Problems with Spam Filters

Content-Based Filters

These systems can work on either the mail server or on your local computer. They work by scanning incoming and/or outgoing e-mail for words, phrases, and other characteristics often found in spam messages. Every time one of these is found, the message's spam score increases. If the spam score is too high, the message goes straight to the junk mail folder.

The trouble with these filters is that these words, phrases, and characteristics are also found in legitimate messages. The spammers learn what words can trigger a spam filter and reword their spams to get through the filters. Many of these filters allow you to set the sensitivity. Setting it so you get most of your important messages also allows most spam through. Setting it so it removes most spam also removes many important messages.

Most of these filters delete suspected spam without any notice to the sender. Thus the sender has no way of knowing that you didn't get the message unless you complain to them that you didn't get a message you expected. Then the filter trashes their second reply.

These filters can help keep spammers in business. The best way to fight spam is to stop it at the source. Reputable internet service providers (ISPs) will terminate the account of a spammer if they receive complaints. For information on how to make such a complaint, see the Tracing Spam page at uxn.com. However, if your ISP uses a content-based spam filter on outgoing mail, you may not be able to do this. The filter likely will flag your spam complaint as spam and accuse YOU of sending spam.

Our Experience with these Filters

• Confirmations that an order has shipped.
• Replies to inquiries about an order.
• Replies to questions about the availibility of a product.

1. We receive an inquiry from a customer about an order.
2. We research the issue and promptly reply.
3. A few days later we receive another message from the customer complaining that we never replied to their inquiry.
4. We look up the previous reply and forward it to the customer a second time.
5. A couple of days later we get an irate message from the customer complaining that we are ignoring their inquiries.
6. We attempt to contact the customer my phone or snail mail. Isn't it wonderful how the internet has improved communication?

Challange-Based Filters

These filters work as follows:

1. The mail server intercepts all mail from e-mail addresses that have not been approved to be received by the recipient.
2. The server sends another message to the sender with a link the sender must click on.
3. Depending on the filter, sometimes clicking on the link will send a message to the recipient informing them that a message is waiting for them and they must add the senders e-mail address to their approved list. Sometimes the sender must enter a code shown on a graphic on the screen.

The trouble with these filters is that they stop all automated messages, not just spam. Many companies send out automated e-mails confirming that your order has been placed, notification that your order has shipped, notification that your order is backordered, etc. These filters can eliminate all these messages. They also change sending an e-mail from a one step process to a several step process. Many people (including us) refuse to jump through these extra hoops, so you may not receive replies to messages you sent to busy people.

Local Blacklists

Most e-mail programs allow you to block messages from e-mail addresses you specify. If you receive repeated spams (or any other unwanted messages) from the same e-mail address this is a good way to eliminate those messages without blocking messages from other people. This does little to eliminate the spam problem because many spammers use phony or other people's "From" addresses and they change them frequently.

Server Blacklists

These filters operate at the server level and bounce back to the sender all e-mail originating from Internet Protocol (IP) addresses in the blacklist. Because these filters work on the numerical IP addresses, they are not fooled by phony "From" addresses, and they do not block legitimate messages just because they happen to use some words spammers often use.

Depending on how the blacklist is compiled and maintained, this can be one of the best or worst anti-spam measures. In the past there have been horror stories of innocent people losing the ability to send e-mail because of a single uninvestigated mistaken spam complaint. If the list is compiled from carefully investigated complaints, this can be a good method for eliminating the majority of spam with minimal impact on innocent senders.

Further Resources