Expand my Community achievements bar.

SOLVED

Adobe analytics audience not included in CDP Profiles

Avatar

Level 4

Hi,

 

I am ingesting Adobe Analytics as a streaming source into CDP. I am trying to create a segment only with Adobe analytics streaming data based on URL parameter contains xxxx in segment and sending ECID attribute into SFTP destination. The website that i'm pulling the Adobe Analytics data from does not have login space, so, hashed Id is not passed through Adobe Analytics into CDP. Only ECID is passing. After creating the segment, there is no profile count coming, even though I can see the Adobe Analytics relevant records in Data Lake which I queried through query service. Can anyone suggest why? Is this because there is no HashedID? 

For reference --> Schema for Adobe Analytics in CDP has Hashed ID and ECID both as Secondary Identifier and the Schema is bounded on HashedID identifier for Profile Stitching with another Snowflake ingested schema.

 

Please suggest. 

Thanks.

1 Accepted Solution

Avatar

Correct answer by
Moderator

Hi @mansibedi 

 

1. If you query the identityMap or explode it you can see values f/t within the object this denotes secondary/primary. ECID is primary by default and AAID is secondary but in case ECID is not generated then AAID is set as primary.

 

arijitg_0-1684144365354.png

(select identitymap from <AA_Dataset> limit 2;)

 

2. I think for first time backfill it takes 24 hours or more for some destinations (this should not be the case going forward if it still takes more than 24 hours feel free to raise a support ticket).

 

View solution in original post

6 Replies

Avatar

Moderator

Hi @mansibedi if I understand correctly your segment has only 1 event attribute of "URL parameter contains xxxx" but no profile attribute and it's still not populating any audience?

 

I got confused by your reference "Schema for Adobe Analytics in CDP has Hashed ID and ECID both as Secondary Identifier" because for Adobe Analytics source ECID is always the primary identity flowing through identityMap did you customize this to make secondary? Also you mentioned "The website that i'm pulling the Adobe Analytics data from does not have login space, so, hashed Id is not passed through Adobe Analytics into CDP" then why this identiy is defined and how is this contributing to identity graph?

Avatar

Level 4

Hi @arijitg ,

 

1. Yes, my segment has  1 event attribute of "URL parameter contains xxxx"  along with another Event attibute Form Start exists. Only these two parameters are there in the segment.

Like I menioned -- Schema for Adobe Analytics (AA) in CDP has defined Hashed ID and ECID both and the Schema is bounded on AA HashedID identifier for Profile Stitching with another Snowflake ingested schema (which has XDM Individual profile Class). That does qualify AA's HashedID and ECID as profile attributes right? 

 

2. ECID and HashedID in Adobe Analytics source was specified as Secondary identity in the Data Model when it was designed initially. Does this create a problem if we specify both as secondary identifier?

 

3. Reg your question --> why this HashedID identity is defined in  AA Schema --> We have 2 websites (one with login space and one without login space) from where we are pulling Adobe Analytics data into CDP as source. We have some segments that we created where we utilised Hashed Id as well. For this particular Segment use case, we are only referring to Website without login space, thus, Adobe Analytics data that we pull from this website doesn't include Hashed ID, so as per our segment definitition, we are checking if URL contains xxx and Form Start Exists, and was to send the ECIDs of those profile counts in SFTP destination. Hope this makes sense.

 

Thanks,

Mansi  

Avatar

Moderator

Hi @mansibedi 

1. Yes, profiles should qualify. As they're from AA dataset I would suggest if you can write a query to mimic the same segment logic and check count from data lake. They won't be exact but you will have an estimate.

 

2. I assumed you're using OOTB source connector for AA in that case dataflow and schema gets auto generated and if you explode identityMap you can see ECID is primary for most of them (Few records can have AAID as primary) 

 

3. If you have 2 websites do you have 2 different report suites for that? and are you trying to host 2 different report suites into a single schema  that will have it's own challenges.

Avatar

Level 4

Hi @arijitg 

 

Thank you for the response. 

 

1. & 2.  I did query the same logic in Query service and it did gave me all the events earlier that were performed on the website. Also, yes i think we have OOTB connector since dataflow and schema gets auto generated. In identity graph I can see AAID and ECID both, but nowhere it is specified it is primary, how do I check that?

 

Also, once update --> Can now see the audience in SFTP after 1.5 days of segment creation.. it shouldn't take this much time for streaming segment to evaluate and activate in destination, any idea on this? 

 

3. No no, we are using a single report suite only. For our segment (which is a test segment) we wanted audience specific to one website, and thus added URL condition in Segment.

Avatar

Correct answer by
Moderator

Hi @mansibedi 

 

1. If you query the identityMap or explode it you can see values f/t within the object this denotes secondary/primary. ECID is primary by default and AAID is secondary but in case ECID is not generated then AAID is set as primary.

 

arijitg_0-1684144365354.png

(select identitymap from <AA_Dataset> limit 2;)

 

2. I think for first time backfill it takes 24 hours or more for some destinations (this should not be the case going forward if it still takes more than 24 hours feel free to raise a support ticket).

 

Avatar

Level 4

Thanks @arijitg  for clarifications for both point 1 and 2. Yes the segments are populating in the destination now as expected. Thanks.