Mirror, Mirror On The Wall: Is Confirmation Bias Skewing My Data-Driven Decisions? - Lift361

Our Thinking

Mirror, Mirror On The Wall: Is Confirmation Bias Skewing My Data-Driven Decisions?

By Ed Higdon. Posted in Uncategorized

July 30, 2015

Retailers turn to big data as a source of objective information. Through numbers, companies can learn how well they stack up to their competition, the ways in which their customers think and approach the buying journey, and gain insight into hidden trends. Numbers themselves are objective. They have no bias or agenda. It is when humans begin to analyze those numbers and use them to drive decision making that things can become subjective. At the heart of that conundrum lies confirmation bias.

Confirmation bias is the conscious or unconscious tendency for a person to seek information that upholds their pre-existing opinions. It leads an individual to ignore information that does not support his or her viewpoints, whether that information is positive or negative. Confirmation bias can prevent us from considering critical information when making decisions, and that failure can lead to poor, ineffective or undesirable outcomes. The good news? Confirmation bias can be avoided, with a bit of practice.

How Can You Avoid Confirmation Bias?

Where any data reporting is involved, confirmation bias is sure to follow. It does not mean that the people looking at the data are manipulative or they cannot understand the information presented. It is merely human nature to seek and find information that supports our own beliefs.

Even though confirmation bias is in our nature, it is also critical to take steps to avoid that trap. The next time you look at your big data to answer questions and drive decisions, keep the following in mind:

  1. Ask the Right Question From the Start. Often, we phrase our questions in a way that allows us to get the answers we are seeking, regardless of what the data actually says. Start at the top to determine your overarching goals, and then decide precisely which key performance indicators (KPIs) must be measured in order to help you paint an accurate, objective picture.
  2. Check Your Ego at the Door. When it comes time to sit down and consume data and information, leave your ego behind. Nobody likes to be wrong, but nobody is right all of the time, either. Approach the exercise with the attitude that the truth is more important than being proven right.
  3. Actively Seek Differing Opinions. Always look at data with a group of peers. Specifically seek out those who disagree with you and ask them provide a detailed explanation of their reasoning, and then listen to and consider those viewpoints. Make the session about sharing ideas, not arguing or proving a case. A healthy discussion almost always leads to new insights.
  4. Remove Caution From the Discussion. If you could make a decision based upon the data in front of you with absolutely no consequences, which track would you take? Caution and fear can often lead to confirmation bias.
  5. Put the Data Down. When you’ve had time to discuss the data with the group, give yourself time to mull over that discussion. Revisit the information after a day or two and see if you feel differently, or if you notice anything new.

Even The Best And Brightest Are Not Immune To Confirmation Bias

It is very easy to fall prey to confirmation bias. Even the world’s top scientists can allow their own biases to creep into their research. In 2015, Dr. John Ioannidis penned a paper entitled, Why Most Published Research Findings Are False, in which he examined confirmation bias among 49 highly regarded medical research findings. According to his paper, 45 of those widely accepted studies showed evidence of bias.

Even billionaire investor Warren Buffet is not immune to confirmation bias. According to Forbes, Buffet once wrote a piece for Fortune magazine where he discussed the lengths he goes to avoid such a bias, including aggressively seeking a variety of contradictory data and opinions before making any major investment decision.

In order for big data analytics strategies to be effective, leadership must respect the truth in the numbers and make decisions based upon that truth, rather than their expectations or their desires. It is not always easy to step outside ourselves and open our minds to differing opinions, but once you identify confirmation bias and take steps to avoid it, you will be able to make stronger, more objective decisions for your company.