When a spam filtering system is suspicious of a message, it generally does one of three things:
While each ISP or anti-spam filter vendor has their own philosophy of where to set those thresholds, the general philosophy is always the same. If the source IP address or other connection characteristics are extremely suspicious, the message is rejected outright during the SMTP conversation. Once the message is accepted, if there’s a known phishing URL, virus attachment, or other particularly dangerous item, the message will be deleted. Similarly, a message may be deleted if it has a lot of spam-like characteristics, or if that particular recipient has created a rule sending it to the trash.
But then there are the messages that have some spam-like characteristics — enough to be suspicious, but not enough to be certain. Those go in the spam folder, which permits the recipient to check the message in a safer environment (usually with images, links, and other HTML layout disabled) and decide for themselves if it’s spam or not. If it is, the recipient can leave it there or delete it. If it’s not, there’ll be a button labeled “not spam” button or something similar. Based on conversations I’ve heard around the industry, it seems that this “not spam” button is generally assumed to have no purpose — but that couldn’t be further from the truth. When a recipient clicks “not spam,” it improves the sender’s reputation! When a lot of recipients click “not spam,” similar messages won’t be marked as spam! Cool, huh?
There are, of course, some caveats. It’s true that many users never even look in their spam folder, so they never have the chance to vote. Reputation systems will take that into account, too, and weight things accordingly. They may even weight some users’ clicks — spam or not spam — as more relevant than others, often due to the interface engagement metrics that have gotten the industry so excited recently. You won’t see any indication of this weighting in your feedback loop, but that doesn’t mean it’s not happening.
The obvious way to mess with the system would be to send your own account at that ISP a lot of mail, and click “not spam” for all of it. Some spammers have been doing this for years, and there are rumors that some legitimate senders have been caught doing it too. That kind of gaming might work for a little while, but not for long — as a friend who works for a big ISP has been known to say, “do they think we’re stupid? Of course we know it’s them.” When the system catches you doing that, it can simply ignore “not spam” votes for your mail; in other words, your reputation will be allowed to get worse, but it will never improve. Definitely not worth the risk to your reputation or deliverability to try this tactic.
But never mind that; let’s focus on the positive. As a reader of this blog, you already know what you need to do to delight your recipients: be respectful, be relevant, et cetera. And if you are a Return Path customer, you have the tools at your fingertips to ensure that your mail has a good reputation. But even so, filtering techniques and thresholds have to change constantly in reaction to the latest spammer tricks. When there are elements in your mail that a filter deems suspicious, and your mail ends up in the spam folder, it doesn’t have to be the end of the world. If you’ve been doing your job right, and your recipients want your mail, they’ll wonder what happened. They’ll go look in the spam folder to see if it’s there. They’ll click “not spam.” And without ever knowing it, you will have won.