Let's talk all about Unsubscribes!


Last time, I wrote about list-unsubscribe methodology specifically, but today I’d like to broaden the topic to all the different methods that senders use to offer an opt-out option for their subscribers. That could include “one click” unsubscribe links, or leading users to unsubscribe screens or profile center pages, or asking a subscriber to reply to a message using a special tagged address, or sometimes people will say “just reply to abc@xyz.example with UNSUB in the subject line.”

Some are better than others. Here’s some pros and cons around different types of unsubscribe methodology, as filtered through my twenty years of watching people send marketing messages (and seeing both a lot of good senders and bad senders along the way).

Unsubscribe via email: Reply-to-unsubscribe with a tagged address.

Maybe your sending platform puts a tagged address in the reply-to header and tells you that if you want to unsubscribe, you should send mail to “repxyz-123-456@reply.spamresource.com” and say “UNSUBSCRIBE” and you’ll be removed. That’s not horrible – the “tagged” (unique) address usually corresponds to your specific email address and even if the mail was forwarded, sending a message to that address should result in the correct subscriber being unsubscribed. But one risk here is that you need to be aware that spambots could get hold of that email address and send spam to it. Does your platform retire those tagged addresses after a period of time? Could a spambot trigger an unsubscribe that wasn’t actually requested by a real subscriber? That can get a bit tricky and is something to talk to your sending platform about.

Unsubscribe via email: Reply-to-unsubscribe without any address tagging or special functionality.

You’ll often see something like this in the email footer: “Just email joe@spamresource.com and put UNSUB in the subject line.” This is problematic. If a user has email forwarded to them, from an alumni account or an old Gmail account, they possibly can receive mail at a certain address, but not send mail from that certain address. Meaning you’re going to end up unsubscribing the wrong person. Then there’s address plussing and aliasing (domain aliasing and Gmail’s dot aliasing), which are even more ways for subscribers to receive mail at more addresses than they can send from. For some of these, you can indeed configure your email account or system to send from the alias address. But not all – or it can be complex. In short, if you email bob2@spamresource.com or bob+mailer@spamresource.com but he sends in an unsubscribe request from bob@spamresource.com, how do you know which Bob address to unsubscribe? If you unsubscribe the wrong one, the emails won’t stop, and you’ll have an unhappy camper who will be likely to start complaining that you’re spamming them.

One click unsubscribe: Just click here to be unsubscribed.

A single click to a web interface that completes the unsubscribe process has long been the gold standard for marketing email messages. But, nowadays, with the rise of security applications and services scanning inbound email messages to look for malware, this type of functionality falls prey to “ghost clicks” from bots and security scans. You don’t want a company’s Microsoft or Barracuda email security service to unsubscribe your subscribers unexpectedly. To me that means that every “one click” unsubscribe process should become a “two click” unsubscribe process. Show the landing page, ask the user to click. Lots of platforms do this today and I’ve seen nothing that suggests it violates any of the common email marketing laws, and it safely tracks unsubs while preventing “false positive” unsubscribe damage due to ghost clicks. Those security bots aren’t going to follow links (or submit form actions) on your landing page.

Using a profile center or unsubscribe center: Click here to update your profile.

As smarter sending platforms get smart about bot clicks and ghost clicks (see above), the line between a “two-click” unsubscribe landing page and a profile center landing page begins to blur. Since a profile center’s landing page generally isn’t going to auto-initiate an unsub action upon loading, you’re relatively protected against “false positive” unsubscribe damage. And it makes good marketing sense to allow subscribers to modify their preferences, to be able to do things like “opt-down” (request to receive mail less often but stay subscribed) if they like.

Keep in mind that ISPs will consider it a bad practice (and you are likely to run afoul of various laws) if you “protect” an unsubscribe process behind a login. Though I’m not a lawyer, my plain reading of CAN-SPAM suggests that it doesn’t allow for this. My friend Mickey Chandler (who is also not a lawyer, but is actually in law school, good for him!) would point you at 16 CFR 316.5 for the specific language in US law that effectively prohibits this.

Clicking “Unsubscribe” in the webmail’s user interface.

If Gmail or Outlook.com show an “unsubscribe” button in their interface, it’s likely that the sending platform used to send that message included a hidden list-unsubscribe header in that email message. This is generally a good thing, but it doesn’t excuse the sender from offering their own unsubscribe mechanism in the email body – that’s still required.

The few ISPs that do offer this functionality, though, are doing so as a favor to both senders and recipients. For recipients, it gives them an easy way to quickly find the unsub button to opt-out, without having to scroll through a long email message that they don’t want to read. For senders, it makes it easier for recipients to choose to opt-out in a way that makes them less likely to click the “report spam” button instead. These “list-unsubscribe” opt-outs do less harm to a sender’s reputation than spam complaints.

There’s a couple different types of list-unsubscribe methodology – for more on that, check out my recent post on list-unsubscribe MAILTO versus HTTP versus POST (1 click). TL;DR? Mailto and simple HTTP/HTTPS list-unsub could fall prey to ghost clicks like other methods described above. IMHO, best to implement HTTPS and POST only, and the HTTPS method should be “two click” in nature, requiring that the user confirm their unsub by clicking a button on the landing page.

What about the “report spam” button in most webmail UIs?

Remember that ISPs happily collect spam reports about the mail you send. Some send reports back to you, if you’re registered with them (Outlook.com and Yahoo and many others), and some do not (Gmail and many others). But regardless of whether or not they send reports back via a “Feedback Loop” process, they still use these reports as spam indicators. Too many reports, too high a percentage of reports, and a sender’s messages are likely to get blocked or relegated to the spam folder. This is but one of many reasons that you don’t want to encourage subscribers to “report spam” instead of unsubscribe – make sure you’re not doing so, even unintentionally. Is it easy for your subscribers to unsubscribe? Big, clear, text? Not 6 point white-on-white hidden in a legal footer? The better your hide the unsub link, the more likely subscribers will be to say “to heck with this” and just hit the report spam button instead.

No comments

Comments policy: Al is always right. Kidding, mostly. Be polite, please and thank you.

Powered by Blogger.