What confidence score should I treat as a reliable match?

There is no universal cutoff; it depends on the cost of a wrong decision. For high-impact actions like blocking payments or locking accounts, require a high score to avoid harming legitimate users, and for low-risk features you can accept lower values. The right threshold is the one that balances your tolerance for false positives against missed fraud.

Why is a visitor's confidence score sometimes low?

Low confidence usually means the current signals only partially match a known device, often after a browser or OS update, a hardware change, or on a privacy-hardened setup that exposes fewer distinguishing signals. It can also happen when a device closely resembles many others, making a unique match harder. In these cases the identifier is still returned, but you should weight it more cautiously.

Risk & Detection

Confidence Score

A confidence score is a numerical value from 0 to 1 that expresses how certain an identification system is that the current visitor matches a previously seen identity. Higher values mean stronger correlation between the current signals and a known device profile.

Start Free Book a Demo

How it works

How Confidence Score works

When a visitor arrives, the identification layer compares the freshly collected signals against stored profiles. The confidence score summarizes the quality of that match: how many high-stability signals line up, how unusual the combination is, and how much the current fingerprint differs from the closest known one.

The score exists because device identification is probabilistic, not exact. Browsers update, users change settings, and two devices can look similar. A confidence value lets the system return its best-guess identifier while honestly communicating how sure it is, rather than forcing a binary match-or-no-match decision that would either miss returning devices or merge distinct ones.

Integrators typically set a threshold appropriate to the action's risk. A low-stakes personalization feature might accept a modest confidence value, while blocking a payment or locking an account should require high confidence to avoid punishing legitimate users on the basis of an uncertain match.

Why it matters

Why Confidence Score matters for fraud prevention

Confidence scores are what let fraud teams tune the trade-off between false positives and false negatives. Acting only on high-confidence identifications reduces the risk of penalizing a legitimate customer whose device merely resembles a flagged one, while lower thresholds catch more repeat abuse at the cost of occasional mistakes. Exposing the score keeps that trade-off in the hands of the people who understand their own risk tolerance.

With TRACIO

How TRACIO handles it

TRACIO's Identification product returns a confidence score alongside every visitor ID so integrators can calibrate their own thresholds. Because the score reflects how well the current signals match a known device across 130+ inputs, teams can require higher certainty for destructive actions and relax it for low-risk personalization.

Explore further

Identification Identification docs Signal stability across updates

FAQ

Frequently asked questions

Identify every device with confidence

Start with a free plan of 2,500 API calls per month. No credit card required.