Cognitive Biases: What Are They and How Do They Explain How Our Brain Works
Cognitive biases are shortcuts our brain uses to process information. Which ones matter most for product design, and how to mitigate them in practice.
Part of the guide Inclusion and Diversity
Did you know that 95% of cognitive biases occur without us being aware of them?
They can be understood as shortcuts our brain uses to process information around us.
Sometimes, these shortcuts aren’t helpful and lead to incorrect information processing, which we call judgment distortion.
Currently, more than 188 cognitive biases are known. These deviations or errors can appear in various forms. The most common are related to information processing through shortcuts, distortions in how we store events in our memory, limitations in our brain’s processing capacity, social influence, or emotional shifts.
Understanding cognitive biases is important for us as designers because we are constantly creating products and services in an era where “dark patterns” are on the rise, cognitive effort is increasing, and unconscious human error is becoming more frequent. This is especially concerning in environments like hospitals, where human lives can be at risk.
How can we use this knowledge to our advantage
Learning about cognitive biases is just as essential for understanding the people who will use our products and services as it is for understanding our stakeholders. Many stakeholders make decisions with biases behind them, often leading to poor choices on user experience.
We can use this knowledge from both sides: better understand the user, and better understand (and influence) internal decisions.
The seven biases you’ll meet
In product, seven biases show up frequently. Worth being able to recognise them, in yourself and in others.
1. IKEA Effect: we attribute more value to what we built with our own hands than to what we bought ready-made. That’s why IKEA sells you furniture you assemble. In product, this translates to: users value features they customised more. Useful for retention, dangerous for discovery of new features that feel “of the product” rather than “of the user”.
2. Gender Bias: assuming characteristics based on gender. Medical studies show pain is treated differently in men and women without scientific basis. In product: assuming women use health apps more than men, or that men don’t want emotional-care features. The data shows otherwise.
3. Bystander Effect: no one acts because everyone assumes someone else will. In teams, it shows up when no one escalates a problem because they assume someone with more authority will. The fix is direct: name explicit owners for each decision.
4. Confirmation Bias: looking for evidence that confirms what we already think, ignoring what contradicts. The most dangerous bias for research: you Google for data that backs your initial hypothesis. The classic mitigation is the alternative hypothesis: force yourself to look for contrary evidence before concluding.
5. Group Attribution Error: judging an entire group from one person’s impression. “I talked to one user from Lisbon who said X, so Lisbon users think X.” No. The mitigation is sample size, and care with generalisations in presentations.
6. Halo Effect: the first impression colours everything that follows. In research interviews, in stakeholder meetings, in hiring. If the person looks good in the first 30 seconds, the rest is interpreted favourably. If they look bad, the opposite. The fix: structure evaluation against separate criteria before the encounter.
7. Misinformation Effect: the way a question is asked changes the memory of the answer. In legal studies, asking “how fast were the cars going when they smashed?” produced higher estimates than “how fast were the cars going when they hit?”. In research, word choice changes significantly what users report.
How to mitigate (in practice)
Biases don’t go away. They’re automatic processes of the brain. But there are four tactics that reduce the impact:
1. Reflect on past decisions. Look at design and research decisions from the last 6 months and ask: was it my decision, or was it the brain on default? This periodic audit reveals patterns.
2. Include external viewpoints. Bias is harder to see from inside. Asking review from someone in another team, another background, another culture, is one of the cheapest and most effective mitigations.
3. Challenge your own viewpoints. When you’re convinced about something, consciously try to build the opposite argument. Not to convince yourself, to check if the conviction survives.
4. Don’t decide under pressure. Maybe the most important. When we’re pressured, especially emotionally, the brain goes on autopilot. Decisions under pressure are nearly always worse. If pressure is real, the first action is to slow down, not to decide.
The connection to inclusion
Most of the inclusion problems we see in product don’t come from bad faith. They come from cognitive biases operating in cascade. The Halo Effect leads us to hire people similar to us. The Group Attribution Error leads to generalisations about users. The Confirmation Bias leads to research that only confirms the initial hypothesis.
Knowing the biases isn’t abstract. It’s a practical skill for anyone designing products for others who aren’t the same as them.
More on the background in the Inclusion and Diversity guide and in the ethical principles in AI design. On how team diversity is the first line of defence against these biases, see Inclusive design teams.
