Section three: Adjusting Spam Scores

The filtering software assigns scores to email messages based on content within the body and headers.  It looks at small pieces of each message as well as the message as a whole to determine its “spaminess”.  Many hundreds of tests are performed, with each test contributing to the overall score of a message.  Several decisions need to be made once this score has been established.  Users can tell the system what to do with a message based on the score.

Adjusting these settings can be done from the “Spam” link within the “Mail” category on the left navigation menu.  An explanation of each of the settings on the “Spam Options” screen follows.


Tag Level:

If the Spam score of a message exceeds the value specified for the “Tag Level”, a header will be added to that message that gives information about the tests that contributed to the score.  This is an informational header and does not necessarily determine whether a message is considered to be Spam.  The header will not be visible during normal viewing of a message.  It is almost always a good idea to include this header for troubleshooting purposes, and doing so will not hurt anything.  To do so set the value of “Tag Level” to a very low negative number (like -999) to ensure this information is added to the headers of each message.

 

Hit Level: 
The most important setting related to Spam scores.  If a message scores at or above this level the system will consider it to be Spam.  The default setting is 5, and this is the recommended starting point.  If too many messages are getting through, increase the score a little bit at a time.  If too many messages are being tagged as Spam, decrease the score a little bit at a time.  Decimal values are allowed for any of these score values, so adjustments of 0.2 or 0.5 in either direction may be made to start with in order to get a feel for what to expect.

 

Kill Level: 
At the value specified for “Kill Level” the system will not deliver a message to a mailbox.  Instead, it will be sent to a quarantine.  This quarantine is only available through a website at webmail.example.com/mailzu, again where example.com is replaced with the user's domain name.  A valid email address password is required to view a user's messages within the quarantine.  Once signed in, messages can either be deleted or released.  If they are deleted, they are gone forever.  If they are released, they will be delivered to the mailbox, most likely to the “Spam” folder.  USE THIS SETTING WITH CAUTION!  Messages that are quarantined will be automatically deleted after fourteen days.  There will be no way to retrieve them after that.  If a user wants to make use of this quarantine, setting the “Kill Level” to a value between 20-25 may be a decent starting point.  There is a high likelihood that a message is Spam if it scores that high.  To avoid messages being sent to the quarantine, set “Kill Level” to a very high positive score (i.e., 2000).  If “Kill Level” is set to a value lower than “Hit Level”, messages will go directly to the quarantine rather than a “Spam” folder.  Such a configuration is not recommended.

 

Rewrite Subject:
If this option is enabled, the string “***SPAM*** “ will be added to the Subject header of messages considered to be Spam.  Modifying visible portions of an email in such a way is not recommended, so this is disabled by default.  However, some people do like to have a visual indicator.

 

Skip Spam Checks:
This option overrides the other settings.  If this is enabled, Spam filtering will be disabled.  This is not recommended, but users who receive low volumes of email may not want their messages to be filtered.  NOTE:  This does not guarantee message delivery.  A number of tests are done at the email gateway before messages are ever sent to the Spam filter.  Such tests are system-wide and not configurable by the end user.  They are in place to verify the validity of the email servers sending the messages to our users.