Expand my Community achievements bar.

SOLVED

Profile dataset from AEP into CJA

Avatar

Level 7

Hi there,

 

We are looking into implementing CJA with our AEP ecosystem. We have customer profile attributes set up in AEP, where they are batch files updated once daily (Full file). 

 

I am wondering when we connect these datasets into CJA, will CJA only be looking at the latest snapshot of the data we import? For example, 

TimestampPersonIDProfileAttribute
Aug 11A
Aug 21B
Aug 31C

 

When we create our reports in CJA, say I want to deepdive into PersonID = 1's traffic in Aug, and I breakdown by the dimension "ProfileAttribute", will the latest record (which is Aug 3) overwrite everything and we will only see ProfileAttribute = C overwrite all the historical records in the past? ie. Aug 1, ProfileAttribute = C, Aug 2, ProfileAttribute = C?

 

I was hoping we could see the historical changes of the profile attributes.

 

Thanks

1 Accepted Solution

Avatar

Correct answer by
Employee

When you use a profile lookup dataset in a CJA Connection, the attributes in that dataset will be available in CJA reporting for matching PersonIDs. 

 

To answer your question, the last values in the profile lookup per PersonID are applied to all of that PersonID's events - similar to how record lookup/classifications work. Profile lookups/customer attributes enrich reporting for PersonIDs similar to how record lookup/classifications enrich a dimension.

 

Being able to analyze the historical changes of the profile attributes, often referred to as "slowly changing dimensions" would require you to ingest those same changing profile attributes as an event dataset, with timestamps. Then you would be able to better associate those profile attribute values closer to when the Person's other events happened. 

 

Hope this helps.

View solution in original post

3 Replies

Avatar

Correct answer by
Employee

When you use a profile lookup dataset in a CJA Connection, the attributes in that dataset will be available in CJA reporting for matching PersonIDs. 

 

To answer your question, the last values in the profile lookup per PersonID are applied to all of that PersonID's events - similar to how record lookup/classifications work. Profile lookups/customer attributes enrich reporting for PersonIDs similar to how record lookup/classifications enrich a dimension.

 

Being able to analyze the historical changes of the profile attributes, often referred to as "slowly changing dimensions" would require you to ingest those same changing profile attributes as an event dataset, with timestamps. Then you would be able to better associate those profile attribute values closer to when the Person's other events happened. 

 

Hope this helps.

Avatar

Level 7

Hey @bryan_skelton, thanks for clarifying. 

 

So what I am hearing is, the schema class of the customer profiles need to be in "XDMExperienceEvent"? Just wondering what are some of the application of XDMIndividualProfiles then? Wouldn't it then makes more sense to have everything as event dataset as opposed to profile attributes, as that would include historical changes + current values in CJA, at the same time you can build the same audience with the events dataset. 

 

Thanks

 

 

Avatar

Employee

I would not be able to say one way or the other what is best for you, only presenting options to your problem of being able to analyze changing profile attributes. Many factors to consider in whether to use both a profile lookup and an event version or one or the other, for example:

  • The Profile lookup dataset may be used elsewhere in your AEP efforts
  • impact on person and event counts in CJA by using an event version of your profile data.

Good luck in your evaluation.