Entity resolution

Entity resolution is the data-matching process that decides which records, in one dataset or several, refer to the same real-world thing, such as a person, household, business or product, and links them under one ID.

Customer data rarely agrees with itself. The same shopper appears as "J. Smith" in a loyalty file, as an email address in web orders and as a phone number in a call-centre log. AWS describes its Entity Resolution service as matching, linking and enhancing related records stored across applications, channels and data stores. It offers rule-based matching, machine-learning matching and matching led by data service providers (AWS Entity Resolution User Guide, accessed 9 November 2026).

Most pipelines follow the same steps. They standardise fields such as names, addresses and emails, group likely candidates so not every pair is compared, score each pair, decide with rules or probabilities, and cluster the matches under one ID. Two kinds of error matter. A false merge joins two different people. A false split leaves one person as two records, which inflates reach and breaks frequency caps. Identity resolution is entity resolution applied to people and their devices for marketing.

Example: A retailer's loyalty record "J. Smith, 10 Elm St" and a web order from john.smith@example.com share a phone number. A rule that matches on normalised phone plus surname links them, so the retailer counts one customer instead of two.

Related terms: identity resolution, CDP, hashed email, household graph, clean room, RampID.

Authoritative source: AWS, "What is AWS Entity Resolution?" (user guide, accessed 9 November 2026).

Read the full guide: AWS Clean Rooms vs Snowflake vs LiveRamp: entity resolution and activation.