{"slug":"flaconi-personalizes-beauty-product-recommendations-in-real-time-with-databricks","url":"https://findausecase.com/use-cases/flaconi-personalizes-beauty-product-recommendations-in-real-time-with-databricks","title":"Flaconi personalizes beauty product recommendations in real time with Databricks","description":"Flaconi, a European online beauty retailer with 2.3 million customers, used the Databricks Data + AI Platform on AWS, with Delta Lake, MLflow, and SageMaker's DeepAR, to build a streaming machine learning pipeline for real-time product recommendations, reducing prediction latency from 20 minutes to 300 milliseconds, cutting model deployment time-to-market by 200x, reducing ML staff costs by 40%, and increasing net order income by 5%.","company":"Flaconi","industry":"Retail","aiCapabilities":["Recommendation & Personalization","Predictive Analytics"],"technology":["Delta Lake","MLflow","Amazon SageMaker","Tableau"],"deployment":"Public Cloud","problemStatement":"Flaconi struggled with massive volumes of streaming data and infrastructure complexity that was resource-intensive and costly to scale, while needing to support different infrastructure requirements for data analysts and data scientists to deliver a personalized online shopping experience.","solutionApproach":"Flaconi implemented the Databricks Data + AI Platform on AWS, using Delta Lake for reliable streaming data pipelines, MLflow for model lifecycle management, and Amazon SageMaker with the DeepAR algorithm to forecast demand, enabling streaming analytics and machine learning for real-time product recommendations.","businessValue":"Infrastructure setup time fell from two days to 15 minutes, letting the ML team deploy recommender models 200x faster; prediction latency fell from 20 minutes to 300 milliseconds; ML staff costs decreased by 40%; and the company expects increasing overall revenues, with net order income increasing by 5%.","evidence":{"band":"high"},"sourceUrl":"https://www.databricks.com/customers/flaconi","dates":{"publishedAt":"2026-09-29T05:46:40.731Z","publishedAtSource":"pipeline","updatedAt":"2026-09-29T05:46:40.731Z"},"license":"Open for reading and citing with a link to https://findausecase.com/use-cases/flaconi-personalizes-beauty-product-recommendations-in-real-time-with-databricks. Bulk republication requires permission."}