INTEL_PROTOCOL_02 // PROTOCOL_SPEC
Identity Resolution Rules
Merge/split logic and match definitions that prevent audience duplication across platforms—so reach, frequency, and measurement reflect reality.
The Duplication Tax
Without deterministic identity rules, the same account appears as multiple entities across CRM and DSP graphs—driving waste, fatigue, and false performance lift.
CRITICAL FAILURE MODES
- ✕ Accounts double-counted across audiences
- ✕ Frequency caps fail because identity isn’t stable
- ✕ Measurement over-attributes wins to media due to duplicate exposure
Merge / Split Logic
MATCH CANDIDATES
Domain
Company Name Variants
Subsidiaries
DUNS/NAICS
CRM Parent/Child
→
RESOLVED ENTITY
Canonical Account ID
Logic: [Parent/Child Rules] + [Domain Truth] + [Exception List]
PLATFORM ON (DSP Identity Graph)
We define canonical account IDs and map them into DSP account audiences to reduce duplicate reach and enforce consistent frequency controls.
Rule: Subsidiary domains map to Parent Account unless exception = true
PLATFORM OFF (CRM Identity First)
We establish canonical CRM account identity (parent/child + domain truth) and push stable IDs into downstream activation lists.
Rule: If ParentAccountId exists → use ParentAccountId as CanonicalID
Artifacts You Receive
- ✓Canonical Account ID Spec (fields + precedence order)
- ✓Merge/Split Rules + Exception List
- ✓Duplicate Detection Queries (before/after validation)
- ✓Frequency Integrity Checklist (cap + overlap QA)
Implementation Steps
WK 1
Identify duplicate clusters across CRM and activation audiences; define canonical identity fields.
WK 2
Publish merge/split rules and exception logic for subsidiaries, acquisitions, and edge cases.
WK 3
Enforce identity in exports + reporting; validate frequency control and deduped reach metrics.