Case Study: Notion achieves fresher data and over $1 million in annual savings with Onehouse

Optimizing Change Data Capture (CDC) for Fresher Data and Million-Dollar Savings

Notion, a provider of an all-in-one enterprise workspace, faced significant data management challenges as its user base and data volume grew exponentially. Their existing infrastructure, built around a sharded PostgreSQL database and Fivetran for syncing to Snowflake, struggled with unpredictable data loads and expensive, slow syncs. This made it difficult to support analytics and reporting, prompting them to seek a new solution.

The company implemented a universal data lakehouse architecture using Apache Hudi, a choice supported by Onehouse. This solution provided efficient incremental processing and handling of random data updates. The results were substantial, including over a million dollars in annual cost savings, a reduction in Fivetran sync times from one week to just two hours, and the ability to build new AI-powered features for their users.

Notion

Thomas Chow

Software Engineer, Data Platform Team