What are the implications of using a combination of batch and streaming web data? | Community
Skip to main content
N
NickLe11
Level 1
July 10, 2026
Question

What are the implications of using a combination of batch and streaming web data?

  • July 10, 2026
  • 1 reply
  • 50 views

We currently have 3 report suites in Analytics which we have set up the Analytics Data Connector to pull those into AEP so they can be used in CJA. 2 of the 3 report suites have migrated from legacy Launch to SDK and are now being captured in our AEP SDK dataset. I am thinking to remove the 2 ADC datasets from CJA and replace them with the AEP SDK dataset but I’m conscious the AEP SDK dataset will be real-time and the ADC dataset will continue to be batch.

 

What are the considerations we should be aware of when looking at reporting data that is partially real-time and partially batch?

1 reply

Devyendar
Devyendar
Level 7
July 22, 2026

​@NickLe11

I would first validate whether the ADC datasets and the AEP Web SDK dataset contain overlapping events. If Web SDK is sending data to Adobe Analytics and also into an AEP SDK dataset, then the Analytics Source Connector can bring the Adobe Analytics copy of those same hits back into AEP. If both datasets are included in the same CJA connection for the same report suites and date range, CJA may treat them as separate event rows.

Key considerations:

  • Duplicate data: Avoid including ADC and SDK datasets for the same report suite and overlapping dates if they contain the same events.
  • Metric inflation: Duplicate rows can inflate page views, events, orders, revenue, calculated metrics, and other reporting totals.
  • CJA usage/licensing impact: Duplicate rows can also increase reportable row volume in CJA.
  • Cutover strategy: Use ADC for historical/pre-migration data and SDK dataset for post-migration data, with a clear cutover date and no overlap.
  • Real-time vs batch freshness: SDK data may be available sooner, while ADC data may lag. Current-day reporting can look incomplete or uneven until the batch data catches up.

I would test this in a separate CJA connection/data view first and compare key metrics before changing the production connection.

Terms & ConditionsAccessibility statement

Loading footer...