Understanding data is one thing. Being able to effectively analyze it is another. Effective data analysis consists of 1) proper techniques to disseminate data and 2) being aware of cognitive biases that can skew our analysis.
Cognitive biases are lapses in logical and rational thinking, remembering, or other cognitive processes. These often occur as a result of maintaining one’s beliefs and preferences even when practical, logical, and/or contrary information is present.
These powerful forces can heavily skew our perspective and interpretation of data. We can't remove biases, so we need to learn how to acknowledge them and work with them.
Here we walk thru some of the most common (and powerful) biases we have to be aware of when performing any interpretation of information.
Biases can be grouped into 4 main categories: biases born from too much information, not enough meaning, the need to act quickly, and the limits of memory. The image below is a beautiful grouping of the different types of biases.
Are Biases Bad?
Everybody has cognitive biases. Despite their reputation, not all biases are negative. Many of them serve an adaptive purpose are actually essential to speed up decision making.
While it is impossible to remove all biases (it is part of being human), it is critical that we put effort towards identifying and acknowledging potentially misleading prejudices so that we can conduct objective analysis. Below are the most common biases in the data & analytics world:
Remember that it is healthy and useful to stress test your viewpoint with different perspectives.
A Strategy For Addressing Biases
Over the years we have developed a number of reliable strategies for acknowledging biases and preventing group think in a business environment. 2 strategies that should work for anyone:
1) Ask "What do I WANT the data to say?"
This forces you to address your bias up front. An example might be you are looking to give a customer a price increase. With this bias, you will naturally look for the information that supports giving a price increase and will discount or ignore contradicting data that suggests a price increase may not be needed.
Spending a few minutes up front understanding your agenda can help ensure you are offering up as objective of an analysis as possible. The benefit is you avoid misleading your audience and/or yourself.
2) Assign a contrarian to the analysis
Never underestimate the value of having someone in the group who's role is to be the contrarian.
The job of this person is to poke holes in your methodology, approach, and analysis, even if they strongly agree with your angle. Their job is to challenge your analysis in the spirit of making it better/stronger.
This is a difficult role that requires a pretty unique mindset, so if you find someone capable of doing this be sure to leverage them on a regular basis.
While it seems counter intuitive, these 2 basic strategies for identifying and acknowledging biases will:
Every piece of analysis we do at Strategy Titan follows this simple framework. The benefits far outweigh the costs. I would strongly encourage you to do the same. Is this approach a little bit more work? Yes. Is it worth the effort? I believe so.
The next piece in this series, Data Literacy Playbook: A Path To Business Superpowers, outlines a basic data literacy playbook you can start using today.
Be sure to check out the previous installments in this intro to data literacy series:
Part 1: Why you MUST Be Data Literate In Business Today
Part 2: A Company’s Most Valuable Asset? Data
Like this content? Be sure to join our newsletter list and follow me on LinkedIn.