Create Unified Customer Profiles
The unification process
Unification creates the unified customer profile (golden record) by combining records from one or more source tables. It has three logical stages: source selection and mapping, match, and merge. In matching you choose the tables to include, designate the primary key, and define match rules using exact or fuzzy matching on attributes such as email, phone, or name. You then set the match order so the strongest rule runs first. In merging you define merge policies per attribute, choosing which source wins (for example, most recent or a ranked source priority).
Concepts to master
Know the difference between exact and fuzzy matching and when each is appropriate. Understand that deduplication within a single table is a supported scenario. Recognize customer ID generation and that unified profiles persist across runs. Understand relationships that connect activity tables to unified profiles, and the activities view. Know that unification can be run on demand and reviewed before publishing results downstream.
Traps to expect
A classic trap is assuming fuzzy matching guarantees correctness-it produces candidates and can create false positives, so review match results. Another is confusing match rules (who is the same person) with merge policies (which value to keep). Questions may imply that adding a source later requires rebuilding everything; in fact you can add sources and re-run. Also note that changing a primary key after unification can invalidate downstream segments.
Rule of thumb
Match on stable identifiers first (email, loyalty ID), then fuzzy attributes. Always inspect the match review output before publishing.