Dealing with SPAM

If you are hosting your e-mail or web-site with NeatClubs.COM, here are the top things you can do to avoid receiving SPAM.

a) Make sure you do not use "info", "sales" or other easy to guess mail names that are forwarded to your e-mail account. We will help you change this if you need help. We had setup several customers this way initially, but In the recent year we've seen such common e-mail names become targets for SPAM.

b) Consider using "contact us forms" on the web-site with "captchas" enabled. (Captchas can be turned on under System Setup / Registration Preferences). Using a form for inquiries protects your e-mail address from web-site visitors but still allows you to receive legitimate customer inquiries.

c) If your e-mail is hosted with NeatClubs.COM, make sure the SPAM filter is enabled (we can help if you're not sure how to do this)

d) Consider using a recent version of Microsoft Outlook and accessing your mail via POP or IMAP as opposed to our webmail service - Recent versions of Outlook have more sophisticated SPAM filters

e) If we are simply forwarding e-mail to your ISP, make sure that SPAM filtering facilities with your ISP are enabled.

There are books written about dealing with SPAM, so we'll provide a an abbreviated explanation of the topic here.

SPAM is the bane of most companies in our business.  Approximately 70% of the compute cycles on our servers are dedicated to dealing with SPAM, and the remaining 30% to doing everything else.  This is sadly true of most internet providers.  Said simply, SPAM is a "huge" issue.  A web-site visit may take 30 milliseconds of processing time on our web-server. Filtering an incoming e-mail, whether it is SPAM or not, will take approximately 2 seconds of real compute time per e-mail - Approximately 60 times more than computer system resources than the web-site visit.

The good news is, there are excellent tools and techniques to reduce the amount of SPAM that you receive.

Most internet providers (including us) provide SPAM filtering capabilities.  We can't help you with SPAM coming through your internet account because those spam filters are managed by your ISP.  Almost all ISPs will provide you with a mechanism to login and control your SPAM filter and policies.

Approaches to control SPAM are described below:

1. Use an e-mail client capable of filtering SPAM.  In our experience recent versions of Microsoft Outlook are effective at this.

2. Filtering SPAM on the server.  This is where we come in, so we'll spend a little more time explaining this.  Again, this only applies to customers who are hosting their e-mail accounts on our servers as opposed to with their ISP.  We cannot filter SPAM on mail-forward accounts.

We license a commercial version of SPAMAssassin software on our servers that is easily configurable through PLESK (our server management interface).  If you have only one or two mail accounts that are unchanging, we will set this up for you.  If you are managing multiple accounts however and experiencing difficulties with SPAM, you will need to know how to do this through the PLESK mail interface.

A particular challenge with SPAM is that it evolves.  Different types of organizations tend to be targeted with different types of SPAM, and depending on the business you're in, your tolerance of different SPAM related keywords will vary.

The SPAMAssassin software has a default set of behaviours, but for SPAM filtering to really be effective it needs to "learn" with time what you consider SPAM and not SPAM.  Bayesian spam filtering is incorporated into many modern SPAM filters including Microsoft Outlook as well as the SPAMAssassin software that we operate. You can learn more about the underlying technology here:

http://en.wikipedia.org/wiki/Bayesian_spam_filtering

For practical purposes, when managing your e-mail accounts using our PLESK interface, you can take the following steps to control SPAM.

When you select the SPAM icon in the mail setup for any account with a Mailbox (this means it is a natively hosted account in PLESK terminology), you will see the screen below.

SPAMAssassin uses a scoring system to decide what is SPAM and what is not SPAM.  The default threshold is "7".

E-mail messages with a score higher than this are almost certain to be SPAM.  Mail below 7 is more likely to be legitimate.

One approach to SPAM filtering is to append some text to the subject of an e-mail as shown above.  In the example above, an e-mail with the subject "Buy Viagra on-line" would be translated to something like "*NWX:SPAM 23* Buy Viagra on-line".

Note that _SCORE_ is replaced by SPAMAssassin's calculated SPAM score for any received e-mail.  This makes it easy to create a mail filtering rule in your mail client into a suspected SPAM folder. (in this case anything with *NWX:SPAM in the subject automatically gets shunted to a SPAM or Junk folder.

The other approach is to simply delete SPAM e-mail as it comes in on the server before if ever gets to your e-mail client.  Before deleting SPAM e-mail, you should be confident that you are properly classifying SPAM.  We recommend running the SPAM filter for a week or so before select the checkbox to delete mail on the server.

The training tab will show you the subject of your received e-mails and let you indicate whether received messages are SPAM or not SPAM.  This will greatly increase the accuracy of your SPAM filter by allowing the BAYES filter to learn the vocabulary inherent in your typical e-mail communication with your customers and colleagues, and it can temper the standard SPAM filtering rules with this additional knowledge.