Ensuring the Evidence Is Evident

© Can Stock Photo Inc.
© Can Stock Photo Inc.

As I’ve covered in past posts, it takes years of painstaking academic inquiry to gather the right data and perform the right analyses to support your evidence-based investment strategy. Then, there’s step two: Effectively communicating what you’re doing to everyday investors. As fond as I am of the written word, data visualization – those tables, charts and graphs that accompany your prose – is also essential to ensuring that investors internalize the “what” as well as the all-important “why” that will help them stay the course in turbulent times.

Data visualization is for sharing data in ways that our dominant sense — sight — can readily interpret. Or, as Edward Tufte, an early leader in data visualization, says, data visualization occurs when “clear and precise seeing becomes as one with clear and precise thinking.”

A few years ago, to learn more about data visualization, I attended a Visual Business Intelligence workshop offered by Stephen Few of Perceptual Edge. Few, an author and expert in applied data visualization, cuts to the essence of effective delivery with a three-step process:

  1. Determine your message.
  2. Select the best medium to display your message.
  3. Design all components of the display to showcase the data.

That’s it. These steps seem blindingly simple, don’t they? But much like evidence-based investing, these basic tenets can be a lot harder to implement than you might think. They also have a major impact on the success or failure of your efforts. If you’re going to work hard at communicating effectively, you might as well shoot for the best possible outcomes, don’t you think?

Step One: Determine Your Message

Again, it may seem obvious that it’s important to determine your message. And yet, I can’t tell you how often I’ve reviewed materials in which the main point is either entirely absent, or so deeply buried that it might as well be. Without a clear theme, you may be sharing data, but it’s not yet knowledge. It doesn’t foster Tufte’s “clear and precise thinking.”

A vital part of determining your message is identifying your audience. Is the message intended for well-established clients? Brand new prospects? The media? A specialized niche? What are the details of the subject that are likely to matter the most to your specific audience? The more narrowly you can define to whom you’re communicating, the more easily you can craft your message to deliver that ah ha! moment when you show them the numbers.

Step Two: Select the Best Medium To Display Your Message

By “medium,” we’re talking about which type of chart or table is best for which types of data. Academics and practitioners like Tufte and Few have been looking at the evidence on that, and here is what they have found:

  • Tables are most useful when the user will want to look up or compare precise values — such as dollar amounts in each fund held. It’s easiest to read precise values in columns/rows.
  • Charts function well for analyzing relationships among the shapes being viewed – such as the popular and ubiquitous “Growth of $1” chart. Imagine trying to view all those numbers in a table and still gain the same insights about stocks, bonds, bills and inflation data, ongoing since 1926!

Specific charts are useful for specific subsets within that general function. For example:

  • Bar charts tend to be best for part-to-whole comparisons or rankings
  • Line charts are most effective for time series
  • There also are times you can combine various types into a single illustration.

Like building a balanced portfolio from its asset class parts, delivering effective data visualization requires recognizing the right role for each charting tool, as well as the interrelationship among them.

Here’s a surprise to many who are new to effective data visualization: pie charts are almost always a poor choice. This is mostly because it’s harder to compare “slices” versus bars, and it becomes even harder if you add 3-D effects. In Figure 1, the pie and bar charts each display the same data values in the same order. Arguably, the pie chart may be “prettier.” But your goal is to communicate, not decorate.

Figure 1: Compare the effectiveness of a bar chart vs. a pie chart. Each displays the same data, but note how much faster you can compare values using the clean bar chart, versus the “pretty” pie.

Step Three: Show the Data

Once you’ve selected the media, the final step is to polish your data visualization until it glows. There are far more techniques for effective design than could fit into a blog post. Tufte’s landmark book, “The Visual Display of Quantitative Information,” condenses the theory behind data visualization into these basic rules:

  • Above all else show the data.
  • Maximize the data-ink ratio.
  • Erase non-data-ink.
  • Erase redundant data-ink.
  • Revise and edit.

“Ink” means anything appearing in the chart or graph. Ink overload is where most data designs go awry. Overall, the less unnecessary ink, the less time it takes to find the message, which leads to more effective communication and happier people:

  • The data-ink that is “showing the data,” should be the most obvious – by far – and even it should be applied as efficiently as possible (the less required to make the point, the better).
  • Supporting information should be visually secondary to the data and only included as essential.
  • Remove ink that serves no purpose besides decoration. It only slows down comprehension.

For example:

  • Supporting information such as labels, tick marks, keys and background coloring should be there only if needed, with minimal disruption to the main attraction, which are the bars, lines, scatter plot or other devices that are displaying the data.
  • If information is duplicated – such as a value being identified by both label and coloring – select the best method and eliminate the duplication.
  • Avoid including illustrations unless they clearly support the data.
  • In color selection, don’t forget that approximately 7-10 percent of men are red-green color-blind.

With these essential concepts in mind, it becomes easier to build new data presentations as well as improve upon existing ones. For example, For example, the left-hand illustration in Figure 2 shows the results if I simply accept all the default settings that PowerPoint selects for me when I create a basic bar chart. At right, I’ve applied a few simple data visualization techniques to add polish and clarity to my message.

Figure 2: Before and after. At left, we used PowerPoint’s default settings to create a basic bar chart. At right, a few basic data visualization techniques can add significant power to your PowerPoint message.
Figure 2: Before and after. PowerPoint’s default settings (left) vs. applying a few  data visualization techniques to eliminate inky overload (right).

Applying Strategy to Reality: Why Bother?

It’s fun to know a little bit about how professionals create effective data visualizations. (Well, I think it’s fun, anyway.) Your audiences don’t need to know the gory details that go into proper design, but they’ll benefit greatly from being able to understand your messages quickly and easily through thoughtful design. Not to mention, you’ll stand out from the crowd in terms of the quality of your materials. While effective design need not be more work once you get the hang of it, it’s far less frequently applied.

If you’re thirsting for more on the subject, I recommend Tufte’s book for the theoretical underpinnings and Few’s “Show Me the Numbers” for applied understanding. And to help with the words in between, don’t forget about yours truly.