The Legend of Spamalot and the Nobility of Words
Spamalot? You might if you’re using the wrong words in your email marketing.
Years ago, the role of the email inbox gatekeeper was one that pivoted on the propensity of email to actually be spam and do away with that message before either the ISP had to process it or the recipient fell prey to the cloaked message. As time passed, receivers became wiser to the game spammers played and created more robust checks including whether messages had been marked as spam, network lists containing known origin information about bad senders (blacklists) and measures of email engagement as of most recently. But, one of the bulwarks that has made it through to this day is the checking of spammy words when email content is analyzed. It is still a critical piece of email deliverability that everyone should be aware of but can be easily dismissed for more sophisticated checks. If I had a dime for every time a legitimate sender had mistakenly created a cacophony of spammy sounding words simply by selecting those which were not alarming by themselves but when added to others made the message appear as spam, I wouldn’t have to order off the kid’s menu anymore. This usually is followed up by a discussion on how to avoid being dinged next time around. You won’t find many major email receiving networks weighting word choice heavily (although it’s still built into the pre-delivery check at most), but as you start to send into the long tail of smaller domains, you can find yourself really being tripped up as spam checkers have become virtually free and easy to use, and most importantly, are still effective in ferreting out bad mail. So, how do you avoid your messages becoming false positives that get relegated to the spam folder?
Hopefully the following tips will help you find a way with your email marketing to do just that.
Most ESPs these days have a built in content check you should take advantage of. I’ve seen some that are seamlessly built in to campaign creation workflow whereas others require a few extra button pushes before seeing results. But, however your ESP has these tools implemented, you should take advantage of them. It’s what you pay them for and a clean creative is imperative to good deliverability. Usually you can get some assistance in interpresting the results from the experts on staff as well.
Some ESPs outsource message checking to 3rd party deliverability services altogether. Or, if your ESP isn’t integrated with one of these vendors, you can sign up for your own account to get a better read on the type of content you’re sending out. Be proactive, though. Willful ignorance will most likely cost you. Companies like Return Path and Pivotal Veracity have tools which will review your content and tell you if there are any red flags. Of course, there are a slew of other smaller companies which offer this type of service at a discount as well. One of the most advantageous aspects of doing a check in an automated fashion with one of these services is the report you’ll get which then can be distributed around to the different teams responsible for the email creation process so the results are easily seen and shared.
By far, this is the most popular (and free) application available to email receivers that scans and assigns a numerical score to each word found that’s on the “bad” list. There are a ton of other rules in place that Spam Assassin checks for as well, but word usage plays a prominent role in labeling your email as spam. The more words you have that are on Spam Assassin’s list, the higher your score. Depending on the threshold set by mailbox administrators, your mail will be quarantined, bulked or outright blocked. This is also the engine that most content checking tools use for their basis of determination of the nature of the creative – will an ISP see this as spam? For a complete list of rules that Spam Assassin checks for (including those around word usage), check the latest listings at the Spam Assassin website here.
Sometimes you’ll come across ISPs or email rendering applications that will post rules around what they define as spam. For example, Road Runner uses a special bounce code for messages coming in that triggers their anti-spam filters based on content. If you send enough email out that has spam words in it (i.e., a large campaign going to enough volume at a single ISP), you run the risk of all your mail leaving that IP address being blocked. Microsoft has also provided a healthy list of words and patterns to avoid in the content so recipients behind Outlook are protected. Here’s a list of words extrapolated from spam resources as well.
Free self service tools
If you want something lightweight you can use on your own without being plugged into an ESP or a deliverability vendor, you can download MailingCheck and install to your desktop. From my experience using it, you get a decent overview of what, if any, content triggers exist. It will provide the actual listings but you’ll have to do the leg work and research what exactly the impact of a certain word or combination of words is.
When you find words designated as spam in your content, you should first decide if the language is enough to cause harm to your email marketing deliverability. Just because I use the word “free” in an email marketing promotion won’t get me blocked, but it will most likely show up in some content checkers. However, using the word 30 times and with accompanying words like “hot” and “young” (think of a back to school campaign) anywhere in the content will most likely shunt the mail permanently to the spam folder. If you must use a word, use it sparingly. Also, try using words that describe the same intent but do not show up on the spam hit list. Test repeatedly to find the right mix of verbiage to convey your message without offending the spam checkers, which are computer driven and don’t get the subtleties of written language as easily as we do (read Bayesian filtering if you’re up for the more technical details).
Takeaway: Spam checkers use content analysis of words to determine whether a message is legitimate or not. While the content may look benign to the human eye, a machine scanning your message may disagree and inadvertantly mark your message as spam due to the perceived toxic combination of words used.