Anti-spam programs can detect unwanted email and prevent it from reaching users’ inboxes.

These programs use a combination of methods to decide whether an email is likely to be spam. They can:

• Block email that comes from computers on a blocklist. This can be a commercially available list or a local list of computer addresses that have sent spam to your company before. 
• Block email that includes certain web addresses. 
• Check whether email comes from a genuine domain name or web address. Spammers often use fake addresses to try to avoid anti-spam programs. 
• Look for keywords or phrases that occur in spam (e.g. “credit card”, “lose weight”). 
• Look for patterns that suggest the email’s sender is trying to disguise their words (e.g. putting “hardc*re p0rn”). 
• Look for unnecessary HTML code (the code used for writing web pages) used in email, as spammers often use this to try to conceal their messages and confuse anti-spam programs.

The program combines all the information it finds to decide the probability of an email being spam. If the probability is high enough, it can block the email or delete it, depending on the settings you choose.

Anti-spam software needs frequent updating with new “rules” that enable it to recognize the latest techniques used by spammers. Many users worry that anti-spam software will delete personal or useful email. In fact, your email is safe, and you can even see selected spam if you wish. Anti-spam programs can be very accurate. Typically, they may block less than one genuine email in ten thousand, or even a hundred thousand.

Even if the program does incorrectly identify an email as spam, it can be configured to place it in a “quarantine” area, rather than deleting it. An administrator can then decide whether to let the mail be delivered or to delete it. Some programs let each user reclaim any quarantined mail that they want.

Some anti-spam software is “adaptive”: it learns which subjects you find acceptable and which ones you don’t. Suppose that a pharmaceutical company installs anti-spam software. At first, the software tries to spot spam by looking for words like the following: credit, free, consolidate, debt,
mortgage, drugs, prescription, medication, doctor. It blocks email with too many of these keywords, but allows individual users to retrieve mail that they want to read.

Someone in the research department finds that genuine mail about new drugs has been blocked, and asks for it to be released. The software learns that that user frequently receives email about drugs – and so gives less weight to drug-related words when checking for spam. In the finance department, users reclaim email with financial terms in it, so the software learns to give less weight to these words – but still blocks drug-related email for that user.