Reporting queries can put a significant load on the db, to the point that it interrupts service.
Analytical queries (like "find the average sales price across all orders over the past year, grouped by store and region") can be 100-1000x faster in an OLAP database compared to Postgres.