I have sat in executive workshops where the team spent three hours on anchoring and confirmation bias and nodded through a slide deck on better judgment. Ten minutes later they were back in the monthly operating review, staring at a cross-sell league table and a bonus dashboard nobody wanted to question. That is how cognitive biases in business usually survive.
The room needed permission to question the target and a process that forced the hidden assumption into the open before status settled around it. Bias training stays popular because it leaves the target alone, which suits the sponsor who set it and the executives being paid to hit it.
Cognitive biases in business are predictable errors in organisational judgment caused by hidden assumptions that go untested inside targets and status hierarchies.
Cognitive biases in business are usually built into the target
In business, bias often arrives before the meeting starts, because the scorecard already tells people what counts as success. The cleanest modern example is Wells Fargo. The U.S. Department of Justice settlement says the bank paid USD3 billion after employees opened millions of accounts or products without customer authorisation, and says senior leaders knew about "gaming" as early as 2002. That misconduct ran until 2016. Sixteen years is not a lapse in self-awareness. It is an institution defending the wrong frame.
Once success is defined as product volume, every later discussion bends around it. Complaints become noise and dissent looks disloyal. Branch staff learn quickly that a customer without another product is a failure, while a dubious extra account looks like initiative. What many people call confirmation bias is often something larger and uglier: managers teaching the organisation which evidence is welcome because their rankings and pay depend on the target surviving. If the decision had been framed around customer value rather than cross-sell numbers, much of the later rationalisation would have died on the table. That is why I start with Frame the decision. At Wells Fargo, the cross-sell target framed the room before anyone opened their mouth.
Cognitive biases in business survive when challenge is ceremonial
Boards fail on bias when disagreement is treated as theatre rather than interruption. The APRA prudential inquiry into Commonwealth Bank of Australia described complacency "from the top down" and a consensus style that reduced constructive criticism. The sharpest line in that report was the missing "should we?" question. Plenty of organisations ask whether they can hit a target. Far fewer ask whether the target deserves obedience when reopening it would expose the executives who sold it to the board, and the directors who applauded it in the minutes (always easier than admitting it should have died earlier).
The same weakness appeared at Carillion. The parliamentary report said the company entered liquidation with nearly GBP7 billion in liabilities and just GBP29 million in cash, while the board stayed focused on bonuses and flattering accounts of the business. Non-executives and auditors were present, but presence is not challenge. Once a board starts protecting the story that made it successful, contrary evidence has to fight status as well as optimism. I do not get much comfort from tidy labels like halo effect bias. The harder question is who in the room can stop sign-off before the nice story eats the company.
Why cognitive biases in business survive training
They survive because training asks individuals to inspect their minds, while the organisation keeps rewarding the preferred answer. Hino Motors is a blunt example. The Department of Justice resolution valued the global case at more than USD1.6 billion and said Hino improperly imported and sold more than 110,000 diesel engines in the United States from 2010 to 2022. Engineers fabricated test results, and software effects were kept out of sight. Nobody there needed a seminar on integrity. Certification had to "work" because production schedules and bonuses were sitting on top of it. The fiction was costly to challenge, which is why it lasted.
In my experience, that is what awareness sessions miss. They make bias sound private, something living inside a manager's skull. In practice the stronger force is public: the target is already on the wall, and the sponsor has already signalled which answer is safe. People do not need to become strangely irrational. They just need to read the room and protect the story the room is rewarding. Sorting that into System 1 and System 2 adds vocabulary. It does not add a checkpoint.
Once the answer is fixed, evidence stops being evidence and becomes raw material. I have watched that happen in boards and audits for years. A bad number appears and is explained away before the meeting moves on. Roger Estall and I wrote Deciding because the failure was usually embarrassingly plain: nobody had written down the assumption carrying the preferred answer, and nobody had named the margin miss or audit finding that would reopen it. Overcoming cognitive biases is a structural job. In the Universal Decision-Making Method, the step that matters here is Recognise assumptions. Name the bet. Write the trigger that kills it.
The fix is to test the decision, not the personality
My practical fix is plain. Frame the decision around the real purpose, not the internal target, then write down the assumption carrying the answer and the review trigger that reopens it. If the assumption is that customers will tolerate a price rise, say the churn number that proves you were fooling yourself. If the assumption is that a control issue is isolated, say the next audit result that sends the decision back upstairs. That is where cognitive biases in decision making stop looking psychological and start looking organisational. If the room cannot say this in simple language, it is not deciding. It is defending a belief system with better stationery.
I also want an audit chair or non-executive with actual standing to stop sign-off when the assumption starts wobbling. I do not care whether that person is popular. I care whether the process gives them the power to delay the paper and make the sponsor answer in public. If challenge is left to good manners, the senior voice wins and everyone calls it alignment, which is a handsome word for obedience. Give challenge a recognised seat, and the room is far more likely to catch the assumption it was about to smuggle through.
When I talk about cognitive biases in business, I mean a process willing to humiliate the target before the target humiliates the firm.
You could enter your next review and still let the target think for you.
Work through your decisionNo sign-up. Just pick your decision and start.
Grant Purdy is the co-author, with Roger Estall, of Deciding (2020), and the architect of the Universal Decision-Making Method.