Spam filters are software that is designed to block out unsolicited bulk emails. These filters are usually part of an email client and analyze mails as they come into a user's mailbox. The techniques used by these filters are many in number and start from the simplest technique of information sharing to the more complex use of mathematical functions to calculate the possibility of a mail containing spam.



Some of the more simple filters will simply scan one's address book and allow users from trusted domains and known contacts to contact you. This is a very basic format and was used in the first generation of filters. Technology moved up a notch after that to include more complex mechanisms like DNS blacklisting. In this method, a possible spammer is looked up as a DNS entry and the domain name of origin of the mail is looked up, if there is anything fishy about this entry or if the entity is a known spammer, then the mail is blocked out. There are some more collaborative means of blocking out spam as well. Some of these filters will use a collective intelligence of many spam recipients. In this method, when a mail is marked as spam, the information is then passed on to a central server and the information shared with other users of the spam filter software.

However, all these are not very efficient methods and to be honest, are quite rudimentary in how they work. The best spam filters with a nearly 100 per cent success rate are those that use complex mathematics to calculate probability and actually learn from the user to identify spam mail. There are two back end logics that are used in these filters. Bayesian filtering is a method use by some filters. This method works by identifying the occurrence of some keywords in every mail that is marked as spam by the email user. Over a period of time, the software then builds up a statistical probability of a mail being a spam mail. Eventually, the software can work completely independently. The other method that is used is called Markovia discrimination. This uses the same premise as the previous method but is more powerful because it checks for the occurrences of whole phrases. This is important because some slick writing to mask the occurrence of a keyword could slip through in the Bayesian method.

Understanding how spam filters work is the first step to retuning the content of your bulk email material. The fundamental changes that you need to make to your mailer material is to ensure that you never unnecessarily repeat information be it words or phrases. The other requirement is that you ensure that you acquire a customer's opt-in. This will make nearly any mail that comes from you immune to filtering in a spam checking mechanism. This is because most filtering mechanism do a pass of your sent folder and any material sent from you folder is kosher.