Reproduce “Billion-user Customer Lifetime Value Prediction: An Industrial-scale Solution from Kuaishou” 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

Kunpeng Li, Guangcui Shao, Naijun Yang, Xiao Fang, Yang Song

Researchers at Kuaishou developed the Order Dependency Monotonic Network (ODMN) framework, an industrial-scale solution for Customer Lifetime Value (LTV) prediction across billions of users. ODMN effectively models severely imbalanced LTV distributions and enforces ordered dependencies between multi-time-span LTVs, leading to a 14.7% uplift in 30-day online ROI compared to a baseline.

Key insights

  • Decomposing a highly imbalanced LTV distribution into a hierarchical structure of sub-distributions and fine-grained buckets effectively transforms a complex problem into more manageable sub-problems.
  • Explicitly enforcing monotonic relationships between LTV predictions for different time horizons through dedicated network units and calibration losses ensures logical consistency and improves longer-term prediction accuracy.
  • A novel Mutual Gini metric provides a more comprehensive evaluation of how well a model fits the entire LTV distribution, particularly the crucial high-value segments, beyond traditional point-wise errors.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.