Probabilistic Matching

A method that uses varied data sets and statistical algorithms to identify the same user across different devices and applications with high probability.

Because each device carries its own unique identifier — such as a device ID or cookie — it is inherently difficult to determine whether activity on two separate devices belongs to the same person. Probabilistic matching addresses this by inferring connections based on signals like IP address, browser type, location patterns, and timing, without requiring a shared login or account.

It is typically employed when deterministic matching is not an option — for example, when no common identifier (such as an email address or social media account) exists to definitively link a user to multiple devices. Probabilistic matching can also be used alongside deterministic matching to improve overall match rates and enrich user profiles with cross-device behavioural data.