Identity stitching is the foundational capability that resolves customer identities across devices and channels, creating a single, unified profile. Without accurate identity resolution, data remains siloed, making effective personalization and cross-channel orchestration impossible. While the other options are important functions, they rely on the unified profile created by identity stitching.

For real-time profile updates from streaming data, the configuration of streaming segments is key. Specifically, the ‘membership expiration’ setting dictates how long a profile remains in a segment after a qualifying event. If this is misconfigured or if streaming evaluation isn’t properly set up for the relevant segments, profiles won’t update in near-real-time. Batch schedules and merge policy priorities don’t directly control the latency of streaming data ingestion into the profile.

A core function of identity stitching within a merge policy is to link identifiers across a customer’s journey. When the user logs in, the system recognizes the ECID and CRM ID as belonging to the same person (often via a shared private graph or authenticated event). The merge policy, prioritizing the CRM ID, will stitch the ECID and its associated data to the primary CRM ID profile. This ensures all behavior, both anonymous and known, is unified into one profile.

Adobe Real-Time CDP’s segment builder includes functions like ‘Top Percentile’ specifically for these use cases. This function dynamically calculates the threshold (e.g., the 90th percentile value for order amount) and evaluates each profile against it. This is far more efficient and maintainable than trying to set a static monetary threshold (option A) or using an incorrect logic type (option B), and it’s fully possible within the platform without external work (option D).

This is a key architectural distinction. ‘Segment Export’ is the standard method where you publish a specific audience segment (e.g., ‘Cart Abandoners’) to the destination. Only profiles in that segment are exported. ‘Profile Export’ (sometimes called ‘Dataset Export’) is used for specific use cases where you need to send all profiles that match a general schema mapping to the destination, regardless of segment membership, often for custom analytics or machine learning purposes.

Data governance is a two-step process. First, you classify data by applying ‘Data Usage Labels’ (e.g., ‘Sensitive Health Data’, ‘No Advertising’) to fields and datasets. Second, you define ‘Marketing Actions’ (e.g., ‘Export to Advertising Destination’) and create ‘Policies’ that block actions on data with certain labels. The enforcement layer then prevents labeled data from being used in segments or exported to destinations that violate the policy. The other combinations do not create an enforceable governance control.

When multiple source datasets provide a value for the same attribute (e.g., an email from CRM and a different email from a webinar sign-up), a conflict arises. The ‘timestamp’ conflict resolution method dictates that the value from the source record with the most recent timestamp will be chosen for the unified profile. This is a common strategy for attributes where the most recent information is considered the most accurate.

Segment templates are a powerful feature for enabling business users. A technical user can create a reusable template (e.g., “Viewed Product X in Last Y Days”) with the core logic locked. The business user can then instantiate this template, simply filling in parameters like the specific Product ID (from a dropdown) and the number of days. This balances governance, speed, and ease of use far better than training on full logic building (A), using APIs (B), or creating bottlenecks (D).

The ‘Preview’ (or ‘Estimate’) function is a crucial planning tool. It runs the segment logic against a sample of profile data and provides an estimated count and a sample list of qualifying profiles. This allows marketers to refine audience criteria (e.g., adjust time windows or attribute values) to hit a target audience size or composition without actually creating and processing the full segment.

Identity stitching is the foundational capability that resolves customer identities across devices and channels, creating a single, unified profile. Without accurate identity resolution, data remains siloed, making effective personalization and cross-channel orchestration impossible. While the other options are important functions, they rely on the unified profile created by identity stitching.0