![rw-book-cover](https://miro.medium.com/v2/resize:fit:1200/1*YkbQNpHKMzO5nZ9vBNsWnA.jpeg) ## Metadata - Author: [[quentin-fernandez|Quentin Fernandez]] - Full Title:: How We Made Our Reporting Engine 17x Faster - Category:: #🗞️Articles - Document Tags:: [[adtech|Adtech]] [[data-lake-alternatives|Data Lake Alternatives]] - URL:: https://medium.com/teads-engineering/how-we-made-our-reporting-engine-17x-faster-652b9e316ca4 - Finished date:: [[2023-04-15]] ## Highlights > ![](https://miro.medium.com/v2/resize:fit:1400/1*jaaSuNaZdzzxC3HXuNT4_Q.jpeg) ([View Highlight](https://read.readwise.io/read/01gy2jhet31q68p0qg45rsy0rv)) > For example, for the `creative_type` dimension we can now group directly in BigQuery without joining, instead of pulling all creatives and checking their type from the dimensions data ([View Highlight](https://read.readwise.io/read/01gy2jjxt200bgcvhmhesg6hpk)) > we were able to reduce the cost of the reporting engine by a **factor of 10** by removing the EMR cluster and only focusing on a small service that forwards query and stream results. ([View Highlight](https://read.readwise.io/read/01gy2jmxe60gtf13ffrwzk08re))