Publication Date

Spring 5-3-2026

Document Type

Student Project

First Advisor

Jeon, Yongwoog (Andy)

Degree Name

B.S. (Bachelor of Science)

Department

Department of Marketing

Abstract

Brands in the beauty industry rely heavily on visual content to differentiate themselves and communicate their identity on social media. However, it remains unclear how visual complexity affects consumer engagement and perceptions, and limited research has combined visual entropy, sentiment, and engagement.

This project examines how three beauty brands: Glossier, Rhode, and Clinique, use color and visual complexity in their Instagram content and how these different strategies relate to emotional tone, brand identity consistency, and consumer engagement. A dataset of approximately 150 posts was collected, and deep learning-based visual analysis was used to extract the dominant colors and calculate visual entropy. Comment sentiment was evaluated, and a regression analysis was conducted to test the relationship between entropy engagement and sentiment.

Findings indicated a statistically significant negative relationship between visual entropy and eposts, suggesting that more visual complexity reduces users’ likelihood of sharing content. No significant relationship was found between entropy and likes, comments, or sentiment, although similar negative trends were observed. These findings suggest that simpler visual content is more effective at driving engagement, supporting the cognitive load theory.

This study aims to provide insight into how visual design influences digital engagement and offers practical insights for brands seeking to optimize their social media strategy.

Download

Share

COinS