Understanding Results
When you call the SiftlyAPI, you receive a trust score (0-100) and a risk level. The score is calculated by analyzing multiple signals from the signup data you provide. Here's what each signal category means and how to interpret the results.
Trust Score Formula
The overall trust score is calculated using a weighted formula that prioritizes the most reliable signals:
Trust Score = (Email Score × 0.40) + (IP Score × 0.30) + (Velocity Score × 0.20) + (User Agent Score × 0.10)Email Intelligence carries the highest weight (40%) because email reputation is the strongest predictor of signup quality. Network Analysis (30%) and Velocity Checks (20%) provide important context, while Device Fingerprint (10%) adds additional signal.
How email reputation, domain type, and registration patterns affect the trust score.
- ✦Business email domains boost trust
- ✦Disposable email providers lower trust
- ✦Domain age and reputation are considered
- ✦Professional email patterns increase confidence
How network origin, geolocation, and proxy detection factor into the trust score.
- ✦Residential and corporate IPs are trusted
- ✦VPN and Tor connections raise flags
- ✦Datacenter IPs receive moderate scrutiny
- ✦Geographic context informs risk assessment
How signup frequency and patterns from the same source affect the trust score.
- ✦Normal signup rates are trusted
- ✦Rapid signups from one IP raise flags
- ✦Bulk registrations from one domain are flagged
- ✦Always analyzed in real-time, never cached
How client fingerprint and device classification affect the trust score.
- ✦Standard browsers are trusted
- ✦Bot and automation signatures lower trust
- ✦Missing user agents are flagged
- ✦Device type provides additional context