Color Analytics for Data-Driven Brand Communications (in cooperation with Daria Dzyabura)

What are the colors of Sophistication? Is Tiffany’s “Robin Blue” really glamorous? What if McDonald’s yellow arches turn green? Color plays a significant role in branding activities, ideally serving the brand’s strategic positioning goals. However, brand color decisions are largely qualitative, based on creative and aesthetic considerations, intuition, and trial and error.

Current research focuses on specific brand elements (e.g., logo), analyzing a limited number of standalone colors. We propose an approach that is rooted in the consumer perspective, incorporates the full complexity of the color space, including color combinations, and maps a large number of brand metrics into the color space. We use a large dataset of individual-level digital collages and ratings on brand personality and equity characteristics for 303 major U.S brands to map the color space into self-organized data-driven color palettes.

We identify and validate relationships between characteristics and color palettes and examine them relative to designer palettes and to the brand-generated communication. For example, Family Oriented is positively related to a palette with dominant shades of green and yellow and negatively related to palettes which are predominantly purple, blue, and pink.