Eighteenth century Scottish engineer and economist William Playfair was one of the founding fathers of presenting data visually. He is credited with inventing the ubiquitous line and bar charts. That’s more important than it sounds, given that research since shows humans process visual information faster than verbal information. And we do it with a part of the brain that requires less energy.

But Playfair also, a few years later, gave us the pie chart, a highly popular form of graphical display that, as a form of meaningful communication of financial data, fails miserably.

Playfair, operating a couple of years before the term “biology” had even been coined, didn’t have the advantage of years of scientific research on the workings of the brain and how it processes visual information. “He understood the intuitive ability of a sub-divided circle to represent part-to-whole relationships, but not the perceptual problems that we encounter when trying to compare its parts,” wrote Stephen Few, a consultant with 25 years in business intelligence and information design, in his Visual Business Intelligence Newsletter.

17Dec_VisualDataPlayfair didn’t realize (or care, perhaps) that a pie chart often makes it hard to figure out the exact magnitude of a data point (a slice) and uses a lot of text to display very little data. It also forces readers to rapidly move their eyes back and forth between the legend and the graphic to interpret the data. A simple table can be a lot more elegant, experts say.

Armed with scientific research and better charting tools, corporate executives, 200 years after Playfair, are still using pie charts. And they are also committing a host of new sins in the name of holding employees, colleagues, and bosses rapt when presenting financial data.

By employing poorly conceptualized and designed graphics in presentations, dashboards and reports, executives often fail to convey information clearly, accurately, and efficiently. They are confusing their audiences, and, worse, may be leading decision makers astray. The simplest (but not necessarily obvious) things in graphical presentations can distort data analysis or distract the audience from the important numbers, leading to bad or biased decisions.

For example, shapes and colors that “pop out” in a chart or have the greatest variation in size will be more heavily weighted in decision making, according to a 2007 paper, “Visual Representation: Implications for Decision Making,” by Nicholas Lurie, of the University of Connecticut, and Charlotte Mason, of the University of North Carolina. Further, researchers have found that the arrangement of data in a visual display influences the interpretation: When a company’s sales results chart is ordered from highest to lowest, it may lead to higher overall sales estimates than when it is arranged from lowest to highest, say Lurie and Mason, because of the “tendency to place greater weight on initial information.”

Just the choice to display information in a graphical format as opposed to text can skew the meaning. On pages of a report that include text and graphics, the audience is likely to lend greater weight to the graphical information, say Lurie and Mason. The perception of risk is higher, for example, when it is presented graphically and the cost of reducing a risk is presented in text format, compared with when both are presented in text format, studies have shown.

17Dec_VisualDataExampleSome of the sophisticated information-visualization tools available today worsen the problem — the 3-D pie charts and “exploding” 3-D pie charts available in Microsoft Excel, for example, only add to the data distortion and ambiguity in the traditional pie chart, according to Gary Klass, author of “Just Plain Data Analysis” and a political science professor at the University of Illinois. Excel can produce charts and graphics that are beautiful but add little to the understanding of a data set.

One place mistakes in data visualization show up is in dashboards. Many corporations have adopted dashboards so that executives can monitor key metrics through a single platform and avoid wading through multiple spreadsheets or other report formats. But many corporate dashboards are rife with design flaws. Stephen Few, in a whitepaper called “Common Pitfalls in Dashboard Design,” unearthed some common mistakes:

• “Displaying excessive detail or precision.” To display a number or amount too precisely ($3,848.305.93 instead of $3,848,306 or $3.8M) “forces the viewer to process levels of data that are irrelevant to the task at hand,” Few writes. “Every unnecessary piece of information results in wasted time.”

• “Expressing measures indirectly.” If the dashboard viewer only needs to know by how much actual revenue differs from budgeted revenue, instead of displaying the actual revenue and budgeted revenue amounts and leaving it to the viewer to calculate the difference, the presenter can display the variance amount directly, says Few. Also, in many cases, rather than displaying the actual variance in dollars it’s more direct to simply express the variance as a percentage.

• “Ineffectively highlighting what’s important.” The user should be able to look at a dashboard and have his eyes immediately drawn to the most important information. When everything is equally prominent visually, as in many dashboards produced by the latest software applications, nothing stands out.

There’s also the basic problem of cluttering the screen with bells and whistles to make the dashboard look like the inside of a fighter jet. “Too often we smear a thick layer of gaudy makeup over data in an effort to impress or entertain,” says Few in his whitepaper.

So what’s the key to good information-graphic design?

Some chart types should just be forbidden in finance departments or entire organizations. Besides pie charts, stacked bar charts are to be avoided, says Klass, “because they make estimating the values of the variables on the top of the bars difficult.”

Graphics that require estimations of area also befuddle users. Lurie and Mason analyzed a famous General Electric/McKinsey matrix used for business portfolio analysis. In the graphic, circles represent business units and the areas of the circles are proportional to market size. But according to what researchers call the “size effect,” decision makers will underestimate the magnitude of the difference between larger and smaller circles (business units), the professors note.

A visual display is “only successful to the degree it encodes information in a manner that our eyes can discern and our brains can understand,” writes Few in “The Encyclopedia of Human-Computer Interaction.” A good graphic represents quantities accurately; makes it easy to compare them and see how they relate to one another; and makes it easy to see a ranked order of values. Perhaps most important, says Few, it “makes obvious how people should use the information — what they should use it to accomplish — and encourages them to do this.”

Sounds simple. But the average PowerPoint presentation or Excel-generated report clearly shows that it’s not.

, , , , , , , , , , ,

Leave a Reply

Your email address will not be published. Required fields are marked *