The general consensus about how Gmail makes inbox placement decisions is “engagement,” in addition to more traditional reputation.
Typically, deliverability experts mean two different things when they say engagement:
Subscriber-level engagement: Clients that positively interact with messages (read them, move them to folders, reply to them) from a domain on IP address are more likely to receive mail from that source in the future.
Global engagement: If the overall engagement with mail from a source across all subscribers is low, then all mail from that source may have a delivery problem.
If “engagement” is the root cause of a deliverability problem, the obvious solution should be suppression of disengaged subscribers, as this will help your global engagement and get more mail into the inbox. As teams at my company have taken a much deeper look at inbox placement at Gmail across our panel of over 2 million mailboxes, it becomes clear that the rules used by Gmail are more complex and nuanced than previously thought. Standard solutions may not always work.
Let me give you two examples that we’ve recently run across, which provide a potential explanation of what might be causing the problem, and potential solutions.
Case #1: Engaged subscribers not getting emails delivered, while disengaged subscribers are. For this client, many of their most engaged customers were not getting emails delivered, but a larger portion of their least-engaged subscribers were making it into the inbox.
Weird, right? A deeper review showed that the highly engaged segment were less engaged with the sending domain than with other domains that they received mail from—although they were very engaged relative to other subscribers on this sender’s list. Does this mean that relative engagement is an important factor in Gmail inbox placement decisions at the subscriber level? Clearly pulling back on the least-engaged subscribers wouldn’t have been the answer here. The likely answer is to find ways to further increase engagement among the most engaged segment.
Case #2: Deliverability tanks, but opens and clicks stay constant. This case is a little easier to explain. A review of the panel data showed that the marketer was sending more mail to the least-engaged subscribers than to more-engaged subscribers. The most-engaged (and profitable) subscribers continued to see inbox placement rates and continued to open and click at high rates. Although a smaller portion of overall messages are getting delivered, this deliverability problem isn’t causing an economic problem for the marketer. I would argue in this case, the right “treatment” is for the marketer to keep doing what they are doing (though they are running a little risk with their global reputation).
I hope that these two examples are sufficient to illustrate that Gmail deliverability is complex and nuanced—likely the result of very sophisticated algorithms that take many factors into consideration that are far more complex than most simple mental models.
Understanding inbox placement by general mailbox engagement with all email, engagement with the sender, and economic value of each subscriber will help you make better decisions. However, there is no one-size-fits-all solution. As always, doing your best to build engaging email programs tailored to the needs of subscribers will drive better deliverability and better email program results.