Each e-mail is assigned a numerical score, generated by our anti-spam engine. The initial score of a message is "0". We use many techniques to scan each message to see how "spammy" it is. The cumulative value of each test becomes the spam score of the message.
We have two thresholds, defined for each domain, that determine what happens to each message. The more spammy a message is, the higher the score. If the score reaches the tag threshold the e-mail will be tagged. If the score reaches the reject threshold the e-mail will be rejected.
Similarly, we look for evidence that the message is legitimate, reducing the spam score. Thus, the spam score can be a positive or negative number. The higher the number (positive) the more spammy it is; the lower the number (negative) the less spammy.
Tests that result in a high impact are examined first: virus scanning, black/white listing, sender history, etc. These tests take precedence; they can set the message result by themselves and may cause other tests to be skipped.
Some very expensive tests can get very good information about the sender; but they are done last and only if the test can change the disposition of the message.
We examine the traffic patterns between the sender and recipient. For legitimate senders, as their traffic history accumulates, their spam scores drop until the sender becomes implicitly white listed. This ensures their messages will never be blocked in error.
If the message is not accepted or rejected by the high impact tests, it is then classified based on it's spam score and the Tag and Reject thresholds defined for the recipient.
PerfectMail™ uses three categories when scoring messages:
Note: Messages containing viruses, unwanted file attachments, or known Phishing (fraudulent) messages are always rejected.