Unveiling the power of social networks in demand prediction for substitutable fashion products

Many fashion-apparel firms use social media for marketing, but very few studies have explored the potential of leveraging social-network data as an analytics tool in the fashion domain. We examined the contribution of social-media data to forecasting demand for substitutable products. We analyzed a year’s daily-level sales and social-media data (likes, comments, and shares on the retailer's Facebook page) from a fast-fashion retailer regarding two substitutable products (sneakers in two colours). The data collection was complemented by regression-based modelling to evaluate the impact of social-media data on prediction accuracy. We found that data on customers’ social-network activity significantly enhanced prediction accuracy, beyond data regarding product sales. The study thus contributes to the theory by demonstrating how social media data can enhance demand prediction models in scenarios where consumer substitution behaviour plays a critical role. The findings suggest that social media data can serve as a strategic tool for inventory planning.