Dec 12, 2013
How Gmail Tabs Affect Your Email Marketing: An Update
More than half a year has passed since Gmail announced the new tabbed inbox. My initial research hinted that open rates took a hit, but at that point I was just looking at a few weeks’ worth of opens. Now that things have settled down, we can look at months’ worth of all kinds of engagement data. To be specific, I pulled 29 billion emails, 4.9 billion opens, 4.2 billion clicks, and 43.5 million unsubscribes to find out how Gmail’s tabs are affecting open rates. (The Ops team wasn’t exactly happy with me for a few weeks there. Apparently, they’re very protective of our backup databases—and the backups of the backups.)
After looking over all the data, my conclusion is that Gmail’s tabbed inbox is working as intended. Open rates are definitely down, but there are reasons to believe Gmail users might be happy with the changes.
A matter of timing
One of the most telling stats about Gmail’s tabbed inbox is the time it takes users to open an email. I looked at every open for every campaign and measured the percentage of opens that come in minute by minute after the campaign is sent. The density of opens by minute tells us how quickly or slowly subscribers are getting to the emails in their inbox.
At first, I plotted every week since the beginning of May, and three distinct groups immediately stood out. Those groups lined up perfectly with the weeks before, during, and after the rollout of the tabbed inbox.
As a control measure, I applied those same time periods to some other email service providers (hereafter referred to as ISPs to avoid confusion). Gmail’s numbers are clearly a little wonky.

What you see is that during the rollout period, Gmail subscribers were actually getting to their emails faster than they normally would. That makes sense—there was a lot of buzz about the new inbox at the time, and everyone was excited to try it out.
After everyone got tired of clicking on tabs for the sheer fun of it, there was a marked shift in opens. Fewer opens come in around the 18 minute mark, while slightly more subscribers are opening their emails 40-55 minutes after the campaign was sent. Meanwhile, the other ISPs experienced little to no change. This tells me that Gmail subscribers are definitely using tabs to ignore certain emails.
Opens
Using the same time periods as above for "before" and "after," you can see the Gmail opens are down. My initial research encompassed a 6-week period around the release of tabs, but this data represents a 25-week period from May to October.


Opens from Gmail subscribers have fallen off nearly twice as much as any other ISP. If you’re surprised that other ISPs are also seeing lower opens, consider that MailChimp starts experiencing a lot of holiday-related volume in October. You might be familiar with a seasonal trend we call holiday email fatigue.
Clicks
To study clicks, I divided the number of clicks by the number of opens. That way, I could look at clicks in isolation from any changes to open rates.


I see another example of holiday fatigue in the data. It looks like Gmail click rates are down compared to Hotmail and Yahoo, but it’s not as bad as AOL or all the other ISPs out there. Maybe this has something to do with your ISP personality?
All in all, I’d call this one neutral for Gmail tabs. Clicks are down for everyone this time of year, and Gmail subscribers don’t look particularly affected.
Unsubscribes
This is where the data really started to paint a picture. Again, I looked at unsubscribes as a percentage of opens, and Gmail stands out as having no shift from May to October.


As a function of normal holiday fatigue, subscribers at other ISPs are starting to disengage with the newsletters they signed up for. They’re opening less, they’re clicking less, and they’re starting to unsubscribe more…except for Gmail. Gmail’s opens are down, the click rate is neutral, but unsubscribes haven’t gone up.
What does this mean? People unsubscribe around the holidays, because they get frustrated at the sheer volume of email hitting their inboxes. For Gmail subscribers, that frustration isn’t showing up in the data. If the data tells us a story, it’s that Gmail tabs are working as intended and helping people manage their inboxes.
Mileage may vary
Ultimately, the effect Gmail’s new inbox has on your email marketing depends on the relationship you have with your subscribers. I can’t create charts for each and every user, but I can narrow down the industries that are most affected.

The education and training industry really stood out to me. When kids go back to school in August, we usually see a huge bump in engagement to education and training campaigns. You can see that reflected from every ISP except Gmail. If you’re in any of these four industries, you might need to make an extra effort to reduce irrelevance.
At the end of the day, it seems people are using tabs to ignore certain emails, and it’s actually making them less frustrated with their inbox. Gmail would consider that a win. Is it possible they’re missing emails that would otherwise be important to them? Sure—but maybe they’re just getting to those emails a little later than they used to.
If you’re just catching up and want some context, read my July 23 post about Gmail tabs and open rates.
Orlando
Great insights – that’s a lot of number crunching done! If people unsubscribe less because of the tabbed approach that is definitely a win for GMail as it keeps the clutter away in a way.
12.12.2013
Chris
Excellent analysis!
I’m wondering – is this just to gmail addresses, or are google apps domains included in the gmail data?
Grateful that you’re doing this kind of work, and publishing the results.
12.12.2013
Matthew MailChimp
For the sake of simplicity, I only looked at gmail.com, hotmail.com, yahoo.com, and aol.com email addresses directly.
Trying to hunt down and include all the secondary and tertiary services those four companies control is like falling into the rabbit hole. I know that skews the “other” category a little bit, but the effect should be minimal.
12.12.2013
Christine
Very informative! Glad my clients tend to favor Gmail more than others. Even then, majority of people who have AOL, Hotmail, etc. Seem to be older and don’t want to switch to newer companies. Considering Google is known for spamming people by their searches, favorites and keywords said in everyday conversations, I’m surprised they’re numbers haven’t dropped dramatically since the tabs have been introduced. As a Gmail account holder myself, I’ve noticed advertisements of just sayings come across the top when I type certain words. No matter the plug-ins, protection apps, etc. that you use, Google still finds a way around it. As great as they are for our email campaigns, Gmail is not so wonderful anymore for personal use.
12.31.2013
Maxwell D.
As for Christine’s problem with Google, I think it is probably ad-ware. Look around for some malicious plugins/add-ons in your browsers.
04.07.2014
Jean
Actually I came here to see if you guys have anything to say about gmail now proxy’ing images… I was happy as a user, but unhappy as someone who wants to track email opens :(
12.12.2013
Joe
Thanks for the great analysis! I’m actually kinda surprised to see the gmail click rates so low. I’ve been hoping (maybe naively so) that the segregation to the marketing tab would help to put the recipient into more of a shopping mode when they do actually open – sure, the opens will plummet but the quality view might become a little richer… doesn’t look like that is the case. Still, its great to see that proved out one way or the other.
12.12.2013
Matthew MailChimp
Hey Joe, I had the same story in mind when I started looking at clicks. I’m not sure why we don’t see that behavior in the data though.
Still, we normally see opens and clicks decline as we approach December, and I’d say the change in Gmail’s click rate struck me as very neutral. Unlike opens or unsubs, Gmail is right in the middle compared to other ISPs.
12.12.2013
Erroin
Awesome analysis!
12.13.2013
jenny
Fantastic analysis across the board very insightful information have noticed though a lot more of my prioritised email not getting through at all since it changed doesnt even go to trash not sure why but a few of my colleagues have found the same
12.16.2013
Mary Ann
Matthew,
Very interesting. I would like to know how Gmail decides what is primary, social or promotional email. Any ideas?
12.16.2013
John MailChimp
Hi Mary Ann, There’s definitely a lot of thoughts and opinions about how Gmail is categorizing email for tabs. That being said, Google hasn’t really tipped much in regards to their process and I’m sure much like their search algorithms, it’ll continue to evolve over time.
12.18.2013
Adrian
Have you considered mobile? On my (Android) phone, Gmail is setup as default to not give notifications for tabbed folders (especially updates & promotions) so I only see these emails when I physically go into the Gmail app. Of course I could change the setting so I would get notified, but the majority of people wouldn’t.
This does mean that it takes longer for me to see the emails in these folders, sometimes as much as a few hours after they’ve actually arrived. But it’s not because I’m choosing to ignore/hide them!
12.17.2013
John MailChimp
Hi Adrian, I’ll pass this along, but I’m not sure there’d be any way to really research that particular piece. Without Google sharing the percentage of mobile users that have the setup as you describe, I don’t think we’d have the necessary data. Thanks for the suggestion.
12.17.2013
Adrian Willings
Well no, but it is food for thought. :)
12.19.2013
Maggie Georgieva
Good stuff! Thanks for the thoughtful post. Unsubscribes seem to have remained flat, but I wonder about deletes. Now it is much easier to bulk delete a lot of the emails in those other tabs. Any insight on that?
Cheers,
Maggie
12.18.2013
John MailChimp
Hi Maggie, The actual deletion of emails from the Gmail inboxes would be something only Google would know, but I think the impact on open rates could be an indicator of that behavior on some level.
12.19.2013
Jason
This is great analysis, and quite comforting information to know. :) I’ll focus on writing great emails that people look forward to receiving and not worry so much about whether or not they’re being seen.
-jason
12.23.2013
Nathaniel Cassidy
Fascinating insights. If Google’s tabbed inbox means users begin to dip into different types of content when they choose over time this could mean a change in strategy for email marketers – less time driven content perhaps?
12.30.2013
Raj
This analysis is flawed for following reasons…
1. Without seperating mobile vs desktop open rate this analysis will skew the results…as gmail tabs dont work on iphone mail app
can you take out mobile and show case the results?
also
what is the % open rate drop of gmail users who were engaged before vs after tabs to the same campaign (again taking out mobile users ) and then for just mobile users only
01.15.2014
Shahid
Thanks Matthew. Another really insightful post. Helpful post especially comparison between other giants.
01.15.2014
junkcharts
Two comments:
can you check your clickthrough rates? I can’t believe that the average rate over billions of emails is 85%.
also, that click rate chart should be a line chart. Column charts need to have axis starting at zero.
01.18.2014
Adam
Hey Matt,
Can you double check those click-through rate numbers? 85% seems very, very high.
Your major competitor sees click through rates 10-20% (varies by industry)
http://img.constantcontact.com/training/tutorial/industryavg_621.pdf
01.19.2014
Matthew MailChimp
Hey Adam,
I definitely used a different method to calculate click rates for this study. Normally, you would say clicks / sends = click rate. In this case, I was interested in seeing if the total click engagement changed before and after the introduction of tabs, and I wanted to control for changes in overall delivery and open rates.
To isolate click rates from changes in delivery and open rates over time, I decided to divide clicks by unique opens. Since I was interested in the total amount of click engagement, I did not use unique clicks (one click per subscriber). Instead, I counted every click even though I was only interested in the first open.
Measuring clicks / opens like this really tells you the quality of the opens you are getting. If this kind of click rate had plummeted for Gmail, I would have concluded that Gmail tabs were a complete failure. It would have indicated that subscribers who used to be very interested in a particular email weren’t seeing it anymore after tabs were introduced. Instead, this click rate remained completely steady, which tells me that subscribers who like a particular email are still seeing it and getting to it.
So clicks / sends probably does get you 10-20%. On the other hand, total clicks / unique opens is more like 85%. At least, in my data it was :)
01.21.2014
junkcharts
Matthew: even clicks to unique opens would be nowhere near 85% on average! That would imply that 85% of the people who open emails click on something. Your numbers may be skewed by say clickbots. Maybe you should try unique clicks / unique opens.
click / send of 10-20% on average is an astounding number that seems high by one or two orders of magnitude.
also, you should start column chart axis from zero, or use a line chart.
the spirit of the blog post is great though.
01.21.2014
Artur | JuicyCanvas
Our Open Rates dropped to 2% after Tabs… and remained as such for 4months.
Unfortunately, we’ve had to switch to a more home-made solution for our email blasting… one that easily goes around gmail’s tab and gets right into the user’s inbox.
Now our open rates are back to 35%.
No other way to read those #’s for a growing startup.
01.21.2014
John MailChimp
Hi Artur, Hrmmm… something sounds a little amiss. Unless something else changed with content or something else, that sounds like an awful big drop. If you wouldn’t mind, it’d be great if you could reach out to our support team at: http://mailchimp.com/chat They’d be able to take a closer look and maybe offer some suggestions to get you back on track.
01.22.2014
Humphrey
Brilliant analysis! This helps a lot!
01.24.2014
Dan
Great analysis. How is the change impacting “social” emails specifically?
02.19.2014
Kaleb
This is a great analysis. Is there anyway to do a year-over-year analysis in order to remove the holiday fatigue that you mentioned…or at least normalize it to some extent? Also, I agree with Raj’s stipulations about further refining and analyzing by smartphone, tablet, and desktop since those could skew the data given mobile’s 51% penetration of email opens?
02.27.2014
John MailChimp
Hi Kaleb, The team is always thinking of different ways to crunch the numbers, but I’m not sure this in on the list. No guarantees, but I’ll pass it along to see if it sparks any new ideas for them to dig into.
02.27.2014