{"slug":"quartile-uses-patented-machine-learning-to-optimize-e-commerce-ad-bidding","url":"https://findausecase.com/use-cases/quartile-uses-patented-machine-learning-to-optimize-e-commerce-ad-bidding","title":"Quartile uses patented machine learning to optimize e-commerce ad bidding","description":"Quartile, the world's largest e-commerce cross-channel advertising platform, built on six patented machine learning technologies, automates and optimizes ad bidding across Google, Facebook, Amazon, Instacart, Walmart and other channels. Its patented ad-performance algorithms run on the machine learning persona in the Databricks Data + AI Platform. After migrating to Databricks and Delta Lake, Quartile reduced the runtime of its daily bid-optimization job from 7.5 hours to 45 minutes and cut data storage by 80% compared to traditional data warehouses.","company":"Quartile","industry":"Technology & Software","aiCapabilities":["Predictive Analytics"],"technology":["Databricks Data + AI Platform","Delta Lake","Databricks SQL"],"deployment":"Public Cloud","problemStatement":"Quartile faced challenges storing and processing data for multiple advertising channels, batch processing over 10TB of data on a single job that applied all transformations required for data reporting, causing server unavailability, late data deliveries, and individual jobs running up to 7.5 hours every day.","solutionApproach":"Quartile evolved its data architecture from traditional SQL databases running on Azure cloud to the Databricks Data + AI Platform, using Delta Lake, Databricks Auto Loader, Databricks SQL, and Lakeflow Jobs; its patented algorithms for improving ad performance are implemented utilizing the machine learning persona in Databricks.","businessValue":"When migrating the data from a traditional SQL database to Databricks, they saw a considerable reduction in data volume, mainly due to the optimizations of Delta with version control and Parquet compacting, resulting in storage reduction from 90TB to about 18TB, and individual optimization jobs that used to take 7.5 hours now run in 45 minutes on the lakehouse.","evidence":{"band":"high"},"sourceUrl":"https://www.databricks.com/customers/quartile","dates":{"publishedAt":"2026-09-25T05:47:36.609Z","publishedAtSource":"pipeline","updatedAt":"2026-09-25T05:47:36.609Z"},"license":"Open for reading and citing with a link to https://findausecase.com/use-cases/quartile-uses-patented-machine-learning-to-optimize-e-commerce-ad-bidding. Bulk republication requires permission."}