SpamAssassin was one of the first spam filters. An open-source antispam technology, it's still used as the basis for many anti-spam programs, despite its age and limitations.
As with many antispam programs, Spam Assassin relies heavily on "heuristics" to guess the probability that an email is spam by examining one or more spam-like characteristics. These are typically things like the words "Viagra", "Free Offer", large letters, periods between letters (to disguise the word), etc.
Each characteristic is assigned a "weight"; and if the sum of the weights exceeds some threshold, then the email is flagged as spam. The problem is that spammers quickly and easily defeat such systems by disguising words differently (for example, "V!agra"), which the system must learn, over time, is another spam characteristic.
One attempted improvement has been the addition of "Bayesian" algorithms to help automate this learning process. Based on user feedback of missed spams (false negatives) and legitimate emails that were blocked (false positives), the system changes its weightings in the hope of becoming more accurate in the future.
The result of all this fine-tuning is difficult to predict. Often, it starts blocking legitimate emails for no apparent reason or misses obvious spam.
By itself, SpamAssassin is no longer effective at stopping spam. Typical out-of-the-box performance might block less than 89% of spam and more than 14% of legitimate email. That's unacceptable, especially today, when spam volumes are so high and email has become such a critical lifeline for business.
With a few months of feeding it known spam and legitimate email, SpamAssassin can improve its performance. But what business can afford to wade through months of spam, or never see 10-15% of its email? The cost, especially for some like doctors, hospitals, attorneys and law firms, could be staggering.
Despite those issues, SpamAssassin is still used as a basis for many antispam products today. As always, there is no free lunch.
SpamStopsHere™ does not use SpamAssassin. In fact, SpamStopsHere works much differently and is more accurate out-of-the-box.
SpamStopsHere successfully blocks about 20,000 emails monthly for us, with no user intervention and very accurately.
SpamStopsHere uses a completely different approach. We flag spam based on many complex criteria, the most important being the URL (web address) of the "click me" link that appears in almost all spam emails.
Spammers can (and do) easily send identical spam from new "zombie" computers. That's why "blacklists" are not very effective. But spammers cannot easily add new destination URLs (the "click-me" link) in those messages. Doing so would be difficult and expensive. They would have to register and pay for the domains. So, a much better anti-spam technique is to block messages that contain known spam click-me links. That's called "URL" filtering.
Blocking spam based on such URL filtering has been considered far superior in theory, but also technically impractical. SpamStopsHere solved that problem with proprietary filtering technology, a 24/7/365 live review team. and "honeypots" that receive millions of spam messages each day.
When you use SpamStopsHere, you'll benefit from one of the most complete and accurate antispam knowledge bases in the world. Your spam should nearly disappear and you should start receiving almost all of your emails. Many of our customers stop thinking about spam altogether.
SpamStopsHere works wonderfully for us. Over 98% of email sent to us is spam (or viruses). You stop almost all of it. We used to check the 'hits' to look for any real mail that you called spam by mistake (false positives). We never found any, and have stopped checking. Bottom line - Your service is amazing.
Note: The SpamAssassin Logo is licensed under the Apache License, Version 2.0 (the "License"). You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0