PerfectMail is 99.9+% Accurate!

Our Approach is Different, so our Results are Different
Most antispam appliances either use spam signatures or Bayesian filters to score a a message. These techniques when combined can achieve a maximum of about 95% to 98% effectiveness, but don't offer consistently accurate filtering.

PerfectMail combines traditional techniques with a data warehouse based reputation system. PerfectMail's data warehouse records:

• Details of every message received including sender, recipient, score, etc.
• Scores for every message received from every connecting mail server
• RBL status of every connecting mail server
• Much, much more.

Think of PerfectMail's reputation system like a credit rating system... for e-mail. PerfectMail strongly favors e-mail from established peers in a 2-way e-mail relationship with a protected users. PerfectMail also favors messages from persistent 1-way senders (such as e-mail newsletters). If the sender has no prior history, then the messages' disposition is determined by the reviewing and verifying the content of the e-mail.

PerfectMail's reputation system is unique and it is key to our achieving 99.9% accuracy.

The PerfectMail Approach

"The simple truth is that there is no single, definitive test for Spam."

PerfectMail uses a combination of techniques to quickly and accurately scan incoming and outgoing mail for spam and viruses.

• Traditional Antispam Tests
  PerfectMail uses over 100 traditional tests when examining an e-mail. These tests include verification of all e-mail header and envelope fields. PerfectMail also does a complete scan of the e-mail body and all attachments looking for obvious signs of spam.

  PerfectMail's antispam tests are very efficiently implemented - most complete in 1/1,000 ^thof a second or less.

  PerfectMail performs Baysian scans of e-mails whenever the spam disposition of the e-mail is in doubt. This extra scanning helps ensure that PerfectMail only tags or rejects e-mails that are truly unwanted. Training for the Baysian filter is automatic.

• Next-Generation Antispam Tests
  PerfectMail includes a number Next Generation antispam tests that are unique to the product. These tests go well beyond traditional antispam techniques and have proven themselves to be exceptionally good predictors of spam.

• Historical Tests
  PerfectMail records all e-mail activity into a number of very efficient databases. These databases contain information on past connections, sender activity, malicious e-mail activity and much, much more.

  PerfectMail uses these databases to help determine the E-Mail Worthiness of a particular e-mail. Just like a merchant will consult past-history when making credit decisions, PerfectMail consults past activity by the sender before determining the final disposition of an e-mail.

• Black/White Lists
  PerfectMail implements Black (always unwanted) and White (always wanted) lists. Black/White lists let administrators override PerfectMail's default behavior so that e-mail from a specific source (domain, network or user) will always be handled as desired.

  Traditionally, black/white list handling has been a real bother - because administrators and users spend a lot of time maintianing the list. With PerfectMail, Black and White lists are available but probably won't be needed. If PerfectMail ever makes a mistake, you can always override is decision by updating the Black/White list.

And This E-Mail's Status Is...

Before tagging or rejecting an e-mail, Perfectmail consults historical databases to examine the past activity of the sender. If the sender has an established history of sending legitimate e-mails, PerfectMail will let the current (questionable) e-mail through. This is done to help ensure that no legitimate e-mails from known e-mail peers are ever rejected.