Skip to main content

Hi,

We use Amplitude to track activity on our B2B which requires a login, so we set our UserID to be the email address, we also store a duplicate user property called Email (just in case we want to change the UserID later).  So for every event we get pass the user’s email twice.

I’m considering data retention and wondered if there’s any way to pseudonymise these values if I want to forget a particular user, but not lose their event history?

Hi @Mark Johnson 

As a best practice, it’s recommended to avoid setting the user’s email id as the user_id since it’s both a PII and a value which might get changed in the future.

If you intend to change the user_id of an existing user in the future to their new email id, then Amplitude will count that new user id as a totally new unique user.

Re. the Pseudonymization, I’m not aware of any out of the box solution in Amplitude which can enable your specific use case.

  • There is the User Privacy API which can delete user related properties and event data with some caveats, but I don’t any function to pseudo anonymize the user ID.
  • The TTL feature allows Amplitude to delete all event data sent to Amplitude prior to your retention period. The doc mentions that “ it impacts all event data and doesn't impact user data that you have sent to Amplitude.” So not exactly sure what this means for now.

Hope this helps.


Reply