程序化广告转型下的电商营销效能提升路径研究——以京东“知识图谱友好性”广告生态为例

Research on the Improvement Path of E-Commerce Marketing Efficiency under the Transformation of Programmatic Advertising—A Case Study of JD.com’s “Knowledge-Graph-Friendly” Advertising Ecosystem

摘要: 程序化广告正从“流量采购”向“智能匹配”深度转型，知识图谱技术成为提升电商营销效能的关键驱动力。文章以京东广告生态为案例，基于资源基础观、动态能力理论与平台生态系统理论的整合框架，系统分析京东如何通过GRAM生成式检索对齐模型、多模态内容理解体系及AI智能体“京小通”，构建“知识图谱友好性”广告生态并提升电商营销效能。研究发现，知识图谱技术通过“深层意图理解–端到端匹配–全链路自动化”的三层作用机制，实现了电商营销效能的系统性提升：GRAM模型将广告匹配从“关键词匹配”升级为“意图推理”，P99时延压缩至30~50 ms；“京小通”智能体将广告投放门槛从“操作岗”降为“战略岗”，测试阶段提效30%；“稳赚计划”以AI算法保障商家ROI。研究提炼出对大型平台、中小平台及商家的分层差异化策略建议，为程序化广告转型期的电商营销实践提供理论参照与路径参考。

Abstract: Programmatic advertising is undergoing a profound transformation from “traffic procurement” to “intelligent matching”, with knowledge graph technology becoming a key driver for improving e-commerce marketing efficiency. Taking JD.com’s advertising ecosystem as a case study and based on an integrated framework of resource-based view, dynamic capabilities theory, and platform ecosystem theory, this paper systematically analyzes how JD.com constructs a “knowledge-graph-friendly” advertising ecosystem and enhances e-commerce marketing efficiency through its GRAM (Generative Retrieval Alignment Model), multimodal content understanding system, and AI agent “Jing Xiaotong”. The study finds that knowledge graph technology achieves a systematic improvement in e-commerce marketing efficiency through a three-layer mechanism of “deep intention understanding - end-to-end matching - full-chain automation”. Specifically, the GRAM model upgrades advertising matching from “keyword matching” to “intention reasoning”, compressing P99 latency to 30~50 ms; the “Jing Xiaotong” AI agent lowers the barrier to advertising placement from “operational role” to “strategic role”, achieving a 30% efficiency improvement in the testing phase; and the “Stable ROI Plan” uses AI algorithms to guarantee merchant ROI. This paper extracts stratified and differentiated strategy recommendations for large platforms, small and medium-sized platforms, and merchants, providing theoretical references and practical pathways for e-commerce marketing practices during the transformation of programmatic advertising.