Spam Email Tagging Service

Our "spam" (Unsolicited Commercial Email, or UCE) tagging service is available as an option to help you identify junk e-mail as you receive it. We use a scanning program named Spam Assassin which can process an e-mail message and assign a numeric "spam level" ranking to it based on a wide variety of criteria. Each message is run through a barrage of tests to determine its "spamminess" and assign it a score. Once a message's score exceeds a pre-set threshold, the message is considered spam.

If a message is considered to be junk mail, it is tagged, or marked, in several ways. The most obvious is that the prefix of '[[ SPAM ]]' is prepended to the subject line of the e-mail. A report is also included in the headers of the e-mail which gives a brief explanation of the various tests which matched the message and contributed to the final "spam score." Another useful header added to the email is "X-Spam-Level:" and includes a "*" (asterisk) for each "spam point" a message carries (this can be useful for filtering, described later).

There are several hundred tests performed each time a message goes through the process. They include things like checking technical validity of header information, checking content for known spam phrases/links/tricks, checking online databases for known spam gateways the message may have gone through, and an intelligent classifier tool which automatically learns what messages "look" like spam and which don't (this is called the Bayesian classifier). For the most part there is no single test which definitely identifies spam, but instead multiple factors contribute to the total score which makes the tool highly accurate.

It should be noted that while the program is very accurate, it is not 100% accurate. This means you may occasionally get legitimate mail tagged as spam, as well as spam which doesn't get caught. For this reason it is a good idea to occasionally review suspected spam for any real messages.

Spam tagging can be enabled on a per-domain or per-user basis. If you're not sure how to do this, please contact the administrator of your domain or us at support@WorldDesign.com. Once spam tagging is enabled, all suspected junk mail sent to you will be altered as described above. At this point you have a few choices in how you want to deal with the tagged messages.

A) The simplest method is to do nothing and just skip over messages with '[[ SPAM ]]' in the subject line when reading your mail. You can usually sort your message listing by subject so that all tagged spam is grouped together for easy deletion.

B) If your e-mail program supports sorting messages into separate mailboxes, then you could make a new folder (called 'spam' for example) and direct all suspected junk mail into this folder for later examination and deletion. This is typically done by setting up a "filter" in your e-mail program settings. For example the filter can be set so that any message with '[[ SPAM ]]' in the Subject would be directed to the spam folder. You may also use this technique to direct suspected junk mail directly to a Trash folder if your e-mail program uses one.

C) A more advanced technique would be to set up filters based on the "X-Spam-Level" header which is added to each suspected spam e-mail. This header contains a "*" (asterisk) for each "spam point" the message scored during the tests. A message is considered spam at 5 points. A message scoring 7 points, for example, might actually be ligitimate mail you want to keep. However a message scoring over 15 points, for example, is definitely spam. So, one can set up 2 filters based on the spam level asterisks. The first filter would be "if header X-Spam-Level contains *************** (15 asterisks) then send message to trash." And the second filter would be "if header X-Spam-Level contains ***** then send message to spam folder." The first filter sends anything with a score of 15 or more directly to trash, and sends all other suspected spam to a separate spam folder.

If you get mail that was marked as spam but isn't, the best thing to do is to place the sender of the message on a "white list" of known-good addresses. This is simply a list of e-mail addresses we maintain and the spam tagging program checks. If a sender's e-mail address appears on this list, any message from that sender is considered non-spam regardless of any other tests. The most typical situation requiring white-listing an address is newsletters or mailing lists that you may have actually subscribed to but which appear to be "spammy" in some ways (for example they may use a lot of large fonts and have unsubscribe links). We will soon have a system in place for users to maintain their own white lists, but for now please contact your domain administrator or email us at support@WorldDesign.com for any whitelist changes. It is useful to include the full headers of an email message you want the whitelist entry based on.

Please contact the administrator for your domain or us directly at support@WorldDesign.com.