What is the difference between deterministic and probabilistic identity in programmatic targeting?

• A. Deterministic relies on probabilistic signals to infer identity.
• B. Probabilistic relies on known user identity data to match across devices.
• C. Deterministic uses known user identity data to match across devices; probabilistic infers identity based on signals when identity data is incomplete.
• D. They are the same approach.

Answer: (C)

Deterministic identity relies on explicit, known identifiers to map a user across devices—think a logged-in account, email address, or a device ID tied to a single user. Because those identifiers are actually linked to the same person, you can be confident about recognizing the same individual on different devices, which enables precise cross-device targeting and measurement.

Probabilistic identity, by contrast, does not depend on a known, shared identifier. It builds a likelihood that two anonymous devices belong to the same person by looking at signals like behavior, context, timing, and technical attributes. This approach fills gaps when identity data is incomplete or unavailable, but it carries uncertainty because matches aren’t guaranteed.

So the main distinction is whether you’re using verified identity data to map across devices (deterministic) or making educated inferences from signals when that data isn’t available (probabilistic). They’re not the same, and in practice many strategies mix both to balance precision and reach.