Zalando's new AI pricing system boosts profit 6% in minutes

A team of researchers from Zalando, one of Europe's largest online fashion retailers, have published a paper detailing a new high-frequency algorithmic pricing system designed specifically for sales campaigns. The system tackles the unique challenges of fashion e-commerce: volatile demand, rapid decision-making, and the need to balance short-term revenue with long-term profitability. Their solution combines daily-resolution demand forecasting using gradient-boosted trees with a multi-objective optimization framework that maximizes both long-term profit and net merchandise value. This architecture replaces the previous weekly-granularity systems, cutting pricing decision time from hours to minutes.

The researchers validated their approach through 23 A/B tests conducted across 12 markets during the 2023-2024 sales campaigns. The results showed the new system achieved approximately 6% higher profit while maintaining equivalent performance on sales and revenue compared to the previous hybrid approach. Based on these strong results, the algorithm was successfully deployed to production and now handles the majority of algorithmic pricing decisions for Zalando's sales campaigns. The paper highlights how AI-driven pricing can scale to over 5 million articles, enabling near real-time optimization that was previously impractical. This work demonstrates the tangible business impact of combining machine learning forecasting with optimization in large-scale e-commerce.