{"slug":"punchh-accelerates-loyalty-analytics-with-a-databricks-lakehouse-on-aws","url":"https://findausecase.com/use-cases/punchh-accelerates-loyalty-analytics-with-a-databricks-lakehouse-on-aws","title":"Punchh accelerates loyalty analytics with a Databricks lakehouse on AWS","description":"Punchh migrated from AWS MySQL to a centralized Databricks lakehouse on AWS with Delta Lake and Databricks SQL to unify raw and refined data for ML and BI teams, increasing campaign analytics dashboard processing performance by 90x, reducing dashboard costs by 50% per request, and improving customer net promoter score by 12% (40 basis points).","company":"Punchh","industry":"Technology & Software","aiCapabilities":["Predictive Analytics"],"technology":["Databricks Data + AI Platform","Delta Lake","Databricks SQL","Databricks Notebooks"],"deployment":"Public Cloud","problemStatement":"Punchh had outgrown its AWS MySQL production system as data volumes grew and needed a scalable, cost-effective centralized platform; teams and data were siloed, ML models took days to run, and Tableau reporting on raw data was slow.","solutionApproach":"Punchh built a centralized lakehouse on the Databricks Data + AI Platform on AWS, using Delta Lake to serve bronze, silver and gold data tiers, Databricks SQL for fast analytics queries, and Databricks Notebooks for collaborative data engineering and ML work.","businessValue":"Punchh reports 10x faster time-to-insight, 30% lower operational costs, and a 12% increase in customer satisfaction (NPS up 40 basis points), plus a 90x increase in processing performance on its campaign analytics dashboard and 50% lower cost per request.","evidence":{"band":"high"},"sourceUrl":"https://www.databricks.com/customers/punchh","dates":{"publishedAt":"2026-10-05T05:48:25.245Z","publishedAtSource":"pipeline","updatedAt":"2026-10-05T05:48:25.245Z"},"license":"Open for reading and citing with a link to https://findausecase.com/use-cases/punchh-accelerates-loyalty-analytics-with-a-databricks-lakehouse-on-aws. Bulk republication requires permission."}