Trust Score Methodology
Every Scamora Trust Score is built from explainable, auditable signals. No black boxes. No opaque AI. Just transparent rules you can understand.
Every score factor is named and explained in plain English on each entity page.
No opaque machine learning. Rules are deterministic, auditable, and consistent.
New entities start at a neutral position. Evidence builds the score, not assumptions.
The Six Score Components
Average star rating weighted by recency, review volume, detail, and verified reviewer status. A higher volume of detailed, recent reviews from verified users scores highest.
Behavioral signals including burst detection, duplicate content analysis, same-IP clustering, account age, and fraud score. Suspicious review campaigns reduce this component significantly.
Domain age via RDAP, claimed ownership, MX record presence, country of registration, and contact information availability.
HTTPS, valid SSL certificate, HSTS enforcement, security headers (CSP, X-Frame-Options, X-Content-Type-Options), and DNS security records (SPF, DMARC).
Whether the entity owner has claimed their profile, responded to reviews, and maintained a complete public presence.
Basic availability checks. Is the site reachable? Does HTTP redirect to HTTPS? What HTTP status does it return?
If an entity has multiple verified high-severity scam reports, the Trust Score is capped at 40 regardless of other factors. This cap is removed once reports are resolved or dismissed by moderators. The cap is always shown publicly so you know it is in effect.
Confidence Levels
20+ reviews and technical checks completed. Score is well-supported by data.
5–19 reviews or technical checks available. Score is directionally reliable.
Fewer than 5 reviews and limited checks. Treat this score as preliminary.
The star rating reflects user satisfaction only. The Trust Score is a broader measure incorporating security, legitimacy, and authenticity signals. An entity can have a high star rating but a lower Trust Score if it lacks technical signals or has suspicious review patterns.