As someone who has worked extensively with large scale customer databases for Fortune 500 companies, anyone who tells you this either doesn't know what they're talking about or is lying their butt off and has an ulterior motive behind their actions.
Archiving data (not purging, per se) is definitely a thing, but you wouldn't need to do that to user tables, since that type of volume is absolutely dwarfed by transactional tables, such as wager history. An inactive customer is not consuming any CPU or memory resources, so it comes down to storage. Storage is dirt cheap, and user tables aren't consuming much anyways.
Archiving data (not purging, per se) is definitely a thing, but you wouldn't need to do that to user tables, since that type of volume is absolutely dwarfed by transactional tables, such as wager history. An inactive customer is not consuming any CPU or memory resources, so it comes down to storage. Storage is dirt cheap, and user tables aren't consuming much anyways.