Cognitive Biases in Decisions

Cognitive biases are systematic errors in thinking that affect every person, regardless of intelligence or experience. Daniel Kahneman and Amos Tversky identified the foundational patterns in their 1974 paper "Judgment under Uncertainty: Heuristics and Biases," work that later earned Kahneman the Nobel Prize in Economics. The key insight: biases are not character flaws or signs of poor thinking. They are features of how the human brain processes information under uncertainty. Over 200 biases have been catalogued by researchers. This page covers the 10 most relevant to team decision-making in business settings.

Biases are bad enough for individual decisions, but they become far worse in teams. Groups don't cancel out individual errors. They amplify them. Authority bias suppresses dissent (nobody challenges the boss). Groupthink creates false consensus (everyone nods because everyone else is nodding). Anchoring cascades through the room (the first number mentioned becomes the baseline for all subsequent estimates). Cass Sunstein and Reid Hastie documented this amplification in "Wiser: Getting Beyond Groupthink to Make Groups Smarter" (2014). Their finding: teams make worse decisions than the average team member would make alone, precisely because social dynamics turn individual biases into group behaviors.

The counter-strategy is not "be less biased." That is impossible. The human brain is not a software bug you can patch. The counter-strategy is process design: build decision processes that reduce the impact of biases before they affect the outcome. Anonymous voting before discussion neutralizes authority bias, groupthink, and bandwagon effect in one move. Independent scoring before group review prevents anchoring. Devil's advocate roles force the team to consider evidence against the preferred option. Each bias card below includes a specific counter-measure tied to a concrete process change, not just awareness.

A product team at a 45-person B2B company held their quarterly planning meeting to choose between 4 feature candidates for Q3. The VP of Product opened the meeting: "I talked to our biggest customer last week and they specifically asked for Feature A, the advanced reporting dashboard. I think this should be our top priority."

Three biases fired in sequence. Bias 1 (Authority + Anchoring): The VP's statement combined two biases at once. As the most senior person in the room, their opinion carried disproportionate weight (authority bias). And by naming Feature A first with a customer story, it became the reference point for the entire discussion (anchoring). Every subsequent comment was positioned relative to Feature A.

Bias 2 (Groupthink): Two engineers had concerns about Feature A (it required a data pipeline rewrite that would take 4 months, not the "a few weeks" the VP estimated). Neither raised the concern because contradicting the VP in front of the team felt risky.

Bias 3 (Confirmation Bias): When someone mentioned that a user survey from last month showed Feature C (improved onboarding flow) was the most requested improvement, the VP dismissed it: "That survey had a low response rate." The team accepted this reasoning without noting that the "biggest customer" was one data point while the survey had 340 responses.

The PM suggested running the prioritization with anonymous scoring. Each team member rated all 4 features on Impact (1-5) and Effort (1-5) independently.