E-Mail Address Forensics

When is an active email address no longer worth the time and resources it takes to send messages to it? I have seven active email addresses (not including my SMS account). Some I’ve had since 1996 and they receive more than 50 messages (mostly spam) per day. It’s a daunting task to keep up with (without getting into why I have so many accounts). For a marketer, it’s a gamble whether or not they’ll connect with me when I check my accounts.

Rather than discuss the merits of email change of address services (of which there are many), I’ll describe the factors that should determine migration of an email address from an active state into one of suppression.

Activity Monitoring

You may have considered goals for email marketing performance, but have you considered setting thresholds for lack of activity? You might consider setting an open rate threshold for 90 days of activity. To do this, divide the number of opens (which do not have to be unique opens) by the number of emails sent. If a user opens seven out of 10 emails, they’re at a 70 percent open-rate average. You might then consider setting the benchmark at 50 percent, the threshold at 10 percent. Perform this exercise every 90 days and segment users who fall below the benchmark threshold. Consider sending these users enhanced messaging that may receive a better response. It could re-establish their status with the rest of your data.

The next area to monitor is unsubscribes and complaints. It should be clear any time you receive a larger than average number of unsubscribes there’s a problem with content, frequency, or other messaging issues. What you may not have considered is many recipients simply refuse to click an unsubscribe button. These are the ones you should be concerned about. Complaints and apathy aren’t far removed from unsubscribes. Your objective should be to try to identify these recipients and suppress them from your list. When a campaign generates an inordinate percentage of unsubscribes, take a close look at non-opens. Consider segmenting those addresses for enhanced messaging, or tweak their unsubscribe options to further evaluate their activity status with subsequent messaging.

AOL’s new 9.0 release adds Bayesian filtering to its anti-spam arsenal. The filtering technology is a probability generator based on user activity. It monitors and eliminates messages the user may not be receptive to. It does this by tracking message attributes, then automatically filtering future messages based on those attributes. Sender name, email address and words in the message are all factors (in AOL’s case, this applies to senders not in the recipient’s address book).

Bayesian filtering can almost be considered a benign form of blacklisting. A sender may have little information other than lack of open rate activity to determine the recipient’s email address is no longer viable. As a result, you may want to consider monitoring AOL recipients in the months following AOL 9.0’s release.

Acquisition History

Many state laws (and pending legislation) place a time limit on the definition of “prior business relationship.” It may be time to ascertain if you have a date and time for the point of acquisition of each address on your lists. If not, begin adding this information to the collection process. If addresses lack this information, consider storing a date and time for any point at which the address was responsive either by opening, clicking, or passing along a particular message. When prioritizing these metrics, note that clicks may be most important. They have a stronger correlation with a consensual email relationship.

The Longer Sales Cycle Message

What should you do with an email address you haven’t mailed to in a long time? In many cases, a sales cycle or product life cycle is longer than the life of an average email address. Car buying is a perfect example. People do an incredible amount of online research prior to purchase, then don’t require that information again for at least two or three years. If your product has a long sales or life cycle, consider sending intermittent messages simply to say hello. Offer relevant content, or re-confirm the permission relationship prior to re-engaging the recipient with commercial email. This keeps the relationship active.

Multichannel Options

Most email addresses have a correlating postal address. Consider an offline promotion with a heads up about special email offers to encourage email opt-in, or simply in tandem with an email offer.

If you plan on sending email regardless of the length of time between messaging, consider sending small batches to test the validity of the addresses, and segment free email providers and AOL accounts from other domains. These addresses have a higher likelihood of churn.

Post-Suppression Activity

After adopting some of the recommendations above, you may decide to suppress a number of email addresses. The question then: are these addresses really dead, like bounces or unsubscribes? There may not be a clear answer, but you may want to segment these addresses from those in traditional bounce and suppression files. These addresses may show signs of life in offline marketing efforts. Who knows? Maybe you’ll suppress my address due to inactivity. Then, some day, I’ll respond to your great offer that’s sent to another of my seven email accounts — through a third party subscription.

When do you think an email address is dead? Send me your thoughts!