When to use it
Server-side identity-based activation is the right path when:- The Permutive SDK can’t be deployed in the activation environment — for example, in some CTV apps, smart speakers, server-rendered properties, or headless setups.
- Client-side bidstream activation isn’t viable — for example, in CTV environments (typically server-side ad-serving setups such as Prebid Server with SSAI), or when you want activation independent of any specific session.
- You want to bridge audiences across environments — make web-built audiences available to mobile or CTV partner targeting.
- You want to share audiences across business units or use cases — activate a cohort once, have it available across multiple partner workflows.
How it works
1
Identity collection
Permutive collects identity values from your environments — typically via the Permutive SDK (cookie ID on Web, IDFA on iOS, AAID on Android, hashed email when authenticated users are present), or via Direct API integrations (IP address, hashed email).
2
Scheduled upload to the partner platform
On a scheduled cadence (typically daily), Permutive uploads cohort memberships to the partner platform via its segment-management API. Each upload is keyed by the identity values you have configured.
3
Segments created in the partner platform
The partner platform creates or updates segments / audiences for each Permutive cohort, populated with the identity values of matching users.
4
Targeting in the partner platform
Those segments appear in the partner’s targeting UI alongside the partner’s other audience segments, and can be used for campaign targeting, PMP deal packaging, or any other activation the partner supports.
Delivery guarantees
Scheduled uploads are delivered to the partner platform in batches, and Permutive deduplicates batch processing internally: a batch is applied once even if it is signalled for processing more than once, so routine operation does not produce duplicated segment uploads. If a batch fails partway through (for example, during a service restart), Permutive automatically recovers and reprocesses it. By default, recovery favours completeness — at-least-once delivery — meaning the recovered upload can re-deliver some rows that had already reached the partner platform before the failure:- For platforms that ingest segment memberships through an API keyed by identity and segment, re-delivery is idempotent and has no targeting impact.
- For file-based destinations (such as S3 or GCS buckets), re-delivered rows can appear in more than one file for the affected batch.
Identity types
Identity-based activation works wherever an identity can be collected and the partner platform accepts that identity for segment uploads. Typical identity types:
The right identity depends on what your environments can collect and what the partner platform supports.
Trade-offs vs. client-side bidstream activation
Many integrations use both approaches in combination — client-side for real-time freshness on Web, server-side identity-based for environments where the client-side path doesn’t apply.
How to request a new integration
1
Contact your Customer Success Manager
Outline the partner platform, the inventory environments you want to activate against, and the identity types you can collect.
2
Identity collection review
Permutive confirms which identity types are already being collected and what additional collection (if any) would be required as part of the build.
3
Scoping and build
Permutive scopes the integration against the partner platform’s segment-management API and builds the connector. Lead time depends on the partner platform’s API maturity and any partner-side enablement required.
4
Activation in the Permutive dashboard
Once built, the integration appears in the Permutive dashboard alongside other activations. Activate cohorts as you would for any other integration.