When you have data quality problems, you might make better decisions by guessing.

A cartoon on page 50 of the July-August 2013 Harvard Business Review (HBR)depicts one business guy telling another “We cut our marketing budget in half by just guessing.” Have data quality problems let you to a similar decision-making approach? Let’s have a look at some decision-making challenges that can only be made worse with data quality problems.

“Before You Make That Big Decision…”

That is the title of the insightful June 2011 HBR article by Daniel Kahneman, Dan Lovallo, and Olivier Sibony. If you have an HBR online subscription, you can read it here. The thesis of the article is that our decision-making process is laden with bias, some of which we can detect and eliminate, and some of which we cannot. Which is which? Please read on.

 

 

The Most Common Decision is…

The authors posit that the most common executive decision is to accept, reject, or pass up the line a recommendation for action prepared by the executive’s team. OK, how would the typical executive react when receiving a recommendation? According to the authors, the response sequence would be:

1 – Get the relevant facts from the people with the details

2 – Determine whether the facts have been “massaged” by the recommenders

3 – Evaluate based on the facts obtained, team bias analysis, and prior experience

While our authors focus on decision-making bias, I want to inject the thought that in an environment of pervasive data quality problems, making a data-based recommendation, and evaluating that recommendation would amount to guessing. When the data that formed the basis of a recommendation or evaluation is seriously flawed, bias might not be the greatest threat to making a good decision. See my prior blog post to learn more about data quality problems.

What are the three sources of decision-making bias?

It seems that, not including data quality problems, there is a triple-threat to good decision-making. As decision-makers, we bring personal bias to the decision, our recommendation team has biases, and our organization has biases. I will discuss the personal bias challenges in this post, and save the other two for future discussion.

We are blind to our own personal biases

This doesn’t seem surprising, but let’s see why we can’t self-assess. It seems we have two distinct but concurrently operating thinking processes in operation as we experience the world. These are the intuitive and reflective thinking processes. Intuitive thinking happens as naturally as breathing, and produces a consistent and comfortable view of the world as we interact.

This intuitive thinking is so smooth that when it leads us astray, we are, in the moment, unaware. Worse still, “post-game” analysis of flawed intuitive decision-making doesn’t lead to improved future results. When we engage in reflective thinking, we are applying rules, calculations, and similar structured approaches to decision-making and problem-solving.

As the authors point out, everyone remembers when they worked hard to solve a problem and got the wrong solution. “Post-game” analysis of flawed reflective thinking will, ceteris paribus, lead to future improvements.

The Bottom Line

So, it is the insidious yet supremely comfortable intuitive thinking that places personal bias elimination out of our reach. If we force ourselves to make every decision through reflective thinking, we might avoid personal bias, but we would never get through the decision-making day.

Worse still, if we face data quality problems in applying reflective thinking to the truly important decisions, we will have traded personal bias problems for fact-based problems. In my opinion, the latter class of problems are as pervasive, but are amenable to elimination. Do you believe you have data quality problems? Contact us to learn more about our Poor Data Quality survey, and how we can work together to detect and correct current data quality problems, and avoid them in the future.