If you’ve visited any forums or read any blog posts related to the email industry in the past few days, you’ve probably seen mention of Google’s Postmaster Tools. Specifically, the tools have been experiencing a partial outage, leaving many marketers and deliverability professionals searching for solace or news about a fix (count us among that group!).
To help pass the time until that sweet, sweet Google data is flowing again, here’s a primer on Google Postmaster Tools (GPT) — what they are, how they’re used, and how they complement your own (or your ESP’s) data and services to optimize delivery to Gmail.
Gmail’s unique challenges
Google does email differently, and that can be a challenge for senders trying to deliver mail to their users. For example, Gmail doesn’t offer a traditional feedback loop (FBL) that forwards spam complaints to the sender so they can be removed from their mailing list. Recipients who mark your messages as spam will continue to receive them, potentially damaging your delivery to other recipients and adding a layer of complexity to your list hygiene. Gmail was also the first provider to rely heavily on engagement metrics to filter mail — if recipients aren’t engaging, your mail doesn’t reach the inbox.
In the time before GPT, senders would closely monitor their open rates at Gmail, watching for trends and often removing those recipients who didn’t open regularly. But if this didn’t work, they had little recourse to dig deeper and virtually no way to contact Google for assistance.
Google opens the curtain (a little)
In July 2015, Google released Postmaster Tools to provide senders some insight into their reputation and performance when sending to Gmail. These tools allow you to see data from Google in 7 categories:
Spam rate is the percentage of messages Gmail recipients marked as spam out of all mail that reached the inbox for the given date. Messages that land in spam are not considered in this rate, and only mail signed with DKIM is counted.
IP Reputation gives the sending reputation of the listed IPs on the given date. Each IP is assigned a status of High, Medium, Low, or Bad.
Domain Reputation gives the reputation of the domain for the given date. Each domain is also assigned a status of High, Medium, Low, or Bad.
FBL provides the rate of FBL complaints received by Google by identifier. Senders must be registered with the Google FBL program to receive data.
Authentication is the rate of traffic that passed sender authentication checks (SPF, DKIM, and/or DMARC).
Encryption shows the rate of mail to Gmail that was TLS encrypted.
Delivery Errors is the rate of mail that was rejected or temporarily deferred (tempfailed) by Google on the given date. Clicking on a date will show more specifics on why the mail failed delivery.
Senders can register their domains with GPT to view all these metrics, with a couple of caveats. Only days in which significant traffic is sent to Google will show data (there is no specific volume threshold, but our experience is at least a few hundred a day). Google may also withhold some or all data if they don’t believe you to be a “trusted” sender. Again, there is no specific criteria for this, but their algorithms evaluate whether you are sending mail they deem legitimate.
GPT data is also generally a day or so behind real time (i.e. the data that’s uploaded today will provide insight into yesterday’s sending), so plan accordingly.
Putting the pieces together
Google’s data provides great insight for senders, but it’s not going to resolve delivery issues alone. Nearly every ESP uses GPT data as part of their deliverability monitoring and troubleshooting regimen. Using the data correctly can help identify small issues before they escalate or ensure your reputation is still intact after a bad send or targeting mistake. Here are a couple of examples where GPT is used to help keep Gmail delivery running smoothly.
Google (and most major providers) monitor both IP and domain when determining how to route incoming mail. Any time you send from a new IP, domain, or both, you’ll have to slow down sending to allow the provider time to acclimate to your new sending infrastructure and patterns. Gmail can be especially tricky during a volume ramp, but GPT helps the process go more smoothly.
When ramping volume to Gmail, you should be monitoring your open and bounce rates for Gmail recipients alongside the IP and Domain Reputation metrics within GPT. Each time volume increases, check the results to ensure your reputation is steady or increasing. Ensure Spam Rates aren’t spiking, and Delivery Errors remain low or nonexistent. Any time you see a spike in Spam or Delivery Errors, dial back volume temporarily. If IP or Domain Reputation drops, dial back volume as well until you see a course correction. Even a drop from High to Medium reputation could be a sign of issues to come. Ignoring this drop and continuing to ramp volume can lead to delivery problems that will take much more drastic measures to correct.
When setting up a new domain or updating your authentication, GPT can be a great way to confirm your new settings are working properly. Check the graphs for SPF and DKIM pass rates; if you see failures, double-check your setup. Since Gmail addresses make up about 40% of the average B2C list, authentication failures here are likely to have a notable impact on your mail’s performance. Google doesn’t automatically route mail to spam for authentication failures, but it is one of the hundreds of signals they use — the more you have in your favor, the better.
Tracking delivery failures
The Delivery Errors view can be helpful if you don’t have direct access to the SMTP bounce responses sent back when your mail is rejected. If you’re seeing bounces from Google on a specific day, you can go here to view Google’s bounce reasons. This can be especially helpful if your ESP doesn’t capture tempfails or deferral bounces in your reporting.
Solving deliverability problems
When delivery to Gmail goes wrong, there’s really no one at Google you can contact to make it right. All the filtering is run by algorithms with little manual intervention. This means GPT data is invaluable when trying to solve problems delivering to Gmail. Typically, the IP and Domain Reputation views are most helpful, as you’ll usually see a dip there correlating to your issues. You’ll usually also see a trend in the metrics — reputation dropped from High to Medium on a certain date, then to Low, then maybe to Bad. When possible, you should start making adjustments any time you see a drop in reputation.
If you’re already in Bad or Low status, you can often identify the date things went wrong. Did you send to a new list that day? Increase volume exponentially? Use creative that might be controversial or disliked by recipients? If there’s no single item you can pinpoint, you’ll likely want to treat your remediation like a warm-up: decrease volume and send only to your most engaged recipients, then slowly increase as metrics improve.
Google Postmaster Tools may not be perfect, but they are a valuable tool to be included in your (and your ESP’s) deliverability arsenal. There’s a lot of buzz out about potential updates to the tools — an API has been discussed for a few years now — but no official word has been released. In the meantime, we’ll keep using the tools we have and keep our eyes open for any shiny new updates.