Restrictions on Segment Evaluation method depending on nature of data | Community
Skip to main content
M
Md__AjazAn
Level 3
November 30, 2023
Solved

Restrictions on Segment Evaluation method depending on nature of data

  • November 30, 2023
  • 2 replies
  • 1706 views

I assume there is NO restriction on Segment Evaluation method depending on nature of data whether it is sourced from Batch Vs Streaming Sources. Once the data is ingested into AEP datalake, it does not matter from where the data came and all evaluation methods viz. Batch, Streaming or Edge can be applied to the Segment.

 

However, logically applying Batch evaluation method to Streaming data makes less sense unless there is a specific need that a particular Segment need to be evaluated in Batches. The same applies in reverse scenario.

 

Kindly confirm. Thanks!

This post is no longer active and is closed to new replies. Need help? Start a new post to ask your question.
Best answer by abhinavbalooni

  Hey @md__ajazan 

 

You can make a streaming segment based on batch data but it will not act as streaming. Cause the data would be processed in a batch manner i.e 24 hours. So even if you set a segment as streaming in the UI, in essence it will not behave truly as streaming. In some instances, if your data is based on a batch source, it may so happen that even if you set a segment to streaming, AEP may revert and set it to batch.

 

 

I would suggest to also have a read of this link: https://experienceleague.adobe.com/docs/experience-platform/segmentation/ui/streaming-segmentation.html?lang=en#:~:text=Streaming%20segmentation%20works%20on%20all,it%20qualifies%20for%20streaming%20segmentation.

 

2 replies

abhinavbalooni
Community Advisor
abhinavbalooniCommunity Advisor
Community Advisor
November 30, 2023

Hey @md__ajazan 

There is a implicit restriction on the kind of segment you can create based on the velocity of data - batch or streaming. Since batch data has an inherent delay, even if from the UI you select edge, it will ofcourse not be real time. What you will get though is that once a user falls in a segment based on the batch data, they can be activated on streaming and edge (near real-time) basis.

 

Also, there are specific scenarios only where you can create an edge segment defined here: https://experienceleague.adobe.com/docs/experience-platform/segmentation/ui/edge-segmentation.html?lang=en

 

Hope the above helps.

 

Cheers.

    M
    Md__AjazAnAuthor
    Level 3
    December 1, 2023

    I hope there is no restrictions for Batch and Stream evaluation as for Edge (URL shared above). In other words, I can apply either Batch or Stream irrespective of nature of data and its source from where it is coming.

      abhinavbalooni
      Community Advisor
      abhinavbalooniCommunity AdvisorAccepted solution
      Community Advisor
      December 1, 2023

        Hey @md__ajazan 

       

      You can make a streaming segment based on batch data but it will not act as streaming. Cause the data would be processed in a batch manner i.e 24 hours. So even if you set a segment as streaming in the UI, in essence it will not behave truly as streaming. In some instances, if your data is based on a batch source, it may so happen that even if you set a segment to streaming, AEP may revert and set it to batch.

       

       

      I would suggest to also have a read of this link: https://experienceleague.adobe.com/docs/experience-platform/segmentation/ui/streaming-segmentation.html?lang=en#:~:text=Streaming%20segmentation%20works%20on%20all,it%20qualifies%20for%20streaming%20segmentation.

       

        P
        PeteToasty
        Level 2
        December 14, 2023

        I found this article really helpful in understanding segment/audience evaluation.

        https://blog.developer.adobe.com/understanding-segment-timing-in-adobes-real-time-cdp-60e19558347a

        "Batch data can be used in a streaming segment, but the ingestion of batch data does not trigger segmentation."

        Terms & ConditionsAccessibility statement

        Loading footer...