The core of Group-Level Unification (GLU) relies on identifiers to recognize and unify users across different panels.
For GLU to work, the value of an identifier needs to match across panels. For example, Brand A uses “Customer Email” and Brand B uses “User Email” as identifiers, and the email address is the same (e.g., sample@insiderone.com). When these identifiers are matched, the system can unify those users in the Unified Customer Database.
Each panel has its own set of identifiers. To unify users across GLU accounts, one identifier is selected to unify users across the other panels. This chosen identifier must be one of the existing identifiers within the individual panels.
Requirements
For GLU to reliably detect common users across brands, the chosen identifier must be configured the same way on every connected panel:
It must be the top (highest-priority) identifier. GLU uses the highest-priority identifier on each panel to match users. If the common identifier sits lower in the priority order, matches will be inconsistent.
Its limit must be set to 1. This ensures each user carries only one value for that identifier, which is what makes a reliable cross-brand match possible. Without this, the same user can end up with multiple identifier values, and matching will not work as expected.
These two settings are essential. They are what turns a panel-level identifier into a dependable cross-brand key.
Data stays where it belongs
No user data is moved or duplicated across individual panels. The data of each brand remains separate, so each panel continues to own its own user profiles and can run personalized campaigns independently.
User permissions across panels
GLU does not override user permissions. Opt-in and opt-out choices remain panel-specific, and each brand continues to respect its own consent settings.
For example, if a user is SMS-opted in on Brand B but not on Brand A, no messages will be sent from Brand A to that user. Even though GLU recognises them as the same person on Brand B. The unified view is used for recognition and targeting logic, not for overriding consent.