Every email ever sent has its digest calculated e.g. we send to alice@example.com an HTML email with the digest edeaaff3f1774ad2888673770c6d64097e391bc362d7d6fb34982ddf0efd18cb on the 14th of August 2022. This is done for every single email.
The actual contents of the email isn’t stored.
You can re-calculate the email message and then verify that the output is the same as what was originally created.
If you stored the digest for every single email sent to alice@example.com, that won’t take up anywhere near the amount of space that storing the full HTML messages would. You just need the email template and the variables that were passed in.
You could look at all the emails your org has ever sent to alice@example.com to understand the messaging she is receiving, whether she gets repeated or conflicted messages or if she just receives a lot of noise.
Benefits
Not storing private information unnecessarily — we don’t actually store the content of the messages.
Save on storage costs.
Able to see the experience from any user’s point-of-view at any point in time.