Zalando standardises data
Zalando said it built a unified data foundation on Databricks to standardise metrics for AI and analytics use across the company. (x.com). The post frames the effort as a way to ensure reliable queries and consistent metrics across dashboards. (x.com)
Zalando said on April 9 that it built a company-wide data foundation on Databricks to make metrics consistent across analytics tools and artificial intelligence systems. (databricks.com) The retailer said the project uses Databricks Unity Catalog, Metric Views and Genie to separate how data is created from how it is queried, so teams can define business logic once and reuse it across dashboards, SQL and AI prompts. (databricks.com) (learn.microsoft.com) Zalando said one recurring problem was “metric divergence”: a marketing dashboard and a finance report could show different figures for the same measure, such as net revenue, because definitions lived in separate tools and pipelines. (databricks.com) A semantic layer is the part of a data system that translates raw tables into agreed business terms, like revenue or active customers. Databricks said its Metric Views let companies define those metrics once at the data layer, with permissions, lineage and metadata that can be reused by dashboards, notebooks and AI agents. (databricks.com) (learn.microsoft.com) That matters as companies push generative AI into internal analytics. Databricks said fragmented definitions that once caused conflicting reports can also lead AI agents to return inconsistent answers at larger scale. (databricks.com) Zalando’s scale helps explain the effort. The company said it serves more than 50 million active customers across Europe, works with more than 7,000 brands and partners, and runs a microservices architecture that streams terabytes of event data into its central data lake. (databricks.com) The retailer has been building AI systems for years. A March 24, 2026 company presentation said Zalando has used machine learning in areas including recommendations since 2011, search and catalogue personalisation since 2019, and a conversational shopping assistant beta since 2023. (corporate.zalando.com) Zalando’s latest annual report, published March 12, 2026, puts the data work inside a broader growth push. The company reported 2025 gross merchandise volume of €17.6 billion, revenue of €12.3 billion and adjusted earnings before interest and taxes of €591 million. (corporate.zalando.com) The timing also lines up with Databricks’ own product rollout. Databricks announced general availability for Unity Catalog Business Semantics on April 2, 2026, one week before publishing Zalando’s customer account of how the system was being used in production. (databricks.com 1) (databricks.com 2) For Zalando, the immediate goal is less glamorous than a new chatbot: get one answer to the same business question, whether it is asked in a dashboard, a notebook or plain English. (databricks.com)