Identity resolution

Identity resolution (user stitching, in practitioner shorthand) is the process of connecting events and records that carry different identifiers into one person: the anonymous pageview, the click ID from an ad, the form fill with an email, the CRM contact, the order. Without it, one customer looks like four strangers across your sources.

#Stitching connects; deduplication removes

Stitching is routinely confused with deduplication, and they solve opposite problems. Dedup removes extra copies of one event (the same purchase sent by pixel and server). Stitching connects different events to one person. What makes a stitching implementation trustworthy is named match keys with an explicit precedence, deterministic matches first (hashed email, then a durable user ID, then click IDs, then device identifiers), each weaker than the last, plus honesty about where it fails: shared devices, consent-blocked keys, sessions broken at domain boundaries. A platform selling "360° identity" without naming its match keys is selling confidence, not resolution; probabilistic matching extends reach for ad platforms, but it's no foundation for revenue data.

Buron ships an identity-map dataset: an actual stitching implementation with exactly that key precedence, which is what its attribution cross-checks stand on.

The real mechanics (which keys match, in what order, what happens on conflicts) are the full page: Identity resolution: how user stitching actually works. The keys themselves are defined at Click ID (gclid / fbclid / wbraid), the substrate is First-party data, and Conversion tracking & signal quality maps the wider territory.