How the Wrong Data Visualization Can Lead to the Wrong Decision

Φεβρουαρίου 04, 2015

See how one Data Viz underdog can help you avoid obstacles to data-driven decision making.

Throughout all the years that I have advocated for the importance of good data visualization within business intelligence, I've done so with the understanding that ultimately it's not a matter of life or death. However, it is the difference between making bad and great decisions. Which in turn can mean the difference between the life and death of your company. Or, as you'll see later on, your job.

Normally I'm not a pessimistic person (just ask all my customers or my boss when I estimate tasks), but I feel that it is important to highlight the consequence of using poor data visualization when making important business decisions. Good data visualization uncovers and exposes your data while poor or no data visualization hides and disguises your data.

I've outlined my 10 best practices for great data visualizations here on the blog.

When data is hidden from sight or you are not aware you have it, making good business decisions becomes increasingly more uncertain and less data driven. Inevitability it gives a false sense of security and leaves your decisions open to more risk that you may have intended.

There are two parts to address here. First, not being aware that the data exists. Second, not being able to find the relevant data when the time comes to use it.

In this post, we will look at how data visualization can enhance your awareness of the data available and its importance to business decisions and a successful BI strategy.

You can download the guide to a better BI strategy here.

Over the last few years, I have been using a technique that has helped me gain a deeper insight into business data and its importance to business decisions. It is a technique that I call "scattering the data." I don't mean throwing it up in the air and hoping that is lands with all the answers facing up. Although I have never tried that and you never know, it might work ;)

Rather, I'm referring to one of data visualization's underestimated tools: the Scatter Chart. The Scatter Chart allows you to confirm the relationships in your data that you already knew existed. And, even more importantly, it helps find relationships you didn't know existed. Being aware of hidden relationships can mean the difference between a bad and a great decision.

What is it a Scatter Chart?

The purpose of a Scatter Chart is to show the type of relationship -- often called "correlation" -- which exists between two sets of data points. A Scatter Chart's correlation can have three states: positive, negative, or no correlation. Each of these relationship states tells you something important about your data. 

Just to be sure you understand, let's have a look at the three different relationship results.

Positive Correlation: As the number of Transactions grow so too does the Revenue.


Negative Correlation: As the Quality Index increases the number of Rejects decreases.


No Correlation: The Average Unit Price does not affect the number of Transactions.


Let me tell you about a situation were poor data visualization led to bad decisions and the impact that these decisions had.

A retail company's branch manager was told by head office at the beginning of the year that his branch had to increase the yearly profit margin by 2 percent. In an attempt to find areas within the branch where costs could be cut and sales increased, the branch manager turned to their BI software in order to make data-driven decisions. I would normally say that this would be a great start to the decision-making process, but the decision this branch manager took would result in a negative profit margin growth for the year.

Breaking down the decision-making process, we understand why it went so wrong and gained a greater insight into how poor data visualization can blind you when making important decisions.

First, the branch manager looked into all the costs that were in their power to reduce. After looking at costs such as employee salaries, rent, assortment, and scheduled discounts/campaigns, it was decided that schedule discounts/campaigns was a good place to start.

After surveying the Discount Management analysis in TARGIT, the branch manager saw that last year's discount amount of $615,063.00 was just over 2 percent of last year's total revenue. Problem solved! 

After looking at the distribution of revenue and discounts per week over the last year, it seemed that there was no direct link between revenue and discounts.


So the branch manager stopped all discounts, promotions, and campaigns. But a year later revenue was down over 5 percent and head office was looking for a new branch manager.

What the branch manager failed to uncover was that there was a positive correlation between discounts and revenue. If the branch manager had only scattered the data on a Scatter Chart, he would have seen this and maybe taken another decision.


Now I'm not saying that knowing this alone would have led to a better decision, but not knowing it definitely contributed to making a bad decision. 

When the new branch manager starts, they'll have a new Discount Management analysis that includes a Scatter Chart so they'll be able to see the same data from different viewpoints, helping them make better, data-driven decisions.


David Hollinsworth

Senior BI Consultant
During my journey as Business Intelligence consultant, I have experienced my fair share of BI tools. Nevertheless, it was not until my first meeting with TARGIT that I truly found my calling. My first impression of TARGIT Business Intelligence was, "it can't be this easy to create good looking and powerful analytics." You could say that it was love at first sight. As..
Continue Reading...

We use cookies to improve your site experience, but they also provide us with information on your use of our website.
To find out more about the cookies we use and how to delete them, see our Privacy Policy. By continuing to browse the site, you are consenting to our use of cookies.