Case Study: Condé Nast achieves improved dataset quality and model performance with Labelbox

Condé Nast - Customer Case Study

Condé Nast, through its Formation research team led by Paul Fryzel, needed a scalable way to manage and improve datasets across 20+ media brands and distributed teams. Building a custom tool would have diverted researchers from modeling work, so Condé Nast adopted Labelbox to provide a centralized, expert-backed platform for labeling, dataset management, and collaboration.

Labelbox delivered a real-time labeling workspace and dashboard used on projects like Vogue’s runway detection, where the team discovered a 10× imbalance of shoes versus dresses and quickly rebalanced labels. That tighter feedback loop, checkpoints, and ongoing support from the Labelbox team improved dataset quality, sped up workflows, and led to measurable gains in model performance and iteration speed.

Condé Nast

Paul Fryzel

Principal Engineer