Poor delivery at Gmail but no where else

I’ve mentioned before that I can often tell what ISP is making filter changes by what my calls are about. The last few weeks it’s been Gmail where folks are struggling to get to the inbox. One of the things most clients and potential clients have mentioned is that they’re not having any problems at the other major ISPs.

meter19
Gmail’s filters are probably the most extensive and complex of all the major webmail providers. Their ability to process data and pull signal out of noise, even when it is all noise, is unmatched in the space. But, still, I wasn’t sure why so many companies were struggling just with Gmail.

I’ve been noodling around with this. Could it be that Gmail is doing something very different than other companies? Are they assigning reputation in ways that are different? Do we need to change our strategies to deal with Gmail? Are there different things we should be telling folks?

Then, I realized the big difference between Gmail and the other webmail providers is their FBL. With the other major webmail providers senders can clean their lists just by removing anyone who hits the “this is spam” button. This lets senders be sloppy with acquiring email addresses without too many consequences. Gmail, however, only tells about complaints, they don’t tell you who is making them. Gmail puts the onus on senders to figure out how to send mail to people who want it.

It also means that Gmail is a more realistic view into what subscribers think about a mail stream. It’s easy to forget that everyone who hits this-is-spam actually gave the sender an email address in the first place. The fact that senders can’t just remove them from the mailing program, means the poor reputation builds up over time. Eventually, the number of complaints is going to go over whatever the appropriate threshold level is, and there’s no easy way to reduce it.

I’m not sure what the solution is. I do know that bad reputations at Gmail take some significant work to repair. There’s no short cut, senders have to get rid of subscribers who don’t like the mail. Identifying those subscribers can be a challenge. We’ve had good luck with some clients, but for others the fixes are a significant challenge and tough for their business model to absorb. The good news is that the pain can be short lived and we’ve been able to ramp mail back up eventually.

Overall, Gmail delivery is harder than a lot of other places. Some of that is because senders use FBLs as a crutch to avoid having effective data hygiene on their signups. That works for those ISPs that send FBL messages, but isn’t so effective at Gmail. Maybe there’s some specific hygiene to do just on Gmail addresses. Working on a better solution than just aggressive hygiene and friction during signup.

Related Posts

Growing your list carefully

Karl Murray wrote a great set of recommendations for growing an email marketing list. I really can’t think of anything I would have said differently. Touching customers and getting contact information from them is great, but there are situations where this gets bad addresses. Too many bad addresses can impact delivery.
So how do you grow your list without falling into a delivery trap? The specific recommendations, as always, depend on your specific situation. But knowing how bad addresses get onto your list will allow you to implement mitigation strategies that actually work.

Read More

Gmail pilots new FBL

Yes, it’s true. Gmail announced last Thursday at M3AAWG that they were piloting a new Feedback loop.
The Gmail FBL is currently for ESPs only. The announcement during MAAWG was that only MAAWG ESP members were eligible. They are requiring a DKIM signature for the FBL, but ESPs using individual customer d= values can get a FBL based on IPs. They are also not providing ANY information that reveals the complainer. Gmail’s intention is only to give ESPs feedback so that ESPs can prevent abuse. They are not giving feedback so complainers can be removed.
The email has a .csv attachment that has 3 columns: date, identifier and complaint rate.
The identifier is an ESP provided customer identifier. One of the ESPs I talked to said they were adding an X-header into their emails.
I’ve heard from beta testers that there is a minimum of 100 complaints before you’ll get any report.
Reports are sent daily if there is sufficient traffic to trigger them.
If you’re a MAAWG member, check the senders list for the signup URL.

Read More

July 2016: The Month in Email

We got to slow down — and even take a brief vacation — in July, but we still managed to do a bit of blogging here and there, which I’ll recap below in case you missed anything.
Sonoma1
At the beginning of the month, I wrote about email address harvesting from LinkedIn. As you might imagine, I’m not a fan. A permissioned relationship on social media does not equate to permission to email. Check out the post for more on mailing social media contacts.
Even people who are collecting addresses responsibly can face challenges. One of the most important challenges to address is paying attention to your existing subscription processes, testing them regularly, evaluating effectiveness and optimizing as needed.
Our most commented-upon post this month was a pointer to a smart writeup about Hillary Clinton’s email server issues. Commenters were pretty evenly split between those who agreed that they see this kind of workaround frequently, and those who felt like regulatory processes do a good job managing against this kind of “shadow IT” behavior. I wrote a followup post on why we see this kind of workaround frequently in email environments, even in regulated industries, and some trends we’re seeing as things improve.
In other election-related email news, we saw the challenges of campaign email being flagged as spam. As I pointed out, this happens to all campaigns, and is nothing unique to the Trump campaign. Still, there are important lessons for marketers here, too, in terms of list management, email content, frequency, and engagement — all of which are inextricably linked to deliverability.
Speaking of spam and engagement, Steve took a look at some clickthrough tracking revealed through a recent spam message I received — and why legitimate marketers should avoid using these sorts of URL referrers.
On the topic of authentication, I wrote a quick post about how seeing ?all in the SPF record tells me one thing: the person managing the record isn’t doing things properly. Need a refresher on authentication? Our most-read blog post of all time can help you out.
And as always, send me your interesting questions and I’ll be happy to consider them as I resume my Ask Laura column in August.

Read More