SCoTER: Structured Chain-of-Thought Transfer for Enhanced Recommendation

Harnessing the reasoning power of Large Language Models (LLMs) for recommender systems is hindered by two fundamental challenges. First, current approaches lack a mechanism for automated, data-driven discovery of effective reasoning patterns, relying instead on brittle manual templates or unstable zero-shot prompting. Second, they employ structure-collapsing integration: direct prompting incurs prohibitive online inference costs, while feature extraction collapses reasoning chains into single vectors, discarding stepwise logic. To address these challenges, we propose SCoTER (Structured Chain-of-Thought Transfer for Enhanced Recommendation), a unified framework that treats pattern discovery and structure-aware transfer as a jointly optimized problem. Specifically, SCoTER operationalizes this through two synergistic components: a GVM pipeline for automated pattern discovery and a structure-preserving integration architecture that transfers stepwise logic to efficient models. Formally, we provide information-theoretic justification proving that structure-preserving transfer achieves tighter performance bounds than structure-agnostic alternatives. Empirically, experiments on four benchmarks demonstrate improvements of 3.75\%-11.59\% over a strong TIGER backbone. Moreover, in production deployment on the Tencent Advertising Platform, SCoTER achieved a 2.14\% lift in Gross Merchandise Value (GMV) while eliminating online LLM inference costs. Overall, SCoTER establishes a principled and production-validated blueprint for transferring structured LLM reasoning to large-scale recommender systems.