Skip to main content
← User Stories
Ant Group · China

Unifying real-time marketing data at Alipay

Alipay uses partial updates and columnar streaming to bring marketing events together and serve fresh data across its marketing platform.

About Ant Group

Ant Group is a digital technology company and the operator of Alipay, which connects consumers with payments and everyday services.

One attribution table, many event streams

Alipay’s marketing platform supports coupons, red packets, payment marketing, merchant offers, and personalized recommendations. Fluss primary-key tables merge events such as coupon issuance, redemption, rewards, and task participation through partial updates. Each upstream job writes the columns it owns, while the storage layer combines them into a shared attribution table.

Fresh data with less processing state

Moving column-level merging into Fluss reduces chained streaming joins and the state held in Flink. The team describes an improvement from multi-minute pipeline latency to second-level data availability. Column pruning lets downstream jobs consume only the fields required for their real-time analytics and feature pipelines.

Connecting streaming and lakehouse analytics

Alipay is building a unified streaming storage layer that tiers data into the lakehouse. Real-time applications read fresh data from Fluss, while batch analytics and next-day validation use the lakehouse representation. This simplifies the previous architecture of separate streaming and lakehouse storage, dual writes, and additional data-copy jobs.

Explore more user stories →