Give Color a Job

The model is done, the content is right, and now comes the fun part: the color picker. A splash of orange here, a friendly green there, maybe a gradient on the header, anything to keep it from looking plain. The result is colorful. It is also, very often, harder to read than the plain version was.

Color isn't the problem. Unemployed color is. Used well, color is one of the hardest-working elements in a visual, but only if you give it a job to do.

The Principle: Give Color a Job

Give color a job means every color in your visual should carry meaning: same color, same group; different color, different role; no meaning, no color. In a logic model, that can be as simple as one soft fill per column, every activity shares one shade, every output another. The reader now knows what belongs together at a glance, and the heavy black borders and divider lines that used to mark the groups can quietly disappear. The fills are doing the enclosing.

Why It Works

Two Gestalt principles are at work. Similarity: things that look alike are perceived as a group, automatically and instantly. Enclosure: a shared field of color creates a boundary without anyone drawing one. Both happen before conscious thought, which means color can communicate structure at zero cost to the reader, no lines to trace, no legend to memorize. But the same automatic processing cuts the other way: when colors vary without meaning, the reader's visual system hunts for a pattern that isn't there. Decorative color doesn't just sit idle; it sends false signals.

A study published in the American Journal of Evaluation redesigned a real program's logic model using exactly this approach; soft column fills in place of black-bordered boxes, alongside better placement and less clutter. Compared to the original, people who viewed the redesign understood the program more accurately, reviewed it faster, and rated it as better looking and more credible.

Then comes the twist. The researchers also tested a black-and-white version of the redesign, with grayscale fills standing in for the colors. It performed nearly as well as the color version, and was rated the most credible depiction in the entire study. The power was never in the hues. It was in using fills to group and enclose. That's worth pausing on: if your color scheme still works in grayscale, you know the colors are doing structural work, not decorative work.

Caption: Same boxes, same layout. On the left, color varies without meaning and the reader’s visual system searches for a pattern that doesn’t exist. On the right, color marks the groups, similarity and enclosure do the work that borders and lines used to do.

Caption: In the study, grayscale fills performed nearly as well as color,  and were rated most credible. The grouping and enclosure mattered, not the hue. (As with the other redesign principles, fills were applied alongside improved placement and reduced clutter, so overall gains reflect the combined redesign.)

Try the Grayscale Test

This finding doubles as a free accessibility check. If your model's grouping survives a black-and-white printout, it will also survive most forms of color-vision difference, a dim projector, and a bad photocopier. So put your next model through a simple audit. Ask each color: what's your job? If it marks a group, keep it. If two colors mean the same thing, merge them. If a color is only there to be cheerful, let it go, then print the result in grayscale and see if the structure still shows. If it does, your colors were never decoration. They were architecture.

Reference

Jones, N. D., Azzam, T., Wanzer, D. L., Skousen, D., Knight, C., & Sabarre, N. (2020). Enhancing the effectiveness of logic models. American Journal of Evaluation, 41(3), 452–470.