Best practices: A Gmail Perspective

At M3AAWG 30 in San Francisco, Gmail representatives presented a session about best practices and what they wanted to see from senders.
I came out of the session with a few takeaways.

  • Gmail spends a lot of time and energy on filtering mail and giving the user the absolute best inbox experience possible.
  • Gmail does per-user filtering, probably more than any other ISP out there.
  • Gmail filters are intentionally aggressive.
  • Authentication is important for good delivery at Gmail.

Gmail mentioned a few specific things that were important for getting mail delivered to the inbox. Anyone who has read here will recognize many of these suggestions.

  1. Make sure your mail is really opt-in. Gmail strongly suggests all senders use a confirmed opt-in process whenever possible.
  2. Comply with RFC 2822/5322.
  3. Use well formatted HTML.
  4. Don’t use public URL shorteners.
  5. Maintain your lists and remove non-responders.
  6. Authenticate your mail. Gmail is waiting for adoption to get high enough so they can throw away any unauthenticated email.

During the session, they specifically called out affiliates as “pretty spammy” and said that they see the highest spam rates by users with promotional affiliate mail. The recommended senders who are going to use affiliates monitor every single campaign. But they said most affiliates have horrible practices and use all sorts of obfuscation techniques. They also called out dating and payday loans as two areas that were hurting a number of otherwise reputable brands.
For ESPs Gmail specifically said they hold ESPs accountable for customer actions. (I’ve seen this with a couple clients where the ESP domain is actually filtered for all their clients.) ESPs must make customers follow delivery guidelines and have zero tolerance for abuse.
The do recommend using separate sub domains for different email streams, but never ever cross the streams. If you have a transactional domain, never send promotional email using that domain.
Gmail also expects you to warm up domains as well as IPs. They did say their filters adjust quickly and that you can start with a low amount of traffic and double that traffic every couple hours.
As I mentioned earlier, they did announce their new feedback loop program. They also announce the presence of an “unsubscribe” link in the email interface.
Gmail_unsubLink
Senders can get interface unsubscribes by providing List-Unsubscribe headers in their emails. Gmail prefers the use of mailto: headers, which will generate an email to the address in the header when clicked. For companies who only provide a http: link, Gmail pops up a box that tells the user to visit the site in the link.
Gmail_Unsub_HTTP
Gmail prefers the mailto: header, as it makes for a more seamless user experience.
This is interesting, as a ‘unsubscribe’ link in the interface is something I’ve heard senders asking for over the years. Will this be adopted well enough that other mailbox providers and mail clients will implement it? Only time will tell.

This month in email: February 2014
Horses, not zebras

Related Posts

Barracuda filters clicking all links

Earlier this month I mentioned that a number of people were seeing issues with multiple links in emails being clicked by Barracuda filters. I invited readers to contact me and provide me with any information or evidence they had. Not only did a number of senders contact me, but one of the support reps at Barracuda also contacted me.
At issue is a part of the Barracuda email filter call the intent filter. There are 3 different modules to this filter.

Read More

Don't unsubscribe from spam!!

Having been around the email and anti-spam industry for a while, I’ve just about seen and heard it all. In fact, sometimes I’ve been around for the beginning of the myth.
One myth that seems to never actually go away is “unsubscribing just confirms you’re a real address and your address will get sold and your spam load will explode.” This is related but orthogonal to “spammers harvest addresses out of unsubscribe forms.” The reality is that both of these things used to be true. Unsubscribing would confirm your email address and increase your spam load. Spammers would harvest addresses out of unsubscribe forms.
But neither of these things have really been true for the last decade.
I have had clients over the years that are spammers. Some of the are names that you probably would recognize. Some of them are companies we could probably all agree are spammers.  Some of them are buying addresses from companies that are spammers. Some of them are companies that have a good mailing program here and then hire snowshoers over there. Sometimes they come to me claiming to be real mailers “with minor delivery problems.” Sometimes they come to me saying that a blocklist has recommended they talk to me about repairing their processes. Sometimes they even actually want to fix things. Sometimes they’re just looking to say that I’ve given them a clean bill of health (which is not something I do).
What that means is that I have lots of addresses on lots of spammer lists. Not just the ones they’ve found, but ones I’ve used to test their systems. I use tagged or disposable addresses for everything. Some of my disposable accounts are only marginally connected to me as I want to see what senders really do for their subscribers rather than what they want me to think they do. The ones I add to their system I use to test their subscription process as well as their unsubscription process.
I have never encountered a situation where unsubscribing one of those addresses caused a “multiplication” (to quote one anti-spammer) of my spam load.
I’ve had cases where my clients have ignored unsubscribes. I’ve had cases where my clients have decided years later to add me to their list again. I’ve had cases where they’ve been bought out and my address has been reactivated by the new owners. I’ve had cases where months or even years of 5xx responses was ignored. I’ve seen just about every bad bit of behavior on behalf of spammers. But I’ve never actually had unsubscribing increase my spam load.
It doesn’t matter how often people demonstrate unsubscribing doesn’t result in more spam in the current email ecosystem. (Ken Magill 2013, NYTimes 2011, dayah.com 2009). It doesn’t matter that many mailers treat “this is spam” button hits the same way they handle unsubscribe requests. The myth still persists.
 
 

Read More

Update on Herman Cain advertising male enhancement drugs

Shawn Studer from newsmax.com contacted me today with a statement about the Herman Cain mailing list.

Read More