Data is constantly or periodically loaded into a data warehouse. The database can grow very big. The bigger it gets, the slower it works. The size of the database is limited by the capacity of the disk storage, so part of it needs to be deleted from time to time. Deletion from a very big table can also be quite slow.
The data that is newer is usually queried more often. Business users could check the reports of the last day every morning, of the last week every Monday, and of the last month at the beginning of the next month. It's common to compare the results of a time period with a corresponding previous period; for example, the current month as compared to the previous month, or to the respective month one year ago. It's unlikely that somebody would query ten-year-old data.
It would be nice to keep the newer, more-queried data in one relatively small...