Hello Community Members,
We are looking to configure a Time-to-Live (TTL) for pseudo-anonymous profiles based on multiple identity attributes.
According to Adobe’s documentation (Experience Platform Pseudonymous Profiles), profile expiration applies only if the identity graph consists of identity namespaces selected by the customer (e.g., ECID, AAID, or other cookies). If a profile contains any additional identity namespace that was not selected by the customer, it will not be deleted.
From this description, it seems possible to configure TTL for profiles, but it's unclear whether this applies only to cookie-based IDs or also to hashed email and phone identifiers.
Can anyone confirm whether this setup is feasible?
@abhinavbalooni @bjoern__koth @brekrut @Harveer_SinghGi1 @RiteshYadav18
Thanks,
Arpan
Solved! Go to Solution.
Topics help categorize Community content and increase your ability to discover relevant content.
@abhinavbalooni @brekrut @Harveer_SinghGi1 - An update here, raised a support ticket to set TTL based on these 4 identities, they successfully did it and the profile with these identities got deleted.
Hi @arpan-garg
Thank you for your detailed question. Pseudo-anonymous profiles are intended to target cookie based profiles. Within your initial post you are considering Email_LC_SHA256 and Phone_SHA256 along with Device ID /ECID.
Both Device ID and ECID can be considered anonymous identifier because we are not collecting any PII related information during the session of the interaction.
Can you please clarify how Email_LC_SHA256 or Phone_SHA256 are considered anonymous? The reason for my questions is these namespaces would require an hashed version of email address or phone number. I would lean toward these identities with a cookie based identifier would be considered a session which had some level of PII in the session.
Views
Replies
Total Likes
Hi @brekrut @Harveer_SinghGi1 - Thanks for your response.
To provide some context, our customers visit the website, browse a few products, and then enter their email to express interest in a product. However, after that, they do not engage further with the brand—there is no app communication or website interaction. Despite this, we still retain their profile in the CDP, containing only their hashed email and hashed phone number.
The reason we store these hashed identities is to enable targeting via Meta or Google Ads. However, we want to delete these profiles after 90 days(ofcourse after targeting them via social channels) to prevent them from consuming our profile license if they are no longer engaging with the brand.
Currently, I don’t see a feature in AEP that allows automatic deletion of these prospect profiles after a certain period. While we can remove them using the "delete records" feature in the Data Lifecycle, this requires external intervention and additional effort. I was wondering if a pseudo-anonymous profile could handle this deletion, which would save us time and effort.
We have a license for approximately 500,000 customers, but around 30% of them do not convert into potential customers. To optimize license costs, we would like to remove such profiles after 90 days.
Looking forward to your thoughts.
Best,
Arpan
Hi @arpan-garg
As @brekrut and @Harveer_SinghGi1 mentioned the pseudonymous profiles would only take into account cookie based namespaces. The documentation, to be honest is a bit vague, but the question under FAQ where it mentions,
it is called out anonymous cookie based namespaces.
As pseudonymous profile applies at a sandbox level, might be a good idea to try it out for a shorter duration (may be a day or two?) in a lower sandbox to test it out.
Based on what I make out of the Adobe documentation, the email and phone number would not qualify for the pseudonymous profiles deletion.
Definitely an interesting one though !
Cheers,
Abhinav
Views
Replies
Total Likes
Hi @brekrut @brekrut - If we for now consider that only cookie id can be used to set TTL , in that case do you think its a good practice to create hashedEmail and hashedPhone as two new identities in AEP of type cookie id and use this 2 new identities instead of Email_LC_SHA256, or Phone_SHA256. With this we can solve all the problems.
Views
Replies
Total Likes
One drawback of this could be that for non-pseudoanonymous profiles that have more then 50 identities in graph, Identity Service will apply a “first-in, first-out” mechanism and deletes the oldest identity. Cookie IDs are first in the priority for deletion followed by Device ID and then others like Cross-Device ID (CRM IDs, loyalty IDs), Email, and Phone.
So you might want to consider if you want to convert hashedEmail and hashedPhone to Cookie IDs.
Views
Replies
Total Likes
Hi @arpan-garg ,
I understand your confusion as all the help resources take cookie based IDs in TTL setup examples. I haven't had the experience of using this option but I'd assume it to work with any namespace that you provide at the time of TTL setup.
I'd request @Saswata Ghosh and @Danny-Miller to chime in as the have written a detailed blog explaining the Cleanup up your pseudonymous profiles in AEP and they might have the information you need.
Cheers!
Views
Replies
Total Likes
@Harveer_SinghGi1 - The confusing part for me here is it mentions cookie based id's + Identities mentioned by customer while setting up TTL.
Any no-where it mentions that the identities mentioned by customer should be cookie based only.
Views
Replies
Total Likes
Any other pointers @arijitg @DavidRoss91 @Parvesh_Parmar @Anil_Umachigi @Ankit_Chaudhary
Views
Replies
Total Likes
@abhinavbalooni @brekrut @Harveer_SinghGi1 - An update here, raised a support ticket to set TTL based on these 4 identities, they successfully did it and the profile with these identities got deleted.
Thanks for the update @arpan-garg ! Nice learning and an interesting use case ! That also calls for the documentation to be a bit less ambiguous with the focus on cookie based ids.
Cheers,
Abhinav
Views
Replies
Total Likes
Views
Likes
Replies