Airbnb has published the first post in a series on Minerva, the metrics platform that powers their analytics, reporting, and experimentation capabilities. In this post, Robert Chang, Amit Pahwa, and Shao Xie describe the problems, like debugging data pipelines and resolving inconsistent KPIs, that Airbnb experienced as its data team scaled and built additional tables and artifacts on top of their core data tables. In response, Airbnb developed a set of certified, vetted, normalized tables and Minerva, which denormalizes these fact and dimension tables and serves aggregated and raw data to other applications. The authors provide a high-level overview of the architecture of Minerva and briefly discuss the other tools that consume metrics from Minerva, including Data Portal, Experiment Reporting Framework, and Executive Reporting Framework. They conclude by describing how Minerva enabled data scientists at Airbnb to quickly understand the impact of COVID-19 on Airbnb’s business.