Home » Insights » Six steps to remove cognitive bias from management decisions

Six steps to remove cognitive bias from management decisions

Businesswoman drawing

The ways in which we perceive the world around us have been challenged by the BLM movement, and an increasing acceptance across society that structural inequities need to be confronted. Leaders need to be aware of how their own biases can be managed and understood so that they can make better judgements about their actions. Tobias Baer, a Contributor on programmes at The Møller Institute, examines ways in which we can all counter our personal biases every day.

Thinking of our colleagues, we see cognitive biases every day – the overconfidence of project managers promising IT migrations faster and cheaper than anyone before, the interest bias of colleagues resisting change, the anchoring of this year’s budget in last year’s, the overoptimism leading to disastrous mergers, and so on. Luckily, we ourselves tend to have the experience, self-awareness, and mental strength to avoid such cognitive biases.

Think again – do we really? Actually, cognitive biases are no character flaw – quite to the contrary, they are hardwired into the human brain and as necessary as the oxygen in the air for our brains to make the tens of thousands decisions small and large we face every day. Each of us has them – but because one of them is overconfidence, we just tend to not be aware. Issues arise only if our brain takes these shortcuts for important decisions where these biases hurt us, our company, or other stakeholders.

Here are six simple techniques I use myself to counter typical biases when making important decisions:

  1. Pre-mortem. This powerful bias-buster is a simple question: “If my decision turned out to be wrong, what would be the most likely reason?” Or if I am about to reject a candidate: “What hidden strengths am I most likely to overlook?” The answer very often informs the next two steps.
  2. Smash overconfidence with data. If I have a strong hypothesis (maybe even think something is a done deal), I try to find some data that contradicts my hypothesis with a quick, often dirty analysis. For example, if I believe “everyone likes to read my blog,” then it’s helpful to find out what percent of my connections actually read it.
  3. Bust the confirmation bias. If the data conveniently supports my hypothesis, then I ask myself whether there could be an alternative explanation. We usually look at aggregated data but often the underlying distribution of that tells a very different story. For example, if the statistics show that your customers, on average, hold 1.7 credit cards then it’s easy to conclude that practically everybody has a credit card, which may influence your product design. A closer inspection of the data might reveal, however, that 40 percent of customers have no credit card at all and that the average was heavily skewed by prepaid gift cards or fuel cards for a huge corporate fleet.
  4. Be your own devil’s advocate. Other people have this wonderful ability to see problems where you see only opportunity. If you don’t have a critical person on hand to challenge you, then try to think of a slightly obnoxious person that tends to disagree with you and ask yourself “What would this person say now?” (Note: if such a person doesn’t exist in your life, you may be at the receiving end of the sunflower effect where everyone follows the leader, in which case extra care is advised!)
  5. Use the wisdom of the crowd. Following the masses is a double-edged sword as you could fall victim to the herd effect – but I do usually look at what others are doing and ask myself how likely it is that so many people got it wrong if their actions contradict my hypothesis. For example, if I am smitten by a new gadget, online reviews can provide a healthy dose of skepticism. If, however, all the people writing 1-star reviews used the gadget in a different way to my own intended use, then I can make a much better informed decision to buy it anyway.
  6. Run a trial. We’ve covered many, though by no means all of the most common biases so far, but if I still think I’m at risk then I have one last golden rule: whenever possible, run a trial first. In a business context, it may be faster to trial something in a mock setup than to spend weeks or months on analysis and debate in an attempt to find the “perfect” answer. It also works in your private life: if you fear that embarking on a two-week cruise could trigger marital strife, why not tag a 1.5-day cruise onto your next traditional holiday trip as a trial?

These are my six simple, DIY-style steps to bust the bias. Do a pre-mortem and a quick data analysis; then challenge your data’s distribution and let both the devil’s advocate and the crowd speak up. And a trial closes the deal.

Related content

Illustration of a number of heads

Belonging and Leadership

picture of ARM leadership

Universal lessons from Army leadership