Google expanded its Data Manager API in two distinct areas: offline conversion events for Google Marketing Platform and additional Customer Match data for Google Ads. These changes are intended for measurement and audience integrations, not general website SEO.
The useful implementation question is whether the API can replace fragmented upload paths without breaking attribution or weakening data governance.
What changed for offline conversion events
Google's Data Manager API announcement says offline conversion events can now be sent to:
- Campaign Manager 360.
- Search Ads 360.
- Display & Video 360.
The AdIdentifiers object adds dclid, impressionId, matchId and encrypted user IDs. Google presents three practical advantages over the older Campaign Manager 360 upload path: one schema across Google advertising products, encryption for identifiers such as email and phone, and the ability to route events to more than one destination in a request.
That can simplify an integration, but it also raises the cost of an unnoticed mapping error. A shared schema should be validated separately for each destination, conversion type and reporting surface.
What changed for Customer Match
Google also introduced IP data in Customer Match uploads through the CompositeData field. Composite data may include IP information on its own or with identifiers such as email, phone and address. Google says pairing an IP address with its observation timestamp can improve Customer Match rates beginning in the third quarter of 2026.
Google continues to support the existing user_data field, but recommends composite_data for future improvements. That recommendation is a technical direction, not permission to collect or upload any available identifier.
Privacy and governance checks
IP addresses and identity data can be personal information. Before implementation, the responsible team should verify the applicable law, Google policy, contractual terms and the organisation's own consent basis. In South Africa, this review should include POPIA obligations and the actual purpose for which the data was collected.
At minimum, document:
| Control | Acceptance question |
|---|---|
| Purpose | Is each field necessary for a defined advertising or measurement purpose? |
| Consent | Does the collection and destination match the user's consent and notice? |
| Retention | Is raw source data kept only for the required period? |
| Access | Can only authorised systems and people send or inspect events? |
| Deletion | Can the organisation honour deletion and suppression requirements? |
| Auditability | Can a reported conversion be traced to a versioned mapping and source event? |
Migration checklist
Treat the change as a measurement migration even if the website interface does not change.
- Inventory existing offline conversion uploads and their owners.
- Map source fields to the Data Manager schema and record transformations.
- Use stable event identifiers so retries do not create duplicate conversions.
- Send a controlled test set to one destination and reconcile counts.
- Compare event time, upload time, attribution windows and currency handling.
- Validate multiple-destination routing explicitly.
- Monitor accepted, rejected, duplicated and late events after release.
Our conversion tracking service explains the wider measurement layer, while the Google Ads management resource covers campaign operations.
How this affects reporting
A successful API response does not prove that the conversion is correct. The event could still have the wrong value, source, timestamp or consent state. Compare API acceptance with source-system records and downstream campaign reports.
Annotate the migration date in dashboards. If event coverage or match rates change, analysts need to distinguish better ingestion from a real change in customer behaviour.
The bottom line
The Data Manager API update can consolidate offline event delivery across Google advertising products and adds a new Customer Match input. Its value depends on a controlled implementation: minimal data, lawful use, accurate mapping, deduplication and reconciled reporting.