Your open rate just got less reliable (again)

Hey, it's Pavel.

Gmail has been making changes that affect every newsletter operator, and most people are either panicking about the wrong one or haven't noticed the second one yet. This issue covers both. By the end you'll know exactly what's happening inside Gmail right now, what it means for your list, and what's actually worth doing about it.

58% of buyers now start their research in ChatGPT or Gemini, not Google. Most startups aren't showing up there yet.

The ones that are get cited by the AI tools their buyers, investors, and future hires already use. And they convert at 3×.

Download the free AEO Playbook for Startups from HubSpot and get the exact steps to start showing up. Five minutes to read.

Get the free playbook

Gmail is opening your emails before your subscribers do

Before we get into what this means for you as a sender, it helps to understand why Gmail does this in the first place, because it's not about tracking and it's not about marketing attribution. It's about security and user experience.

When an email arrives in someone's inbox, Gmail wants to know what's inside it before displaying it to the user. It scans the content, checks images for malicious code, and assesses risk before the subscriber ever sees the message. At the same time, by fetching images in advance, Gmail ensures that when the subscriber does open the email everything loads instantly rather than waiting for external servers to respond. For the person reading their inbox, this is a genuinely better experience. Emails feel faster, safer content gets caught before it reaches them, and the whole thing works more smoothly.

The problem is that your tracking pixel is an image too. So when Gmail fetches everything before the email is displayed, it fires your tracking pixel in the process. Your ESP receives the signal and logs it as an open, even though the subscriber hasn't touched anything yet.

Here's how it works in practice: if someone has Gmail open on web or mobile and your email arrives while their session is active, Gmail prefetches all the images in that email immediately before displaying it in the UI. The prefetch comes from a Google-owned IP address, happens within seconds of delivery, and uses a specific user-agent string that's different from what you see on a genuine open. The subscriber might never actually read that email. But your dashboard already counted it.

For senders, this creates a few concrete problems. Your open rate is inflated, which gives you a false sense of how engaged your list actually is. Any automation you've built that fires on an open event — a follow-up sequence, a re-engagement campaign, a time-sensitive offer — can trigger for people who haven't seen anything yet, which means those subscribers receive messages without any context for why they're getting them. And if you're using open rate to make decisions about list segmentation or sending frequency, you're working from a number that includes people who never opened your email in any meaningful sense.

SparkPost analysed nearly 10 billion Gmail open events and found that between 1 and 6% of opens are false, with the average inflation sitting around 2 percentage points. How much this affects you specifically depends on how Gmail-heavy your list is. If your audience skews heavily Gmail and you think you're sitting at a 35% open rate, the real number is probably closer to 33%. That gap compounds over time, especially when you're building segments and automation logic on top of it.

Worth knowing: this isn't actually new

Gmail has been caching images since 2013, and the specific prefetching behaviour described above was first documented by SparkPost back in early 2022. Most of the posts going around about this right now are rediscovering something that's been quietly happening for years.

That doesn't make it less relevant. It actually makes it more concerning, because it means a lot of senders have been working with inflated open rate data for longer than they realise. But I'd rather you know the full picture than walk away thinking Gmail flipped a switch last week.

The genuinely new development, and the one with more immediate practical consequences for your list, is something else entirely. I'll get to that next.

The question you actually need to answer first

Before you go adjusting anything, you need to know where your mail is actually landing, because prefetching is adding noise to a signal that was already broken. There are only three places your email ends up and each one tells a completely different story.

If your mail is landing in the inbox and open rates are still low, your list is probably stale. Those Gmail addresses are technically valid but the people behind them have moved on and aren't actively using those accounts anymore. If it's going to the promotions tab, open rates will naturally be lower, but here's something worth sitting with: open rate doesn't correlate with revenue the way most people assume, and mail in the promotions tab still converts. The click matters more than the tab it came from. If it's going to spam, you've got a genuine reputation problem that no subject line test or content tweak will fix. The path out is specific: stop sending to anyone who hasn't engaged in the past few weeks or months, let your reputation stabilise, and then figure out why engaged subscribers were telling Google your mail didn't belong in their inbox. Once you've fixed that underlying issue you can start ramping volume back up, but not before.

The simplest way to check where you're landing is to send a test to a fresh Gmail address you control with no history and no prior interaction with your domain, and see where it ends up. That single test tells you more than a month of dashboard data ever will.

Now the update that actually is new

Gmail recently started letting users in the US and UK change their email address and keep the old one active as an alias, with both addresses continuing to receive mail. Google hasn't released any detail on how this interacts with unsubscribe flows, and the only guardrail currently in place is a 12-month limit on how often someone can make the change. The rest of the world is expected to follow soon.

For newsletter operators this creates a specific and genuinely annoying problem. One person ends up with two valid email addresses and there's no obvious link between them anywhere in your ESP. If someone switches their Gmail username and then signs up again with their new address because they forgot they were already subscribed, or simply because your form accepts it, you now have duplicate profiles with no clean way to connect them. You'll send the same email twice to the same human. If they unsubscribe on one address the other stays active, because unsubscribe is tied to the address and not the person behind it. Google hasn't signalled that this will change anytime soon.

The deduplication problem is genuinely difficult to solve cleanly. You can try to match on phone number, physical address, device behaviour or purchase data, but every approach has edge cases. Households share addresses, B2B contacts use different numbers for work and personal, and behaviour can still be attached to the old email even after someone has switched. Any matching you do is probabilistic, not definitive.

That said, I don't think this is as catastrophic as some people are making it sound, and here's what's actually worth doing right now.

Prioritise the most recent engagement signal. If you suspect two profiles belong to the same person, send to whichever address opened or clicked most recently and suppress the quieter one. You're not losing a real subscriber, you're removing a ghost. Watch for soft duplicates using simple logic like the same surname and postcode or the same phone number, and flag them rather than auto-merging so a human can make the call on ambiguous cases. Tighten your disengagement thresholds slightly, because old alias addresses won't go fully inactive right away but engagement will trend down, and moving faster on suppression protects your sender reputation at the same time. And make it easy for subscribers to update their address with you directly. A simple prompt post-purchase or at login asking them to update if they've changed their email lets your most engaged readers connect the dots themselves, which is the cleanest fix available right now.

What you can't do is perfectly solve this, and trying to clean your way out of it completely is a waste of time. Email is no longer a reliable unique identifier on its own, which was probably already true before this update and is just more visible now.

The bigger picture

Both of these Gmail developments are pointing in the same direction. Open rates are getting harder to trust, and email addresses are getting harder to treat as unique identifiers. The foundations most newsletter operators have been measuring on are shifting, and the senders who adapt fastest are the ones who stop building strategy around metrics that are increasingly easy to corrupt.

Opens are directional at best and misleading at worst. If your sending strategy still leans on them for segmentation decisions, automation triggers or re-engagement logic, it's worth rebuilding that foundation on signals that are actually clean. Clicks tell you who's engaged. Replies tell you who's really engaged. Inbox placement tells you whether Gmail is working with you or against you, and those three things are still reliable because they're the signals Gmail itself uses to decide whether your next send lands where it should.

If your numbers are looking off and you can't tell whether you're dealing with a placement issue, a list quality problem or just bad data from prefetching, that's exactly what I work through with beehiiv newsletter operators. A proper look at the data usually makes the diagnosis obvious within the first conversation, so reply to this and tell me what you're seeing.

Before You Go: Here's How I Can Help

Until next time,

Pavel