The end of email open rates: why it’s time to move on
The end of email open rates: Why it’s time to move on
Stop tracking email open rates.
Seriously.
Trying to optimise for them is worse than a vanity stat – it just doesn’t make sense.
I can already hear the wave of angry responses crashing over my opinionated shores, but the data is irrefutable. And to those who really know in email marketing – this has been an open secret for a long time now.
Let me break down why you need to stop obsessing over open rates and start focusing on metrics that actually matter.
How does open tracking work?
Email open tracking hinges on a tiny, invisible image (known as a tracking pixel) embedded in your email. When the email is opened, the following happens:
- Your email client (e.g. Outlook or Mail) looks at the URL of the image – for example https://pathtoanimage.com/image.gif?email=1234567. The end part is important, 1234567 is a unique ID for every email that was sent in a campaign
- The email client then tries to load this image URL
- The web server then reads the unique ID, and logs the fact that the email with that ID has been opened. If it wasn’t opened there would be no reason for this URL to have been called in the first place
- The webserver then returns the invisible image to the email client, leaving the client — or the user, for that matter — unaware of an open being logged because the image is… well… invisible.
Sounds simple, right? And pretty harmless?
Part of the challenge here is that pixel doesn’t just tell you if an email was opened; it’s also a data sponge soaking up your IP address, location, and browser details.
Why it causes privacy concerns
Ten years ago, tracking pixels really were the Wild West of email marketing.
Not only did they track opens, but they also harvested IP addresses, locations, and browser types.
It was like giving someone your house keys because they wanted to check if your doorbell worked.
Understandably, this led to significant privacy concerns. Users were unaware of how much information they were inadvertently sharing, and hackers exploited these vulnerabilities.
This is what prompted the major email providers to do something about it.
Microsoft’s approach
In typical Microsoft fashion, they went route one – the blunt approach. Remember that annoying banner in Outlook asking if you wanted to download images? This is why that came into being.
Without adequate explanation or understanding of why it was there, most people just clicked “yes” to make it go away and stop their email from looking broken. It may have been effective at blocking automatic image loads on paper but did nothing at all for user experience.
Google’s approach
When Gmail grew in popularity, it took a much more user-focused approach.
Around 2020, Gmail introduced image proxying, which fetches images through its servers. It sounds technical, but think back to the four-step process of loading an image pixel: step two involves your email client loading the image.
Google changed this, so instead, a Gmail server loaded it. The location and IP information that was revealed was no longer yours—it was a Google server sitting, most likely, in a data centre in Mountain View.
In one fell swoop, Google solved a major (and largely unknown) privacy concern for all Gmail users. But they didn’t stop there.
Gmail also started preloading images so that when opening an email in your Gmail app or browser, it will load faster. Makes sense for the user but destroys open tracking. The image downloading now happens more like this:
- Email is received by a Google server
- Email is scanned, and images are downloaded and cached
- An open is logged (and potentially IP and location data saved)
At this point, regardless of whether an email has been opened or not…. the open tracking mechanism has been called.
To make matters worse, this is not even consistent behaviour (otherwise, we’d see 100% opens from Gmail recipients) – there is an ever-changing algorithm that Gmail uses internally to determine whether to do it or not. For example, if a user is logged in or an email appears in search results, these can trigger the open being fired without it ever actually happening.
🚨Hot off the press update for August 2024🚨
If there was a further nail needed in the coffin, Gmail has recently started rolling out an Outlook-style banner to prevent the automatic download of images for certain messages.
Whilst it’s not clear exactly what their criteria are (and likely never will be), what is clear is that with the existing false positives and this now creating huge numbers of false negatives… the stats become even more unrealistic and unreliable.
Apple Mail’s approach
Not to be outdone, Apple Mail followed suit in 2022. They implemented similar privacy and caching measures, further muddying the waters for anyone trying to track email opens accurately.
Other providers
While Microsoft, Gmail, and Apple represent a large percentage of email providers and clients, many other email providers have either followed suit or implemented security providers who also automatically scan images in an email before accepting it. All of which cause false opens to be logged.
The impact on open rates: What the data tells us
So that’s the technical explanation done….but what does the real-world data show us?
Gmail open rates
The impact is clear when we look at the data. The chart below shows open rates logged for Gmail, Microsoft, Proofpoint (one of the most popular email security providers) and the average for other providers.
Reported Gmail open rates have consistently been higher, and this trend increased post-2018.
Gmail open rates over time
Maybe Gmail users are just more likely to read emails?
It seems like an unlikely assertion, but again, let’s dive into the data.
Opened in 60 seconds…
Lots of things happen between sending an email and a recipient receiving it, and the likelihood of an email being read within 60 seconds of being sent is low.
Following this logic, if we assume that most opens logged within 60 seconds of sending an email are caused by image preloading or security scanning, we can plot just those on a chart.
This immediately gives a clear indication of what is really happening with your open rates. Between 15 and 20% of all emails sent to Gmail / Gsuite recipients show these false opens, a trend that has been consistent since 2018.
Around 2022, Outlook started rolling out its image proxying, and as expected, the first-minute open rates began to climb.
As this practice of proactively downloading images spreads more widely and is adopted by more providers, first-minute opens (and open rates in general) start to climb accordingly.
Simply filtering out first-minute opens doesn’t help either. If, when a real open occurs, it is served from the cache that was preloaded – guess what? No open is logged. And not all preloads happen in the first minute – this is just a demonstration of the impact of one type of preloading.
We’ve tested multiple combinations of email client and device and have seen false opens occurring when:
- An email appears in search results
- Using a preview pane option in Gmail and loading the next page of emails
Still, think that guru selling you a really clever subject line is what is causing your open rates to skyrocket?
No matter which way you cut it, the makeup of email providers and clients in your user base, coupled with user activity, has a way bigger impact on your open rate than anything else. You just can’t rely on it.
Open rates are a relic of a bygone era. They’re inflated or deflated, unreliable, and do more harm than good.
Focus on metrics that matter
- Genuine click-through rates (Stay tuned for our future article, where we’ll dive just as deep into the world of click tracking.)
- Responses
- Actual engagement
Remember, in the world of email marketing, not all that glitters is gold. Some of it is just glitter.
