Filters, aliases can
make spammers’
job a little harder
Minnesota workers are drowning in a cesspool of e-mail pitches for get-rich-quick schemes, pornography and health products.
Spam, otherwise known as unsolicited commercial e-mail, currently comprises about 50 percent of all e-mail traffic, resulting in an expensive drain on Minnesota businesses’ bottom line. Considering the time needed to deal with it, spam’s cost for a company of 100 employees can easily reach tens of thousands of dollars a month in lost productivity.
While legislators and technology gurus continue to grapple with big picture solutions to control spam, there are steps that employers can take now to reclaim their e-mail boxes.
Check patterns
Are you making a spammer’s job easier? First, take a close look at your employees’ Internet use patterns so they don’t unwittingly place themselves on junk e-mail lists. Employees who place orders online get more spam messages than those who don’t engage in e-commerce. Electronic marketers collect customer e-mail addresses to create mailing lists, just as more traditional marketers do.
Even if you don’t buy merchandise online, the spammer wants you to call a toll-free number or click on the “unsubscribe” link embedded in the message. Once you do, the spammer knows that your e-mail address is valid and will never take you off the list. Worse, the spammer will sell the list with your address on it to other spammers.
Since spamming is so cheap and easy ( for less than $200, your can buy a CD-ROM with millions of working e-mail addresses), it doesn’t take many responses for a spammer to realize profits. In a perfect world, no one would buy anything from junk e-mail messages, and spammers would soon go out of business. With the low costs of sending spam to millions, it only takes a few respondents to make a spammers day. Nevertheless, e-mail users can complain to the administrators of a spammer’s Web site to get that account closed or suspended.
To do so, you’ll need to send your complaint via e-mail to postmaster @(fill in offending domain name here) and to abuse@ (fill in offending domain name here). Legitimate hosting companies do not want to offend their upstream vendors and will quickly take a spam-driven Web site offline.
While these low-tech approaches will keep some spammers at bay, thery’re a bit like teaching someone to swim during a hurricane-you or your employees are barraged with so much junk e-mail that no one has the time to track sown the offenders and complain. Businesses need other tools to protect office computers-and their operators-from drowning in a flood of unsolicited commercial e-mail.
Fake noses
Aliases are the electronic equivalent of a fake nose and mustache. Taking on an alias is a simple way to automatically sort junk e-mail. This strategy costs nothing but a little bit of time, and works best when activating a new e-mail account. As you’re setting up the account, request that the system administrator create an alias address in addition to “real” or public e-mail addresses that employees will use to receive messages.
Use the alias when giving out your e-mail address over the Internet-to register software, peruse a Web site or to place an order. Now, configure your e-mail program to look for messages addressed to the alias and place that pesky spam into the “Deal with these in 2010” folder. When that alias winds up on a commercial address list, messages from the sender will be directed to the alias inbox, keeping workday mailboxes uncluttered for legitimate correspondence.
If spam has become a problem, the best solution might be to change e-mail addresses. How best to handle this situation?
Pick a new address that’s difficult for spammers to guess. For example, rather than bill@visi.com, choose bill-hanson1@visi.com or billh1@visi.com
Set an autoresponder on the old e-mail address to tell senders that mail is no longer accepted there. Give the new address in the text of the autoresponse. Spammers donÕt read returned messages and legitimate senders are automatically informed of the new address.
All the traditional defenses against spam, such as described above, still can be inadequate in keeping mailboxes safe once spammers have employee addresses. Then, there’s only one other thing that can be done to stop it: filtering. There are several types of filtering mechanisms.
Black lists. Software programs that create list of spammers’ addresses and then block them from sending mail to your company’s server. Limited black lists can be maintained by a designated employee or group of employees; large lists are maintained in real time by organizations like mailabuse.org. Once your server is configured to access the database, you can start dumping spam into the “bit bucket”
Matching. The typical junk e-mail has some language that identifies itself as spam, such as “free” or “hot.” Your company’s server can be programmed to watch for these trigger phrases, and to filter or delete spam as it comes in. The danger lies in making filters strict enough to catch spam, but not so strict that you end up with false positives-innocent messages mistakenly identified as spam. For most e-mail users, missing legitimate e-mail is an order of magnitude worse than receiving spam, so a filter that yields false positives is like an acne cure that carries risk of death to the patient.
Recognition weighting. A method of tagging messages as spam, depending on the triggers that match a list of predefined criteria, including specific phrases, keywords and content. The software analyzes e-mail messages, assigns numerical values to the triggers and then totals the values. Quasi-innocent terms such as “lust” or “rich” have a minimal value of one; profanity has a higher value of six, and stylistic components of messages like ALL CAPS SUBJECT LINES have a value of three. If the total exceeds an arbitrary threshold, the message is labeled as spam and is deleted or filtered. This method is best run as a server application, as it requires extensive processing time and needs maintenance to keep the criteria database current. However, weighting has a lower false positive rate than black listing or matching. As an alternative to the server solution, individual e-mail users can run a stand-alone weighting product called Spam Assassin Pro. Even though each user needs to maintain his or her own keyword database, this is a preferred method of spam filtering for many companies.
Statistical filtering. Using the content of e-mail messages to create probability tables that evaluate if any given message is, or is not, spam. Unlike recognition weighting, the statistical method doesn’t need maintenance and updates to the database. The more spam thatÕs sent through the filter, the smarter the filter becomes. Because the filter measures probabilities, this approach considers all the evidence in the e-mail, good and bad. Words that occur rarely in spam, such as “though,” “tonight” or “apparently” contribute as much to decreasing the probability as bad words like “unsubscribe” or “opt-in” do to increasing it. An otherwise innocent e-mail that happens to include the word “sex” is not going to be tagged as spam.
Damming the Flood
Services using statistical filtering like Brightmail and Postini, help e-mail servers keep ahead of spam technologies. The cost can be high (averaging $60,0000 per year), but the hidden costs of spam to companies are higher still.
Finally, if you think that legal regulations are needed to control spam, contact your state and federal legislators. The junk fax law, which has nearly eliminated that once widespread problem, has set a precedent on which to build anti-spam laws.
While none of these spam prevention methods will eliminate 100 percent of the junk e-mail coming to our inbox, they’ll at least dam the flood. For now.
[Contact] Bil MacLeslie is director of operations for the Minneapolis-based Internet service provider VISI.com: 612-395-9000; bil@visi.com
