Many companies perform spam filtering because they have to, and not because they want to. The sheer volume of UCE (or unsolicited commercial e-mail, the "official" name for spam) out there can make an e-mail account nearly useless without some form of protection. Typically this is done at the e-mail server, but it can also be done at the desktop, by filtering directly within Outlook.
Why perform spam filtering in Outlook as opposed to at the mail gateway? Filtering at the gateway may simply not be possible. If your company is using nothing more than simple POP3 mailboxes and can't afford to host its own e-mail server, then filtering may only be possible through Outlook.
You may also run into situations where users are inclined to customize their own spam selection filters rather than use the spam filters provided by an outside authority. Advanced users may need to craft their own filters to be specific to their mailbox since one person's spam may be another person's serious mail.
The most recent version of Outlook came with a built-in junk mail filter, and just last week Microsoft released an updated version, which offers some slight improvements. The filter allows for both white- and black-listing (i.e., senders designated as being always safe or always spam), and the updated version has somewhat more accurate filtering and revised rules. You can download it here.
While the built-in spam filter in Outlook is useful, there are numerous third-party products on the market that can help you strengthen your antispam effort. One product you should consider is a filter known as SpamBayes.
SpamBayes, which is both free and open-source software, integrates directly into Outlook as a plug-in and uses a threshold mechanism to do spam screening. Any e-mail with a more than 90% chance of being spam is sent into a "Spam" folder; e-mail from 15%-90% likely is placed in a "Possible Spam" folder for screening; anything scored as less than 15% is considered legitimate.
Marking an e-mail as spam is as easy as selecting it and clicking a button in Outlook's toolbar. The program can also be trained to recognize good and bad e-mail from mail in existing folders, which reduces the amount of time needed to train the program and makes it effective immediately.
One minor flaw: the program cannot read mail from the "Sent items" folder (as a way to train the filter), but it is possible to copy messages from the "Sent items" folder into a new folder, then train the filter from there. (This would not need to be done more than once.)
Outlook 2003 users can get the best results with SpamBayes by using the "Unread Mail" folder to examine everything that comes in, and setting SpamBayes to mark spam as already read so it isn't present in that view. This speeds up the screening process even further since the user doesn't have to go through two different folders to mark messages.
SpamBayes can also be used in conjunction with existing Outlook message rules. To keep SpamBayes from interfering with rules (for instance, by marking as spam a message that should be processed by rules first), one of the settable options in SpamBayes is the period of time before messages are processed, and a time delay between each message. A start delay of 1 second and a processing delay of about 0.2 seconds should keep rules from being overridden. SpamBayes adds a custom property (called "Spam," appropriately enough) that indicates the spam score for a message, from 0-100%. This can be added as a user-defined field for the inbox or the spam folder (although not the Unread Mail view, unfortunately).
SpamBayes can also be used as a proxy service rather than an Outlook plug-in, with its classification functions accessed through a web browser. This would be a good way to quickly create a spam-screening system for a whole organization, although the multi-user granularity of having SpamBayes run in each user's copy of Outlook would be lost. There's no way to use the Outlook plug-in to talk to SpamBayes running as a proxy service yet. Also, someone would have to assume the responsibility of classifying e-mail as spam or good on everyone's behalf.
SpamBayes is by no means the only free spam-fighting product, but it is probably the most advanced and well-thought-out.
Another powerful system for screening e-mail on the desktop that's free and open-source is POPFile. POPFile was designed more as a generic e-mail classifier, although it contains many intelligent strategies for spam-fighting.
That said, POPFile has several disadvantages when compared to SpamBayes. For one, its default interface is through a web proxy, which makes it useful in a multi-user environment, but only up to a point. There is an Outlook plug-in, called Outclass, but it was written by a third party and is not supported by the authors. The plug-in also does not work in a multi-user fashion; it simply uses the POPFile classification engine inside an Outlook plug-in wrapper.
Also, POPFile does not have a "Possible Spam" classification, so the user will often have to look into the spam folder and manually determine which messages are not spam. POPFile also doesn't have the ability to train from existing e-mail as easily as SpamBayes.
In the same vein as POPFile is PASP, a mail classification filter and proxy written in the Python language. InBoxer is a commercial antispam plug-in for Outlook. Touted as highly accurate and very easy to work with, there's a free 21-day trial version, and the full version is $27.95. Spam Bully, yet another commercial product, works with both Outlook and Outlook Express and includes such features as reverse IP tracing to determine an e-mail's country of origin. It can be used free for 14 days (the full version is $29.95).
Serdar Yegulalp is the editor of the Windows 2000 Power Users Newsletter.
Do you have a useful Exchange tip to share? Submit it to our monthly tip contest and you could win a prize and a spot in our Hall of Fame.