Identity resolution

In a world where customers interact with brands across multiple channels, data about those interactions is collected and stored in different systems, often with different identifiers. A customer who browses on mobile, purchases via desktop, and contacts support by phone may exist as three separate records with no linkage between them. Identity resolution connects those records to a single, unified customer profile.

There are two primary approaches:

• Deterministic matching uses exact identifiers, such as email address, phone number, or logged-in user ID, to link records with high confidence.
• Probabilistic matching uses behavioral signals, device characteristics, and statistical modeling to infer connections between records where no exact identifier is shared, with lower confidence but broader reach.

Identity resolution has significant privacy implications. The same capability that enables a better customer experience can also enable surveillance, data combination without consent, and the re-identification of individuals from nominally anonymous records. Privacy-preserving identity resolution limits matching to contexts where the individual has consented to data combination and maintains accurate suppression of opted-out or deleted records across the unified profile.