Reproduce “Markdowns in E-Commerce Fresh Retail: A Counterfactual Prediction and Multi-Period Optimization Approach” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Junhao Hua, Ling Yan, Huan Xu, Cheng Yang
Researchers at Alibaba Group developed a dynamic markdown pricing framework for e-commerce fresh retail that combines a semi-parametric model for counterfactual demand prediction with a multi-period optimization algorithm. This approach achieved a 92.12% total inventory clearance rate and a 17.14% higher GMV improvement compared to manual strategies in online A/B tests.
Key insights
- Combining a non-parametric component for base sales forecasting with an interpretable, parametric price elasticity model allows for reliable counterfactual demand prediction and causal inference of price-demand relationships.
- Modeling demand uncertainty through Poisson distributions and framing the dynamic pricing problem as an MDP leads to a robust strategy for perishable goods that accounts for sales randomness.
- Decomposing the multi-shop dynamic pricing problem into a two-stage algorithm effectively addresses the 'curse of dimensionality,' making it computationally feasible for large-scale e-commerce operations.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.