Identity stitching

Identity stitching is the process of linking the various identifiers a single person leaves behind — an anonymous device or cookie ID before login, a user ID after login, a separate mobile app ID, and sometimes offline identifiers — into one connected profile, so that behavior can be analyzed as belonging to one person rather than several disconnected fragments.

Deterministic stitching links identifiers using a confirmed match, most often when a user logs in and the analytics system can join a prior anonymous ID to the known account ID. Probabilistic stitching instead infers a likely match from signals like device fingerprints, IP address, or behavioral similarity, trading certainty for broader coverage. It is closely related to identity resolution, a broader term that covers matching identities both within a single analytics system and across separate data sources such as CRM and support systems.

Identity stitching matters because without it, a single customer's web visit, app session, and later purchase look like three unrelated strangers, understating retention, fragmenting attribution, and producing an inflated view of unique users. The main pitfalls are probabilistic matches that are simply wrong, stitching logic that changes over time and silently rewrites historical user counts, and privacy exposure — because building a fuller person-level profile raises the same regulatory considerations as General Data Protection Regulation and similar rules around combining personal data.