Hi folks,
I am trying to bridge offline and online data in CJA. However, weird numbers or paths appear as if something is wrong with the implementation.
Considering that there s no problem with the Web SDK, and we think to have implemented offline data correctly.
what troubleshooting steps can i go through?
if anyone has had a similar experience before, what solution have they found?
Solved! Go to Solution.
Topics help categorize Community content and increase your ability to discover relevant content.
Views
Replies
Total Likes
Hi @Viwuwei ,
Check your identity namespace configuration:
Ensure that your offline and online data share a consistent, common identity namespace (e.g., CRM ID, hashed email, customer ID).
Confirm your Identity Graph or Identity Stitching rules are correctly set up in Adobe Experience Platform (AEP).
Validate identity values are formatted consistently (case-sensitive, trimming, hashing).
Verify schema matching:
Confirm that your offline data schema exactly matches the XDM schema for ingestion into AEP/CJA.
Check for differences in field names, field types, and structures between offline and online data.
Timestamps & Timezone Consistency:
Make sure timestamps from offline and online data are consistent, properly formatted (ISO 8601), and aligned to the correct timezone (preferably UTC).
Verify batch uploads:
Check ingestion logs in Adobe Experience Platform to confirm successful ingestion.
Validate the dataset status (successfully ingested vs. errors or warnings).
Sample offline data QA:
Manually examine some offline data records in Adobe Experience Platform to confirm data integrity and correct formatting.
Confirm your data view includes both online and offline datasets.
Ensure sessionization settings and identity settings within the data view are properly defined (session timeout, data source priority).
Create segmented queries or reports isolating offline-only vs. online-only data to find inconsistencies.
Identify where exactly the weird numbers or unusual paths appear:
Is it across all records or isolated cases?
Is there a common pattern (certain identities, certain events, or certain time periods)?
Utilize Adobe Experience Platform Debugger and Query Service:
Run SQL queries in Query Service to validate data integrity.
Check if paths and identities correlate correctly at raw data levels.
If you suspect mismatches or corrupted identities, reprocess the identity graph and re-stitch identities.
Validate by re-running queries afterward to see if the weird paths/numbers resolve
Thanks.
Pradnya
Hi @Viwuwei ,
Check your identity namespace configuration:
Ensure that your offline and online data share a consistent, common identity namespace (e.g., CRM ID, hashed email, customer ID).
Confirm your Identity Graph or Identity Stitching rules are correctly set up in Adobe Experience Platform (AEP).
Validate identity values are formatted consistently (case-sensitive, trimming, hashing).
Verify schema matching:
Confirm that your offline data schema exactly matches the XDM schema for ingestion into AEP/CJA.
Check for differences in field names, field types, and structures between offline and online data.
Timestamps & Timezone Consistency:
Make sure timestamps from offline and online data are consistent, properly formatted (ISO 8601), and aligned to the correct timezone (preferably UTC).
Verify batch uploads:
Check ingestion logs in Adobe Experience Platform to confirm successful ingestion.
Validate the dataset status (successfully ingested vs. errors or warnings).
Sample offline data QA:
Manually examine some offline data records in Adobe Experience Platform to confirm data integrity and correct formatting.
Confirm your data view includes both online and offline datasets.
Ensure sessionization settings and identity settings within the data view are properly defined (session timeout, data source priority).
Create segmented queries or reports isolating offline-only vs. online-only data to find inconsistencies.
Identify where exactly the weird numbers or unusual paths appear:
Is it across all records or isolated cases?
Is there a common pattern (certain identities, certain events, or certain time periods)?
Utilize Adobe Experience Platform Debugger and Query Service:
Run SQL queries in Query Service to validate data integrity.
Check if paths and identities correlate correctly at raw data levels.
If you suspect mismatches or corrupted identities, reprocess the identity graph and re-stitch identities.
Validate by re-running queries afterward to see if the weird paths/numbers resolve
Thanks.
Pradnya
I will recommend to check below steps to troubleshoot offline and online data issues in CJA:
Check identity stitching – use a common ID (e.g., ECID or CRM ID).
Validate schema and mapping – ensure correct XDM structure and field mapping.
Review timestamps – make sure they’re accurate and in ISO format.
Verify merge policies – confirm correct dataset prioritization.
Ensure data is in event dataset, not profile dataset.
Use debugger tools – Adobe Debugger, Query Service, etc.
Inspect data views – make sure offline data is properly included.
I hope the above reference helps clarify your queries.
Views
Replies
Total Likes
regarding your event dataset statement, that s not completely correct. you must have at least 1 event datasets but you can have profile datasets too.
Views
Replies
Total Likes
Hi @Viwuwei
similar to what @KumarRishii and @pradnya_balvir mentioned,
A few things I would definitely check -
Make sure both online and offline data are using the same ID (like hashed email or CRM ID) and that formatting is consistent (no extra spaces, case mismatches, etc.).
Double-check that your offline dataset matches the same XDM schema as your online one. Even small field mismatches can throw things off.
Make sure all your timestamps are in ISO format and ideally in UTC — mismatched timezones can cause strange pathing behavior.
Confirm that both datasets are included in your CJA data view, and sessionization settings are correctly configured.
These impact which ID “wins” when multiple are present — worth validating if the stitched user journey looks broken.
Use Query Service to look up a few known IDs and confirm their events are stitched together correctly. Helps isolate if the problem is at the ingestion or stitching layer.
Hope that helps!
Views
Likes
Replies
Views
Likes
Replies