Data Platform Evolution and Digital Practice in the FMCG Industry: A Baicaowei Case Study

Introduction The rapid growth of online consumption driven by platform economies and big‑data technologies has created new growth opportunities for fast‑moving consumer goods (FMCG) brands. This presentation explores how data can empower the FMCG industry to improve business capabilities.

1. Data Platform Evolution Insights

1.1 Company Overview Baicaowei is a leading omni‑channel snack brand with over 1.4 billion users and more than 1,000 SKUs, aiming to provide healthy, trustworthy food.

1.2 Data Architecture and Technology Selection The architecture emphasizes metadata management, permission control, and governance, choosing StarRocks for unified data access, Apache Iceberg for table format, Apache Hudi for data lake construction, and Delta Lake for unified storage. Compute frameworks include Presto, Hive, and Spark, while scheduling relies on XXL‑JOB, Airflow, and DolphinScheduler.

1.3 Digital Process A unified data platform (DWD, DWS, ADS layers) is built on structured and unstructured data streams, supporting data applications and visualization. Identified challenges include model changes, low compute efficiency, and long data pipelines.

1.4 Rational Simplification of Architecture By adopting CloudCanal and an MPP framework (ETL/ELT, T+0/T+1, physical‑view hybrid modeling), the architecture is streamlined into ODS → DWD (detail) → DWS (summary) → APP/DWM (data marts), reducing complexity and improving efficiency.

2. Digital Practice

2.1 Metric Governance Business scenarios are mapped to metrics across dimensions such as shipping date, store, SKU, finance, and supply chain, supporting dashboards for warehouse selection, performance evaluation, new‑product monitoring, and more.

2.2 Business Modeling Modeling spans operations, finance, supply chain, customer service, and HR. Data flows from Canal, Kafka, Spark, MongoDB into the data lake, then to applications via RDB. Detailed layers (ODS, DWD, DWS) address raw data, cleaning, aggregation, and topic‑wide data marts.

3. BI and Data Insights

3.1 Business Intelligence Over 20 visual reports cover sales, refunds, and profitability. BI combines historical data with operational rules to generate demand plans, allocate inventory, and create procurement or aging analysis orders.

3.2 User Analysis for Brands A unified dashboard provides user metrics, segmentation, and lifecycle tags, enabling answers to questions about target demographics and new‑old customer behavior.

3.3 Quick‑Win Digital Scenarios A financial reconciliation example demonstrates data extraction, order‑amount matching, and discrepancy reporting, with implementation details on API signing and exception handling.

Overall, Baicaowei's experience shows how a well‑designed big‑data platform, rational architecture, and systematic metric governance can drive digital transformation in the FMCG sector.