Separation of data between accounts makes them fall under different retrieval requirements.

As one account, a request for all of the data from that account contains both chunks. Separation of those accounts separates the need to accommodate requests for data from one on the other.

It can also mean that internally they may have a sufficient mechanism that data that was previously identifying to no longer being identifying (breaking userid to data pairings for example) which is sufficient to “anonymize” the data that it no longer needs to be reported or maintained.