Unlike a number of mailbox providers that offer a standard complaint feedback loop, Google offers the ability for senders to have access to a series of dashboards that display metrics about mail sent to Google users. This access is granted through a service that Google calls Postmaster Tools, and in this article I’ll explain a bit about how to get access to this data. Note that you yourself must have a Google or Gmail account in order to take advantage of this service.
The first step to getting access to this data is to register an authentication domain. An authentication domain is either a domain you use to DKIM sign your mail or your bounce (a.k.a., Return-Path) domain if you publish an SPF record. Registration is simple; enter the domain name on Google’s Postmaster Tools website, create a DNS record specified by Google to prove that you own the domain, validate your DNS record, and you’re done. If you’re a SparkPost Enterprise or SparkPost Premium customer and you’ve delegated DNS ownership of your authentication domain to us, then we will register your domain for viewing by our Deliverability Support team. In such cases, we can share access to your domain’s data with you if you’d like, or if you’re an Enterprise or Premium customer managing your own DNS, you can share your data with us.
The following tabs are available in the Google Postmaster Tools dashboard:
- Spam Rate – What percentage of your DKIM-authenticated mail caused Gmail users to complain
- Domain and IP Reputation – What Google thinks of your domain(s) and the IP(s) from which they’ve seen your authenticated mail
- Authentication – How much traffic using your authentication domain(s) has passed authentication checks – SPF, DKIM, and/or DMARC, as applicable
- Encryption – How much of your mail was sent using TLS encryption between the sending server and Google’s servers.
- Delivery Errors – Various metrics showing errors that your authenticated mail may have experienced.
Feedback Loop – See below
The only requirement to see the data on these tabs, other than the Feedback Loop tab, is to just send mail to Google in sufficient volume; Google doesn’t explicitly say how much, instead using the phrase “a sizable daily volume of email traffic (up to the order of hundreds)”.
Google’s Feedback Loop does not function like a traditional one; there is no FBL address for you to enter, and you do not get copies of messages that generated complaints. Instead, if you insert a carefully crafted header into your outbound mail, you can track FBL data in your dashboard across up to four different categories.
The Feedback-ID Header
The key to getting FBL data from Google is to add a Feedback-ID header to each of your messages; the format for this header is as follows:
The SenderId field is a mandatory unique identifier (5-15 characters long) chosen by you, the sender, and it must be the same for all mail sent on a given mail stream.
The other three fields are optional, and can be embedded to track FBL data at the campaign, customer, or other level, such as in this example:
Aggregate data is generated for each of up to four identifiers in the header, starting from the right-most identifier (the SenderId identifier) and displayed on the Feedback Loop tab of the Postmaster Tools dashboard.
This document isn’t meant to be an exhaustive reference to Google’s Postmaster Tools; Google obviously does a much better job on their website of explaining things in more detail. However, we at SparkPost feel like we have a good understanding of how to enroll and how to get value out of your enrollment, so we stand willing to assist our customers in any way that’s needed.
Insights from Marketing Sherpa & The King Freak Himself: Stephen J. Dubner
The Marketing Sherpa Email Summit is winding down in Vegas; the attendees around me have a kind of ‘thousand yard’ stare from what’s certainly one too many late nights. Those that were in attendance for this morning’s keynote have the same stare but are also walking a little taller and filled with the kind of rich data driven storytelling knowledge that only Stephen J. Dubner himself can impart.
Mr. Dubner’s keynote covered a range of topics, and when I say a range I really mean a range: we moved from competitive eating and the birth of a hot dog champion, to economic monetary models built on teaching Capuchin monkeys the value of money. Anyone familiar with Mr. Dubner’s podcasts or the documentary Freakonomics knows that he’s a data-driven storyteller that squeezes insights from the most unusual places.
In his own words: “we’re good at measuring the what, but we rarely know the why.” To me this means that the difficult part isn’t necessarily the testing, but understanding the intention underlying the results. How does this relate to email and messaging? Well let me go back to something else he said: “think differently and redefine the problem you’re trying to solve.” The most successful practitioners of email routinely approach the challenge of how to better communicate with their customers with equal parts imagination, scientific testing and something that smacks of shock and awe. Repetition without invention or iteration is bound to disappoint and be dismissed in a world where information and messaging is around every corner.
This leads me to consider another point: “the right answer isn’t necessarily the best answer.” Solving a problem isn’t the same thing as innovating a solution that changes the nature of business, or the dialogue and experience between customer and brand facilitated through email or other forms of messaging. Playing off of that, let’s throw this concept onto the pyre: “break down mental barriers, don’t pledge allegiance to the way things are done because everyone else does them that way.” Think about the emails that really catch your eye, or the brands that create a name for themselves: do they follow the herd or do they lead it?
Mr. Dubner’s approach holds a great deal of possibility above and beyond the intersection of psychology, economy and sociology. The doctrine of Freakonomics, as it was dropped on us was simple: “don’t be afraid to ask ridiculous and irreverent questions.” Innovation doesn’t happen unless you’re ready to really break the norms and conventions that define what you think you’re doing, or what you assume is the normal way of doing it.
Here’s one solid relevant example of doing things differently. A few weeks ago at the Email Evolution Conference, Sri Somanchi from Google’s GMail team advised that mailers ramp down messages to unengaged users. We’re all familiar with the concept of IP warm up/ramp up in terms of volume to build a reputation for a new IP, but have you ever considered how volumes need to be cranked the other direction when your emails are deleted and never opened? Sri’s advice is to switch from daily to weekly, if the recipient continues to let the emails go unopened, then send a final re-engagement email and stop mailing all together. This is the same advice that Sri has given to the Google marketing team. If we stop and think about what Sri’s doing, he’s approaching the problem in the simplest of terms: if someone wants the message they’ll open and read it, if not, there’s no reason to send it. The accepted wisdom is to never have an expiration date on messaging, keep sending, don’t stop—maybe, just maybe, this bit of counterintuitive advice fits the mold of irreverent thinking, maybe even heretical depending on which side of the fence you stand on. It’s a solid piece of advice and from an insider, someone deep in the trenches of email. So go on, give yourself carte blanche to do things a little bit differently.
When senders and receivers work together, we all win.
Twitter, Google, Microsoft, AOL and Comcast?
That’s right. We promised that Interact 2013 – The Digital Messaging Industry Conference, would be bigger and better – and we’re making good on that.
The five online titans form the A-list ISP panel that will be spearheading the discussion on Preserving the Email Ecosystem. They’ll be discussing sending best practices and acceptance policies, as well as ways senders and receivers can cooperate to keep the email ecosystem clean and spammers at bay.
Josh Aberant from Twitter will be moderating this event, while the panelist line-up includes:
John Scarrow – Microsoft
Lachian Maxwell – AOL
Olga Gavrylyako – Google
Severin Walker – Comcast
If you’ve ever had burning questions about email deliverability, this is the forum to get those concerns addressed by the gatekeepers themselves.
When it comes to actionable takeaways to increase email deliverability, there’s no better way to learn. Aside from the A-list panel, Interact 2013 features two separate tracks: Digital Messaging Best Practices & Interact Bootcamp, where you’ll hear from email experts – both senders and receivers – from Oracle Eloqua, Groupon, Alchemy Worx, Acxiom, Barclays, OTA, XING and more.
Google, Microsoft and Yahoo are failing your DKIM keys? You’re not alone.
By now, just about everyone in the email/messaging/Internet world has heard about Zachary Harris, the mathematician in Florida who uncovered the fact that Google corporate was using weak 512-bit encryption in its email, and how that discovery has snowballed into most of the big ISPs now rejecting email signed with encryption keys less than 1,024 bits in length. You can read the original Wired story or this blogpost from Return Path’s Ken Takahashi for more context.
The whole episode is a good thing in that it’s shined a light on the problem that many senders are still using weak 512-bit or 768-bit encryption. Yet it’s a bad thing in that since many senders haven’t fully come into DKIM compliance yet (the DKIM standard calls for encryption keys at least 1,024 bits) they’re now seeing mailings fail. We’ve long advised our users to upgrade to the 1024-bit DKIM standard, and now it’s really no longer an option.
To upgrade the strength of your DKIM keys to 1024-bit, here are some helpful instructions:
First, MAAWG has published some best practices guidelines that provide a great starting point if you need to upgrade from 512-bit or 768-bit encryption:
- Use a minimum 1024-bit DKIM key length to increase key complexity, as shorter keys, such as 512-bit, are inadequate.
- Keys should be rotated quarterly to reduce the period of time the key could be used to compromise the integrity of email.
- Signatures should have an expiration period greater than your current key rotation period.
- The “t=y” declaration is for testing only.
- To be able to monitor how receivers are accepting email signed with DKIM, it is recommended to implement DMARC with a “p=none” (a.k.a. “monitoring mode”) policy.
- Domain Keys is a deprecated protocol; use DKIM instead.
- Organizations should be engaged with anyone sending mail on their behalf and ensure that their third party email service providers adhere to these same best practices.
Additionally, you’ll also want to decide if you are going to replace the DKIM keys and selector in place, or change over and start signing with new keys and selector. Starting to sign with new keys is probably the optimal decision. Replacing the keys in place will result in any messages that have already been signed, and are in the queue, to fail DKIM validation after you update your DNS record.
We have documentation on the DKIM module on our support site, and for any of our users, we recommend that you consult that material. If you have any questions, please contact our support team and we’ll be in touch.
Learn more about DMARC, the industry’s email authentication standard with our How DMARC Is Saving Email eBook.