You've seen it happen. A room full of brilliant people somehow arrives at a decision so obviously flawed that you wonder if anyone was actually thinking. The merger that destroyed value. The product launch that ignored clear warning signs. The strategy that everyone privately thought was doomed but nobody challenged.

Here's the uncomfortable truth: intelligence doesn't protect groups from making terrible decisions. In fact, it often makes things worse. The same cognitive horsepower that helps individuals solve complex problems can backfire spectacularly when those individuals gather in a conference room. Understanding why reveals something important about how we can actually harness collective wisdom—instead of collective foolishness.

Imagine you're the fourth person to speak in a meeting. The first three people—all respected colleagues—have endorsed the same approach. You have doubts, but now you're facing a choice: share your concerns and risk looking foolish, or assume that three smart people probably know something you don't. Most of us fold. And that's how information cascades begin.

Cascades happen when people rationally decide to ignore their private information and follow the crowd. The problem is that early opinions carry disproportionate weight. If the first two speakers happen to share the same view—even if it's wrong—everyone else updates their beliefs accordingly. Your dissenting evidence gets buried under social pressure and reasonable-sounding logic. The cascade doesn't care how smart you are. It exploits your intelligence by making conformity seem rational.

What makes this particularly insidious in high-powered groups is that smart people are often better at constructing justifications. They can talk themselves into agreement more eloquently. They're also more sensitive to reputational concerns—looking foolish in front of other smart people carries real social cost. The result? The smarter the room, the faster and more completely the cascade can take hold.

Takeaway

Early voices don't just influence decisions—they can silence the very information that would have prevented mistakes. The sequence of who speaks matters as much as what they know.

Here's a frustrating pattern: put six experts in a room, each holding unique pieces of a puzzle, and watch them spend most of their time discussing information they all already share. The critical insight that only one person holds? It barely surfaces. This is shared information bias, and it's remarkably persistent across contexts.

The psychology is straightforward. Shared information gets repeated, validated, and reinforced. When you mention something everyone knows, heads nod and you feel smart. When you bring up something only you know, you might get blank stares. There's no social reward for introducing unfamiliar ideas—and considerable risk that your unique contribution gets dismissed or misunderstood. Groups naturally gravitate toward common ground, even when the gold is buried in individual knowledge.

This bias turns teams into echo chambers for conventional wisdom. The whole point of assembling diverse expertise is to surface perspectives that no individual possesses alone. But without deliberate intervention, groups default to pooling what's already shared while the genuinely novel insights—the ones that could change the decision—remain locked in individual heads. The tragedy is that everyone leaves the meeting thinking they had a thorough discussion.

Takeaway

The most valuable information in any group is often what only one person knows. But without structure, groups systematically ignore precisely the knowledge they were assembled to surface.

The good news is that group dysfunction isn't inevitable. It just requires what psychologist Daniel Kahneman calls 'decision hygiene'—structured processes that protect against predictable errors without requiring anyone to be superhuman. The key insight is that process can accomplish what willpower cannot.

Start with independent evaluation. Before any group discussion, have each person write down their assessment privately. This locks in their genuine views before cascades can form. Then structure the conversation to explicitly surface unique information—ask each person what they know that others might not. Designate a devil's advocate, but rotate the role so it's not performative. And critically, have the most senior person speak last. Status cascades are real, and the boss's early opinion can shut down dissent faster than any other factor.

These interventions feel awkward at first. They slow things down. They require discipline to maintain. But the alternative is collective intelligence that's actually collective stupidity wearing a suit. Groups can genuinely outperform individuals—but only when they're designed to aggregate independent information rather than amplify early consensus. The structure is the strategy.

Takeaway

You can't think your way out of group bias in the moment—you have to design around it in advance. Good process is what lets smart people actually be smart together.

Intelligence is necessary but not sufficient for good group decisions. Without the right structures, smart people will cascade toward consensus, discuss what everyone already knows, and suppress the dissent that could save them. The failure mode isn't stupidity—it's the social dynamics of how humans share information.

The fix isn't to hire smarter people or try harder. It's to build processes that protect against the predictable ways groups go wrong. Write before you speak. Surface unique knowledge deliberately. Let the senior person speak last. Design for wisdom, and you might actually get it.