Reviewing Decision Processes to Uncover Bias

Definition

Rather than rely on awareness of biases to prevent them, executives can conduct a structured review of the decision-making process that produced a recommendation. A 12-question checklist examines three domains: questions the decision maker should ask themselves (self-interest, emotional attachment, groupthink), questions to pose to the recommenders (saliency bias, confirmation bias, availability bias, anchoring bias, halo effect, sunk-cost fallacy), and questions about the proposal itself (overconfidence, disaster neglect, loss aversion). This approach shifts from individual judgment to organizational discipline, recognizing that while individuals cannot eliminate their own biases, independent reviewers using a systematic process can detect and mitigate biases in others' thinking.

In the Book

Daniel Kahneman, Dan Lovallo, and Olivier Sibony argue that awareness of biases has done little to improve decision quality. Knowing that anchoring bias exists doesn't prevent a person from becoming anchored. The problem is that most cognitive biases operate unconsciously—we don't catch ourselves in the act of making an intuitive error. They use the distinction between System One thinking (fast, intuitive, effortless) and System Two thinking (slow, reflective, deliberate) to explain why: Most decision-making is System One. We construct a narrative, suppress alternative stories, and we're not consciously aware of the distortions at work.

The breakthrough is to move from the individual decision maker to the decision-making process and the organization. While an individual cannot reliably spot their own biases, an independent reviewer can use System Two thinking to scrutinize a team's System One thinking. The 12-question framework operationalizes this. For example:

• Question 1 asks whether the recommending team has self-interested motives (empire building, bonus incentives, career risk mitigation). The reviewer probes: Are their interests materially affected by the outcome?
• Question 5 tests for confirmation bias: Have credible alternatives been considered, or only the preferred option?
• Question 10 addresses overconfidence: Is the base case overly optimistic? Should the team build an "outside view" case using historical analogues rather than just the "inside view" of this specific project?
• Question 11 introduces the "premortem": Imagine the worst has happened; what story explains how we got here? This forces consideration of risks the team hasn't consciously anticipated.

The authors show three corporate executives—Bob, Lisa, and Devesh—applying these questions to different types of decisions. The discipline isn't in getting the "right" answer but in shifting the conversation from "Is this a good idea?" to "How was this idea developed, and did the process account for common sources of bias?"

Why It Matters

Most organizations rely on hope—hope that experienced leaders will make good judgments, that smart people will catch errors, that market discipline will eventually correct mistakes. But hope is not a strategy. By institutionalizing a structured review process based on known sources of bias, organizations can improve the quality of major decisions without requiring leaders to be superhuman. A McKinsey study found that companies that systematically reduced bias in their decision processes achieved investment returns seven percentage points higher. The difference isn't in having fewer biases; it's in having a process that catches and mitigates them before execution.