Big Data-Driven Consumer Behavior Analysis and Personalized Marketing Strategy Recommendation System

In the field of digital marketing, enterprises commonly face challenges such as suboptimal user click-through rates (CTR), conversion rates (CVR), and average order values (AOV). This paper proposes a big data-driven end-to-end consumer behavior analysis and marketing strategy recommendation system. The system employs a Lambda architecture for integrated batch and stream processing, combines dynamic user profiling (enhanced with sequential behavior embeddings and knowledge graphs) and a multi-task DeepFM model to jointly optimize product recommendations and marketing strategy matching. A reinforcement learning framework facilitates closed-loop feedback and continuous optimization, dynamically adjusting strategy weights to enhance long-term returns. The system demonstrates outstanding performance on core metrics, including CTR, CVR, and Gross Merchandise Volume (GMV), achieving peak values of 5.8 % CTR, 3.2 % CVR, and 3.8 million RMB GMV. This paper provides enterprises with an actionable solution for data-driven intelligent marketing practices.