Deliverability toolbox: How to use Google’s Postmaster Tool

Over the years, I have seen Gmail come to make up anywhere between 40-60%+ of most B2C marketers’ subscriber lists.

Google’s reporting application, Google Postmaster Tool, has become a pivotal tool for any B2C marketer and has no comparable peer in its class.

It enables email marketers to reliably monitor their spam rates, authentication success, and domain/IP reputation directly with the mailbox provider.

On top of all of this, it is completely free to use.

While Google’s help doc provides the basics of how it works, they do not answer how you can interpret the data like a seasoned veteran. That’s what I’m here to do today. 

Below you will find the culmination of my 5 years of experience with this tool summarized into how-to’s around:

• How to include it in your weekly marketing review

• How to investigate issues elevated in Google’s Postmaster Tool

• Bonus: Common oddities to be aware of

How to use Google’s Postmaster tool in your weekly monitoring

Depending on the frequency of your traffic, I would suggest checking Google’s Postmaster Tool once or twice a week at a minimum, alongside your usual marketing metric reviews.

When doing this, here are the metrics you will want to see regularly trending:

• Domain reputation - Medium, High

• IP reputation - Medium, High

• Spam rate - <.3% on average across 7 days

• Delivery errors - Every day <.5% 

• Authentication - Success rates >95%

• Feedback Loop - No spikes reported

Here are the indicators for when you should begin investigating in case there is an unexpected issue starting to occur:

• Domain or IP reputation drops outside of normal, ongoing trend

• Spam rates trends >.3% over 7 days

• Spam rate spikes >.5% per day

• Delivery error spikes of >5% per day

• Authentication success rate <95%

• Feedback Loop - Any spikes reported

Here are the emergency indicators that a problem definitely needs to be addressed:

• Domain or IP reputation drops and stays at Low, Bad

• Spam rates trends >.5% over 7 days

• Spam rate spikes >1% per day

• Delivery error spikes >10% per day

• Authentication success rate <90%

• Feedback Loop - Any spikes reported

How to investigate issues reported in Google’s Postmaster Tool

Once you observe an elevated problem like one of those outlined above, you need to know what to do with that information…and what it likely means.

Domain reputation drop

Google’s domain reputation utilizes a lot of data points in its scoring, but some of the heavier weighted pieces are aggregate engagement and average spam rate.

This means that domain reputation drops often occur due to either long-term low engagement (low open rates), low/irregular sending frequency, or average spam rates >.3%.

To investigate:

• Check for a correlating spike in spam rate in at least the 2-3 days leading to the drop, if not 7 days+

• Check your ESP reports for unusual email volume spikes in the 2-7 days leading up to the drop, including signup/onboarding flows for un-secure signup forms being abused by bots

• Check ongoing low open rates specific to Gmail subscribers 

• If nothing has changed, check in with other parts of your organization that use the domain for email traffic, including internal teams like Sales who might be sending cold-lead emails, which have high complaint rates

Based on your findings, the appropriate resolution will likely be (but not limited to) one of the following:

• Removing the new audiences being sent unexpected/unsolicited emails

• Discontinue sending to older, unengaged audiences suddenly being sent unexpected emails

• Removing long-term unengaged subscribers (>12 months since last open) from your normal email traffic

• Correcting the misalignment of content and expectations set with subscribers at opt-in

• Securing signup forms with CAPTCHA to protect against bots adding addresses (and cleaning out any you can identify)

• Stopping internal teams from sending cold email traffic (rare use-case)

IP reputation drop

Google’s IP reputation also utilizes a lot of data points in its scoring, but some of the heavier-weighted pieces are spam rates and delivery errors.

This means that IP reputation drops often occur due to either low/irregular sending frequency, average delivery errors >5%, or average spam rates >.3%

To investigate:

• Check for a correlating spike or trend in spam rate in at least the 2-3 days leading to the drop, if not 7 days+

• Check for a correlating spike or trend in delivery errors in the last 7 days and/or last few large-volume sends

• Check your ESP reports for unusual email volume spikes in the 2-7 days/last few similar-volume sends leading up to the drop, including signup/onboarding flows for un-secure signup forms being abused by bots

• Check for inconsistent traffic patterns, including large stretches of time (30 days+) without sending significant volume (50k+/24 hours)

• If utilizing a shared IP and have verified no correlating metrics on your end, check with your ESP to identify any reputation remediation necessary on their end

Based on your findings, the appropriate resolution will likely be (but not limited to) one of the following:

• Removing the new audiences being sent unexpected/unsolicited emails

• Discontinue sending to older, unengaged audiences suddenly being sent unexpected emails

• Removing long-term unengaged subscribers (>12 months since last open) from your normal email traffic

• Correcting the misalignment of content and expectations set with subscribers at opt-in

• Sending at a more consistent frequency to maintain reputation

• Securing signup forms with CAPTCHA to protect against bots adding addresses (and cleaning out any you can identify)

• Working with your ESP around moving to a Custom IP if on a Shared IP with Low/Bad reputation (rare use case)

Spam rate spike

Google’s spam rate focuses specifically on the number of complaints versus the number of messages delivered to the inbox, and, as seen above, can affect both the domain and/or IP reputation if severe or left ongoing long enough.

To investigate:

• Check your ESP account for unusual campaign/audience sends for that day(s)

• Check your ESP account reports for unusual email volume spikes in the 2-7 days/last few similar-volume sends leading up to the drop, including signup/onboarding flows for un-secure signup forms being abused by bots

• Check for an unusual spike upwards in open rates, as more opens could mean sudden inboxing where not inboxing before, which could earn higher complaints

• If nothing has changed, check in with other parts of your organization that use the domain for email traffic, including internal teams like Sales who might be sending cold-lead emails, which have high complaint rates

Based on your findings, the appropriate resolution will likely be (but not limited to) one of the following:

• Removing the new audiences being sent unexpected/unsolicited emails

• Discontinue sending to older, unengaged audiences suddenly being sent unexpected emails

• Removing long-term unengaged subscribers (>12 months since last open) from your normal email traffic

• Correcting the misalignment of content and expectations set with subscribers at opt-in

• Securing signup forms with CAPTCHA to protect against bots adding addresses (and cleaning out any you can identify)

• Stopping internal teams from sending cold email traffic (rare use-case)

Feedback loop spike

Google’s Feedback Loop focuses on elevating especially problematic complaint spikes related to a common feedback ID, usually in relation to something like a campaign ID. It will only trigger for especially egregious complaint spikes so you can expect (and hope) that you will very rarely see anything reported here.

To investigate:

• Check the relevant ID provided in the report

• Review the campaign related to that ID

• Identify issues between audience and content

  • Is this audience opt-in only?

  • Is it recently engaged?

  • Is the content appropriate for the audience?

Based on your findings, the appropriate resolution will likely be (but not limited to) one of the following:

• Removing the new audiences being sent unexpected/unsolicited emails

• Discontinue sending to older, unengaged audiences suddenly being sent unexpected emails

• Correcting the misalignment of content and expectations set with subscribers at opt-in

Delivery errors spike

Google’s Delivery Errors focuses on elevating both delivery deferrals and bounces, though it is very limited in providing which error types are occurring like the “rate limit exceeded” error.

Significant enough bounces can cause a degradation in reputation (both domain and IP) over time while deferrals can already be an indicator of reputation issues so spikes should be investigated.

However, unlike with the Spam Report, delivery issues can be elevated by your ESP so you can pair information from both GPT and your ESP to investigate the issue.

To investigate:

• Identify the date the errors occurred

• Select the date to see if Google gives information related to these errors

• Check your ESP’s campaign reports for those days to identify the campaign or audience that saw the spike

• Review the Gmail errors observed in the ESP reports to identify/verify a theme 

Based on your findings, the appropriate resolution will likely be (but not limited to) one of the following:

• Removing the new audiences being sent unexpected/unsolicited emails

• Discontinue sending to older, unengaged audiences suddenly being sent unexpected emails

• Removing long-term unengaged subscribers (>12 months since last open) from your normal email traffic

• Securing signup forms with CAPTCHA to protect against bots adding addresses (and cleaning out any you can identify)

Authentication success rate drop

Google’s Authentication checks the success rate for SPF, DKIM, and DMARC for your domain’s traffic. In most cases, your authentication should be good to go once set up so issues here are rarer and likely related to new traffic streams, technical issues in the authentication validation process, or suddenly missing DNS records.

To investigate:

• Identify which type(s) of authentication failed that day

• Review DMARC reports for those days from Google to isolate which IP/mail stream is the likely culprit

• Audit that mailstream to make sure authentication is in place

• Check with support for that mailstream for further assistance if success rate has significantly dropped and stayed down

Based on your findings, the appropriate resolution will likely be (but not limited to) one of the following:

• Adding/fixing authentication for the relevant mail stream

• Setting up DMARC reporting to more easily identify unauthenticated mail streams 

Encryption inbound TLS rate drop

Google’s Encryption reports both the inbound and outbound traffic’s TLS connection rate (inbound is email traffic being sent TO Google and outbound is email traffic being sent FROM Google if you even use Google).

A TLS rate drop can often be associated with limited technical resources available and is not usually a serious issue to worry about. It doesn’t cause deliverability issues so investigating this is less of a priority unless a majority of inbound traffic is failing.

To investigate:

• Identify which days saw the issue and what rate it dropped

• Review which traffic streams delivered on that day

• Test delivery experience to see if emails arrive from that source with a red, unhooked lock insignia on it

Based on your findings, the appropriate resolution will likely be (but not limited to) one of the following:

• Work with the relevant technical support teams to make sure that Google’s TLS best practices are being followed

Bonus: Common oddities to be aware of

Google’s Postmaster Tool is a fantastic tool for what it does, but it does have some oddities that you should be aware of.

They have a breakdown of some FAQs, but here are some of my own observations.

Subdomain setup

If you set up and verify the root domain in GPT, then you don’t have to do the same lengthy setup process for each subdomain. You can then just add the subdomain and it will automatically add to your view since you verified the root domain

Data delay

Data is processed and reported by Google in GPT with a 2-day delay. That can fluctuate from time to time but on average that is the closest recency in data you will get.

Also, each day is based on the UTC timezone.

Thresholds for receiving report data

Different sections only report based on a specific volume threshold and will otherwise show nothing for that day.

IP is the strictest. Anything below ~2k per day (non-official number) Gmail volume will likely not show up here.

Spam rate and domain reputation can be a bit lower threshold.

The Feedback Loop, as mentioned above, will only trigger when significant complaint rates (like >1%) are related to a shared ID (like a campaign ID).

Missing Delivery Errors reasons

At this time, not all of Google’s delivery errors are provided in summary here. Only specific errors, like “rate limit exceeded,” are detailed in the expansion when you click on the date.

Outside of that, it will be very common for you to click on a date’s spike and see the reason percentage not totalling that date’s spike. That is simply how it is at this point in time. You should be able to investigate in more detail within your ESP.

Final word

What an exhaustive list! You are now ready to engage with Google’s tool ahead of the learning curve.

Remember, the above is a starting point. You may encounter strange oddities that don’t make sense. I will admit (along with many of my peers) that I still experience this even after 5 years of using it.

Don’t be discouraged, though. Google’s system is very responsive to senders who follow best practices and make sure their traffic is highly engaged and explicitly opted into. You can repair any damage you find yourself in.

As seen in the guidance breakdowns above, a few answers resolve a lot of issues. Making sure your content is relevant and only sent to subscribers who have explicitly opted into it helps fix most things.

In that way, reaching the inbox is made easier than ever.