The impact of product card visual design on sales: Wildberries case study

Abstract

Amid the rapid expansion of the e-commerce market, the visual characteristics of product cards are becoming increasingly important for capturing customers' attention, building trust, and motivating purchases. This study addresses the limited evidence on how individual visual attributes of product cards affect sales across categories. The aim of this study is to identify design features of product cards on the Wildberries platform that affects sales, accounting for category-specific effects. We compile and analyze a dataset of 76770 product listings spanning six categories: groceries, household chemicals, clothes, shoes, household appliances, and cosmetics. Ten visual features are extracted from the main images: brightness, contrast, color palette, presence of a person, presence and amount of textual information, entropy, object area and vertical and horizontal compositional symmetry. Sales performance is modeled using tobit regression to account for censoring. The results show that salient visual drivers are category dependent. For groceries, both the presence of text on the image and color palette. In clothes, brightness and color palette matter; in shoes, image composition and its color palette are significant factors are significant. In household chemicals and appliances, text is the key factor, with a higher optimal amount for appliances than for chemicals. The cosmetics model yields the most significant coefficients, including color palette, contrast, entropy, object area, and composition.