In 1960, Peter Wason designed an experiment that has shaped cognitive psychology for six decades. Participants were given the number sequence 2-4-6 and asked to discover the underlying rule by proposing their own sequences. Most participants hypothesized "ascending even numbers" and tested sequences like 8-10-12 and 14-16-18. Every test confirmed their hypothesis.

The actual rule was simply "any ascending numbers." Sequences like 1-2-3 or 5-100-999 would also have fit. But almost no one tested those — because testing a sequence that might disprove the hypothesis felt unnecessary. Why look for evidence you're wrong when evidence you're right is so easy to find?

This is confirmation bias. And it is the single most expensive cognitive error in business strategy.

A company decides that its next growth market is Southeast Asia. The strategy team researches the region and returns with a compelling case: rising GDP, growing middle class, favorable demographics, increasing digital adoption. The board approves. Resources are allocated.

What the strategy team did not present — because they did not look for it — was the disconfirming evidence. Regulatory complexity that has stalled three competitors' expansion efforts. Distribution challenges that add 40% to supply chain costs. Cultural preferences that make the company's product category a poor fit for local consumption patterns.

This information existed. It was publicly available. It simply was not sought — because the team was looking for evidence that supported the decision they wanted to make, not evidence that challenged it.

The result, three years and $12 million later, was a quiet withdrawal from the market. The board never saw the disconfirming data because the team never assembled it. The team never assembled it because they were not looking for reasons to say no. They were looking for reasons to say yes.

Confirmation bias does not just affect strategy. It affects which data you look at, how often you look at it, and what you do when you see it.

A marketing director who believes email marketing is the most effective channel will check email metrics daily and social media metrics weekly. The emails may be performing adequately. Social media may be outperforming. But the frequency of attention — driven by the director's existing belief — ensures that email performance is always top of mind and social media performance is always peripheral.

When contradictory data does surface, confirmation bias has a second defense: reinterpretation. "Social media drove more conversions this quarter" becomes "that's an anomaly" or "the attribution model is probably wrong" or "let's wait for a full quarter of data." Each of these responses is reasonable in isolation. In aggregate, they form a pattern of systematically dismissing evidence that challenges the existing belief.

The most senior people in an organization are often the most susceptible — not because they are less intelligent, but because they have more invested in the current strategy. They approved it. They championed it. Their reputation is attached to it. Confirming evidence validates their judgment. Disconfirming evidence threatens it.

Individual awareness of confirmation bias is insufficient. Knowing the bias exists does not prevent it from operating — any more than knowing about optical illusions prevents you from seeing them. The defenses must be structural.

Pre-mortem analysis, developed by Gary Klein, inverts the confirmation dynamic. Before a decision is implemented, the team is asked: "Assume this decision failed spectacularly. What went wrong?" This forces the team to generate disconfirming scenarios — the very scenarios confirmation bias would otherwise suppress. The exercise does not prevent bad decisions, but it surfaces risks that confirmatory thinking would have hidden.

Red team exercises serve a similar function. A designated group is tasked with building the strongest possible case against the proposed strategy. Their job is not to be objective — it is to be adversarial. The collision between the advocating team and the red team produces a more complete picture than either could generate alone.

Data dashboards should be designed to surface contradictory signals, not just confirmatory ones. If the primary KPI is customer acquisition cost, the dashboard should also display customer churn prominently. If the primary metric is revenue growth, the dashboard should include margin trends. Presenting both together forces decision-makers to confront the full picture rather than selecting the view that supports their position.

The most valuable information in any business is the information that makes leadership uncomfortable. The customer feedback that contradicts the product roadmap. The competitive analysis that suggests the moat is narrower than assumed. The financial trend that challenges the growth narrative.

This information is rarely absent. It is rarely even hidden. It is simply unexamined — because examining it requires confronting the possibility that the current direction is wrong. And confronting that possibility is exactly what confirmation bias is designed to prevent.

The leaders who build durable businesses are not the ones who avoid confirmation bias — that is not possible. They are the ones who build systems that force disconfirming evidence to the surface, and then have the discipline to act on it. The data you ignore is always the data that matters most.