Confirmation Bias

How we come to believe things that aren’t true.

This is part of the series on how our biases affect our decision making and what you can do about it.

What is Confirmation Bias?

Confirmation Bias, happens when we look for evidence that supports our views or twist data to ‘prove’ what we think, ignoring data that doesn’t fit our world view.

As with almost all biases it is caused by your brain trying to conserve energy. It is very tiring to keep changing our minds and very pleasing to be certain of something. It feels good to know that you are right and great to tell a receptive audience about it.

This is obviously a problem when running a company. The problems manifest in multiple ways but the Confirmation Bias is at its most dangerous when mixed with HIPPO’s.

HIPPO = Highest Paid Person's Opinion

What is a HIPPO?

Generally a senior person who can overrule the data + other people's ideas. As discussed in our Illusory Truth post, single minded leaders who are good at rallying the troops have a place in every organisation but quickly become an issue when their blinkered view creeps into day to day operations.

In businesses it is easy to find a reason to explain why something performed poorly without digging too deep into the causes. For example, a discount campaign might not work because: “it was targeted at the wrong audience, the creative was sub par, it went to the wrong landing page”.

HIPPOs and Bubbles go well together

The truth is a tenuous thing. Sometimes a truth can be ‘believed’ into existence. We experience this in marketplace ‘bubbles’ where everyone believes something until they are rudely proved wrong, the Tulip Bubble in Amsterdam, Expedition Bubble in Europe, Stock Market in America & Housing bubble most everywhere. Annie Duke, the author of Thinking in Bets, would call this “resulting” which is where a cause is attributed to an outcome without testing if it is correct. E.g. Ice cream vans cause shark attacks as they both go up in summer, you can find more spurious correlations here. Defaulting to causation over correlation is a baddy in many bias stories.

Another way to believe something into existence is the rule of opposites. A good example of this is how our irrational beliefs can get strengthened into phobias.

For example, if you are scared of deep dark water > you avoid it > nothing bad happens. After a few times of this it stands to reason that if you go in the water > something bad will happen.

“If this campaign didn’t work lets try a completely different one”, “we don’t need to optimise the design lets do a whole new design”

Another regular villain is the ‘Umwelt’ our brains create for us. Much like in the Matrix, you are not always perceiving reality as it is. Without telling you, your brain is often filling in blanks with stuff it has made up.

A good way to experience this is through ‘saccades’. Will you are reading this post you ‘jump’ from the end of one line to the beginning of another. Your brain can’t handle the input so it sticks up a blurry image as a placeholder. Try it out!

Obviously, it is very difficult for us to consciously be aware of this and tell our brains to not do it. It is like trying to convince your brain not to be fooled by a visual illusion. Your brain holds onto the core and pulls in the rest from heuristics (shortcuts) & biases based on your worldview. Have an argument with your significant other/ friend and note how convinced each side is of their photographic memory.

Stare at the centre and watch the problems sort themselves out

Do the grey dots really exist?

How does Confirmation Bias affect data programmes

Too many people take the truth for granted

Organisational design is a huge problem so we start to look for explanations that are easier to comprehend.

Sales are not going well because:
IT takes ages to respond to any ticket
Leadership doesn’t set a clear strategy
Ecommerce Team constantly shifts priorities
Marketing doesn’t get TikTok
Product team doesn’t listen to customers

We have a really hard time not giving ourselves an A+ on a task; I am late because I had numerous valid reasons, that person is late because they are lazy. 75% of drivers think they are above average. As we will see this problem compounds with experience/ success.

The Peter Principle

There is a lot of the Peter Principle in businesses - people get promoted to the level of their incompetence. Therefore it is better for them to stick to what they know, to stay safe within the walls of their teams.

If they are challenged by data showing the sales are lacking they will often attack the trustworthiness of the data, or claim their next idea will turn things around or demand that the data tell them how to fix it.

This means that programmes continue as they ever were. For large companies this takes a while, and often a paradigm shift, for this to be a problem. Toilet paper and paint can be sold in much the same way as 40 years ago.

For startups this can be death, or at least the dreaded growth slap. Startups, like careers, need to change, to evolve if they are to grow. They need to walk out into the unknown and start all over again, and do it often and be right most of the time. Most of them suffer from a lack of talent or talent constrained by bias in the organisation design.

Use the scientific method to turn your hunches>hypotheses>tests>learnings>actions

Structure your company to turn your data>insights>initiatives>performance

Learn to spot biases

Do a pre-mortem & post mortem on ideas. Before the meeting try to find out why a test may fail, after the test or initiative send out the results in a standard format then get everyone to write down why they think it happened, what could be improved for next time and what is out of peoples hands.

The key thing here is to write it down. Take a leaf out of Amazon’s book, who start every meeting with a 6 page document that people read in silence before discussing the numbers.

When people talk out loud it allows the most well spoken, confident, senior people to dominate. It rewards talking fast, sounding right and short sentences.

The data and insights come from reflection and often multi step, intertwined causes. We get overwhelmed in meetings and want an easy fix. Prevent this with data ahead of time, force people into the habit of assessing the whole gamut of reasons for something or drill down the focus of the meeting into one specific area.

Focus on how to get more reliable data. Don’t guess and move on. If you want to know how the menu is performing, add tracking so you can see what people are clicking and work to keep that data reliable. Ask users about the menu. Do A/B tests on the menu.

A key piece of organisational design is to focus on getting more granular and contextual data as you grow and your optimisation programme matures.

Don't run from the tough questions

Confirmation Bias supporting cast

When doing analysis and insight beware of the following biases and call them out where possible.

Data without context can be dangerous due to the Clustering Illusion bias which causes even the best of us to overestimate the importance of small clusters or patterns in our data or experiences.

You can spot this in phrases such as - "This is the second week in a row that this has happened" or “Thursday afternoons are the best time or this”

This is often set off by the Anchoring Effect, relying too much on the initial piece of information offered when making decisions.

"The first test seemed OK. Do we need to do any more?", “that other influencer wanted 4x as much so this is fine”

Outside the realm of data be careful of the impact of blogs and tweets causing the Availability Bias, Overestimating the importance and likelihood of events given the greater availability of information.

“We need to get on Tiktok”, “We need to build a specific landing page”

All of these things can build into what I call the Bandwagon effect (a HIPPO reads a blog, finds a bit of supporting data from a politically minded helper and the Confirmation Bias gets underway) which can be very hard to bring down once it gets going.

“This should have worked, it doesn’t make sense that it doesn’t”, "The whole department knows there's no problem here." You can often see varying types of ‘ostriches’ hiding from bad news aboard the bandwagon.

Drill down into the problem

The 5 Y’s/ First principles: Help get to the core of the problem & belief. A first principle is a basic assumption that cannot be deduced any further. Over two thousand years ago, Aristotle defined a first principle as “the first basis from which a thing is known.”

The 5Y’s was reportedly first developed by Toyota who used it to get to the root cause of issues and utilise the knowledge of the people actually doing the job rather than just those in the boardroom.

As James Clear says “In practice, you don't have to simplify every problem down to the atomic level to get the benefits of first principles thinking. You just need to go one or two levels deeper than most people. Different solutions present themselves at different layers of abstraction.”

"Consider the opposite" strategy. In every discussion try to come up with 3 hypotheses for why it could have happened. You can also try physically switching up your location to help shake out some biases - go offsite, do a walk & talk if it is just the two of you.

Face to face, specific time meetings are a breeding ground for combative loud talking and distractions which lead to biases.

Great insights come from consistently, obsessively, writing things down

Consistency means that you waste less mental energy trying to analyse the data. You can focus on the causes and what you are going to do about it. Having a clear, written record before, during and after makes it harder to bias to creep in. Having people write down the 5 y’s or a contrarian view makes them more likely to internalise it.

Once you have your baseline of insights you can work to expand them not constantly recreate them. Work hard on maintaining data quality but accept that it won’t be perfect. Remind people that not having perfect data isn’t an excuse to ignore all new evidence.

Confirmation Bias does not happen because people are lazy or lack the ability to assimilate new information. It happens because they fail to ask the right questions or question their own beliefs.

Start creating your hypotheses in digging into the deeper causes today.

I hope you enjoyed this post is you have anything to add or any questions let me know!

Page updated

Google Sites

Report abuse