Pet Peeve #42: Bad Graphs

Disclaimer: I have a lot of respect for eMarketer. They do great work, and although I’m picking on one of their graphs, I mean them no disrespect.

OK, with that out of the way, take a look at graph below. See anything wrong?

Here’s what ticks me off about it: The 58.4% bar stretches all the way across the graph.

Think about this for a moment. This chart can be translated into words: “Talent was cited by 58.4% of marketers as a web analytics challenge. Actionability was mentioned by 47.3%. Finding insights was the third most frequently mentioned challenge…..” And so on.

What a snooze that would be. That’s why we use charts and graphs — they’re a more efficient (and perhaps effective) way to communicate something.

But when you stretch a 58.4% bar all the way across your graph you diminish the efficiency of the graph because you make viewers work harder than they have to to interpret the graph. A bar that occupied about half the space allotted to the graph would be perceived as having a value of about 50%.

But in the graph above, with the 58.4% bar taking up the whole space, it’s not only not intuitive that the value is 58.4%, but it makes it harder to interpret the other bars.

I see this kind of thing in a lot of presentations. Sometimes it’s the result of not consciously thinking about, but sometimes it is intentional. Like stretching a 20% bar across the space of a graph is going to make people think that 20% is a really big number.

Next time you construct a graph and chart for a presentation, think about how you use the space. And round the percentages down to no decimal places (exception: when the numbers are less than 10%).  When people have to work harder than they should to understand what you’re talking about, you’ve lost a little of your potential effectiveness as a presenter.


2 thoughts on “Pet Peeve #42: Bad Graphs

  1. Wow Ron! Data visualization! Nice! Bad data visualization should always be called out. Why? Because most of us in business were never properly trained. We all grew up with bad data visualization examples.

    Naomi B. Robbins, Ph.D. and Joyce Robbins, Ph.D. in their February 2010 article “Quantitative Literacy Across the Curriculum – Improving Graphs in College Textbooks”,
    Naomi B. Robbins, Ph.D. and Joyce Robbins, Ph.D. address the issues of bad data visualization in social science textbooks. I suggest you and all your readers read these seven pages as they apply to business as well. The PDF can be found here:

    Here are some of my data viz PET Peeves and those of my data viz friends:

    Final Note:

    Ron closed with “When people have to work harder than they should to understand what you’re talking about, you’ve lost a little of your potential effectiveness as a presenter.” So true. There are so many great resources to learn how to use data effectively.

    Many of us complain about bad data visualizations. Are there any good ones? Here is a link to one a like and a great suggestion by a reader.


  2. Thanks Ron!

    This is one of my biggest pet peeves also. I get sick and tired of sitting through long-winded meetings where the presenters either stay off track, never get to the point, or otherwise find ways to waste everybody’s time. I’m a very matter-of-fact kind of person. If you have something to tell me, just tell me.

    You’re definitely correct with respect to graphical representations of data. In an attempt to make things “look better” than they are, or by not being critical with their work, people often make perfectly good data either incomprehensible or irrelevant. Thanks for calling this to attention. Take care and best wishes.

Comments are closed.