A Sift Score is a number between 0 and 100 that is the probability that a user is engaging in “bad” behavior -- 0 indicates a low likelihood and 100 indicates a high likelihood. The Sift Score is the result of all our machine learning analysis on 5,000+ signals.
Sift Scores vary between the different fraud and abuse types, i.e. Payment Fraud, Account Abuse, etc. For example, a user's Promo Abuse score of 97 indicates a very high likelihood that the user is abusing or will abuse a promotion. This score does not reflect the likelihood that the user in question is engaging or will engage in payment fraud, content abuse, or account abuse.