Hey folks, hoping to tap the collective brainpower here. I’ve been wrestling with Marketo Measure (formerly Bizible) offline conversion imports failing *silently* for the past week. The UI shows the upload as "processed," but the attribution data just never materializes in our dashboards. No errors in the UI, no failed job emails—it’s like the data vanishes into a black hole 😅.
I’ve ruled out the obvious culprits:
- **File format:** CSV matches the required schema exactly (using their template).
- **Column headers:** Case-sensitive and spelled per documentation.
- **SFDC IDs:** Validated a sample against our production instance.
My suspicion is leaning towards the **processing logic** on their side, maybe a mismatch in our configured touchpoint date settings versus the timestamps in our import file. Has anyone else hit this?
Here’s a sanitized snippet of our typical import file structure:
```csv
Email,OpportunityId,OpportunityName,Amount,CloseDate,StageName,CreatedDate,AttributionTouchpointDate
lead@example.com,006XXXXXXXXXXXX,Acme Deal,50000.00,2024-05-15,Closed Won,2024-05-10T14:30:00Z,2024-05-10T14:30:00Z
```
We’re using the "AttributionTouchpointDate" column to map to the moment we want credited. Our Marketo Measure instance is set to use "Opportunity Created Date" as the primary touchpoint date—could a conflict here cause a silent discard?
Also, for those using a CI/CD approach to manage these imports (like we do via GitHub Actions), how are you handling validation and logging? I’ve started adding a pre-flight check script that validates the CSV against a JSON schema before upload, which at least catches format issues early.
Would love to hear:
- Any gotchas with date/time formats or timezone handling in the import.
- How you’ve debugged silent failures (are there hidden logs somewhere?).
- If using an API-based approach instead of the UI upload has been more reliable.
Thanks in advance for any pointers! -pipelinepilot
Pipeline Pilot
Been there! That silent fail is maddening. The "processed" status can be misleading - it often just means the file passed basic validation and was accepted into their queue, not that the data was applied.
One thing that's bitten us before: the **AttributionTouchpointDate** column value has to be *later than* the associated Lead/Contact's **CreatedDate** in Salesforce. If your touchpoint date predates the lead record creation, Marketo Measure will often discard that row without a peep. Maybe double-check that relationship for a few sample rows?
Also, have you tried a tiny file with just one or two rows with very recent dates? That sometimes surfaces issues that get lost in larger batches. Good luck!
Pipeline Pilot
The touchpoint date vs. created date check is a solid first test. Another spot to check that's less obvious is the **ActivityDate** field on the underlying SFDC task/event it might be trying to create. If that's mapped incorrectly or falls outside your configured attribution window, it can also cause silent drops.
Since you're already using the template, could you share what your "Touchpoint Date" setting is under Settings > Attribution > Touchpoint Dates? If your import's `AttributionTouchpointDate` is earlier than the window defined there, it might be filtered out during processing without logging.
owl
That timestamp alignment theory is on point. I've seen the same silent drop when the `AttributionTouchpointDate` falls outside the configured attribution model's lookback window. It's not just a date comparison, it's a window check.
Can you run a quick diagnostic? Pull the processing logs via the API, not just the UI. There's often a more granular status (`processing_complete` vs. `touchpoints_created`) and sometimes a count of rows filtered due to "invalid date" that doesn't bubble up to the interface.
Also, check your epoch for custom conversions. If you're mapping to a non-standard conversion point, the date might need to align with a different stage in your revenue cycle.