the cost of objecting personal rather than analytical; stereotyping of out-groups dismisses the competitor, the regulator, or the critic as too weak or too foolish to be worth modeling. The four below are the ones that do the most damage in an ordinary leadership meeting, and each has a tell.
Illusion of invulnerability
Shared, excessive optimism that encourages the group to take risks it would not take individually. The tell is asymmetry: the upside case is detailed and the downside case is a sentence. A group under this illusion does not ignore risk, it discusses risk briefly and confidently, which is a different and more convincing failure.
Self-censorship
Members withhold doubts they judge to be out of step with the emerging consensus. Nothing is suppressed by anyone; each person suppresses themselves, on the reasonable private grounds that raising an objection everyone else has apparently resolved will cost more than it is worth. The doubt does not go away. It goes offline, into the hallway conversation afterward.
Illusion of unanimity
Silence gets read as consent, so every person in a room of private reservations concludes that everyone else is on board. This is the compounding symptom: self-censorship manufactures the silence, and the illusion of unanimity converts that silence into evidence, which then justifies more self-censorship. A group can talk itself into confidence it never actually had.
Mindguards
Members who appoint themselves to protect the group from information that would disturb the consensus, usually while believing they are being helpful. The disconfirming analysis gets filtered before it reaches the room, the awkward customer call is characterized rather than played, and the leader hears the state of the world as the group has already decided it should be.
The Core Failure: False Harmony
Here is the part that matters to a leader running a company rather than a seminar. False harmony is misalignment wearing alignment's clothes. Real alignment is shared understanding that has survived scrutiny, and it is expensive to produce: it requires the objection to be raised, met, and either absorbed or answered. Groupthink delivers the appearance at a fraction of the cost, and organizations reward the appearance, because a leadership team that agrees quickly looks like a strong one. Every incentive in the building points at the counterfeit. Groupthink takes up how a room full of capable people reaches a conclusion none of them would defend alone.
The largest single accelerant is available to exactly one person in the room. A leader who states a preference first converts a decision meeting into a ratification meeting. After that sentence, no contribution is neutral: agreement is cheap and safe, disagreement is with the boss rather than with an option, and the discussion that follows can only produce one outcome, so it carries no information about whether that outcome is right. The leader then experiences a well-run meeting that closed on time with everyone aligned. That experience is the symptom. Why Nobody Speaks Up works through the private arithmetic each person runs before deciding whether an objection is worth its cost, and Why People Follow the Crowd covers the pull of the visible majority even on people who can see it is wrong.
In the Four A's of Organizational Readiness™, groupthink is an Alignment failure that disguises itself as Alignment strength, which is what makes it worth naming separately. Nothing on a dashboard distinguishes a team that agrees because it worked the problem from a team that agrees because it stopped. The distinction lives in the process that produced the agreement, so that is where it has to be measured and that is where it has to be fixed.
The Unanimity Test
Take the last consequential call your leadership team made unanimously and name the strongest argument against it, in the form the person who believed it would have put it. If you cannot, you did not hear that argument, which means either nobody in the room held it or nobody in the room said it. One of those is a staffing problem and the other is a conditions problem, and only one of them is common.
How to Design Against Groupthink
The instruction to speak up does not work, and it is worth being clear about why: it asks an individual to absorb a personal cost in order to fix a group problem, which is precisely the trade self-censorship already declined. Structure changes the trade instead of appealing to it.
The leader states a preference last
Not a style choice, a sequencing rule. The room forms and voices a view before it learns the answer, so what you hear is judgment rather than deference. It costs a leader the pleasure of framing the discussion and returns the only thing the meeting was for: information you did not already have.
Assign a critical evaluator
Give a named person the standing job of building the strongest case against the emerging option, rotating so it never becomes one individual reputation. The role converts objecting from an act of courage into an assignment, which is the entire point: the organization stops depending on someone being brave enough to volunteer.
Collect independent estimates before the meeting
Have participants commit their numbers, risks, and concerns in writing before anyone hears anyone else. Discussion pulls a group toward what everyone already shares and away from what only one person knows, and the unshared item is usually the one that would have changed the decision. Written first, spoken second.
Hold a second-chance meeting
After a preliminary decision and before it becomes irreversible, reconvene with one job: surface residual doubt. Sleeping on it does real work, and the reservation that felt unsayable against a fresh consensus is much easier to say against a decision that has already been provisionally made and can still be changed.
Research basis: Janis, I.L. (1972). Victims of Groupthink: A Psychological Study of Foreign-Policy Decisions and Fiascoes. Houghton Mifflin. Janis, I.L. (1982). Groupthink: Psychological Studies of Policy Decisions and Fiascoes (2nd ed.). Houghton Mifflin.
Frequently Asked Questions
What is groupthink?
The mode in which a cohesive group's drive for unanimity overrides its realistic appraisal of alternatives. Cohesion is the precondition rather than the fault: the groups Irving Janis studied were capable and were trying to get the decision right, and protecting the agreement displaced evaluating the options without anyone choosing it.
What are the symptoms of groupthink?
Janis named observable symptoms, not a mood: illusion of invulnerability, collective rationalization, self-censorship, illusion of unanimity, direct pressure on dissenters, mindguards, and stereotyping of out-groups. Self-censorship and the illusion of unanimity compound each other, since private silence gets read across the room as consent.
What causes groupthink?
Conditions rather than character: a cohesive group, insulated from outside views, without an agreed method for evaluating options, under time pressure. The strongest accelerant is a leader who states a preference first, which converts a decision meeting into a ratification meeting and makes the agreement that follows uninformative.
How do you prevent groupthink?
Structurally. The leader states a preference last; a named critical evaluator argues the opposing case as a duty rather than an act of courage; participants commit independent estimates and concerns in writing before the meeting; and a second-chance meeting surfaces residual doubt before the decision becomes irreversible.
Diagnose Your Alignment Condition
Does your team agree, or has it stopped looking?
The Executive Diagnostic run as a team evaluation compares what each member reports privately against what the group says together - and the gap between those two answers is the measurement groupthink is designed to hide.