Why identity resolution exists
Most online research does not arrive as a clean CRM record. Identity resolution attempts to associate available identifiers or events with a person, household, address, or account. In an intent-data workflow, it is the bridge between a behavioral signal and an audience that can be reviewed or activated.
The bridge is probabilistic or deterministic depending on the inputs. Responsible product language does not treat every result as certain.
The four questions that prevent confusion
First, what entity is matched: person, household, device, address, or company? Second, what evidence supports that association? Third, when was the identity and each delivered field last validated? Fourth, how does the system represent uncertainty, conflicts, and removal requests?
These questions matter more than a single “match rate.” A rate is meaningless until the denominator, entity, field requirements, and accuracy standard are defined.
Field quality is not identity quality
A record may have a valid email address and still be matched to the wrong research event. Conversely, a high-confidence household association may contain a stale phone number. Evaluate the signal, match, and contact fields separately.
Build a data dictionary that states the source, timestamp, transformation, validation method, and null behavior for every field used in a workflow.
Sample-based validation
Request a representative sample for a specific market and category. Measure duplicates, required-field completeness, invalid contacts, obvious household conflicts, geography fit, and suppression handling. If ground truth is unavailable, avoid presenting the sample as a definitive accuracy test.
After activation, measure delivery and response patterns by confidence or source cohort where possible. The operational evidence is often more useful than a vendor-wide headline number.
Governance after the match
Identity resolution does not create permission. Teams still need an approved purpose, channel rules, suppression, retention limits, secure access, and a way to honor disputes or deletion. Do not use a marketing identity match as the sole basis for property, legal, lending, or eligibility claims.
The most mature programs preserve provenance through export. When a record reaches the CRM or advertising platform, the team should still know why it qualified, when the signal occurred, and which restrictions follow it.
Quick reference
Frequently asked questions
What is identity resolution?
Identity resolution is the process of associating identifiers or events with a person, household, address, or account record using available matching evidence.
Is an identity match always certain?
No. Match quality varies by inputs, method, level, and time. Buyers should ask how uncertainty is represented and validate a representative sample.
What is more important: match rate or accuracy?
A high match rate is not useful if precision is poor. Evaluate match definition, precision, required-field completeness, contactability, and downstream outcomes together.
Further reading