← Blog
Deliverability

22 Apple Mail privacy & bot-open statistics that matter: 2026 edition

Apple now drives nearly two-thirds of all email opens, and roughly half of them are machines, not people. These 22 sourced statistics show how Apple Mail Privacy Protection and bots broke the open rate - and which metrics still tell the truth.

11 min readPublished July 11, 2026Sendaloft Research

The email open used to be a proxy for attention: a pixel loaded, so a human must have looked. That link is now broken. When Apple launched Mail Privacy Protection in late 2021, it began pre-fetching the tracking image for every protected recipient the moment mail arrived - firing an "open" whether or not anyone read a word. Security scanners and, more recently, AI assistants do the same at scale. The number in your dashboard labelled "open rate" is now a blend of humans, proxies, and bots.

That matters because so much of email marketing was built on top of that number. Open-based "engaged" segments, open-triggered automations, sunset rules, subject-line A/B tests - all of them assume an open means a person. In 2026, with Apple owning the majority of opens and bots making up roughly half of internet traffic, that assumption quietly poisons segmentation, reputation signals, and reporting alike.

Here are 22 sourced statistics on Apple Mail Privacy Protection, machine opens, and bot traffic - covering email client share, MPP adoption, open-rate inflation, bots and scanners, the damage to automation and segmentation, and the shift to click and conversion metrics - with what each one means for your program.

Key takeaways

Apple and Gmail own the inbox

1. Apple is the dominant email environment, with 64.66% of all opens

Litmus's email client market share for May 2026 shows Apple - iPhone, iPad, and Apple Mail on Mac combined - accounting for 64.66% of all email opens. Nearly two out of every three opens happen inside an environment Apple controls, which is precisely why Apple's privacy defaults reshape the whole measurement picture.

2. Gmail is the clear number two at 24.11% of opens

Gmail is the second-largest email client, capturing 24.11% of email opens in May 2026. Google runs its own image proxy that caches remote content, so Gmail opens carry their own layer of machine activity on top of the human ones.

3. Outlook has fallen to just 6.49% of opens

Outlook now registers only 6.49% of email opens. The historical "design for Outlook" instinct no longer matches reality - the modern inbox is overwhelmingly mobile and Apple-first, and rendering and measurement decisions should follow the audience that is actually there.

4. Apple and Gmail together account for nearly 90% of the market

Add the two leaders and you cover almost the entire inbox: Apple Mail and Gmail together make up nearly 90% of the email client market. Both operate content-proxying privacy features, which means the vast majority of your audience sits behind systems that pre-load pixels and distort open counts by default.

How Apple MPP rewired the open

5. Apple MPP went from 19% of opens to over 72% in fourteen months

Rejoiner's client data tracked Mail Privacy Protection exploding from 19% of total opens in October 2021 to over 72% by December 2022. Few changes in email history have propagated this fast - it took barely a year for the majority of opens to become privacy-proxied, and the open rate has not meant the same thing since.

6. MPP opens made up over 70% of one large 2022 dataset's total opens

In Rejoiner's 2022 numbers, Apple MPP opens represented over 70% of all opens recorded across the dataset. When seven in ten opens are generated by a proxy pre-fetching images, the metric is no longer measuring recipient behavior - it is largely measuring Apple's infrastructure.

7. Machine opens hit 5% of all opens in MPP's first week

Twilio observed that within the first seven days of MPP, machine opens already accounted for 5% of all opens - and that share only grew as adoption spread. Even the earliest data made clear this was a structural shift in what an "open" counts, not a temporary blip.

8. Apple Mail drove 49.29% of opens in January 2025 - about half of them machine-driven

By early 2025, Apple Mail alone accounted for 49.29% of email opens, with roughly half of those being machine-generated rather than human. That single figure captures the core problem: a huge slice of your reported opens are Apple's servers loading a pixel, not a person reading your email.

See how your email really performs, not just opens →

Open rates, inflated

9. Reported open rates rose about 18 points, past 40%, within six months of MPP

HubSpot documented reported open rates climbing roughly 18 points, to over 40%, in the six months after MPP launched. Nothing about recipients changed - the same people opened at the same rate. The dashboards simply started counting Apple's pre-fetch as engagement.

10. One newsletter's open rate leapt from ~28% to 55% overnight

A single newsletter saw its open rate jump from around 28% to 55% once MPP took effect - nearly doubling with no change in content or audience. Anyone who celebrated a sudden open-rate surge in this period was, in most cases, measuring Apple rather than measuring interest.

11. Campaigns at ~28% opens now report ~52% - with identical clicks and conversions

Geysera describes campaigns that tracked about 28% opens before MPP and now show roughly 52% - while click rates and conversions stayed exactly the same. The clean tell is the flat click line under a soaring open line: the extra opens produced zero extra action.

12. Litmus projected a ~75% inflated Apple Mail open rate at peak adoption

Early on, Litmus modelled that once MPP was fully adopted, Apple Mail's apparent open rate could reach around 75% (a 2021 projection). That is not a healthy program - it is a ceiling created by proxy pre-fetching, and treating it as a benchmark would badly mislead any sender comparing against it.

Bots and scanners in your numbers

13. Up to 30-40% of recorded opens can be bots, not people

Beyond MPP, GetResponse reports that up to 30-40% of recorded opens can come from bots - security scanners, link checkers, and automated systems - rather than real recipients. Between proxies and bots, a large minority of your "engagement" was never a human at all.

14. Filtering bots erased 37% of opens and 46% of click-throughs from one cohort

When GetResponse filtered out bot activity, one cohort lost 37% of its opens and 46% of its click-throughs. Note that clicks were hit even harder than opens - which punctures the comforting assumption that clicks are automatically bot-proof. They are cleaner than opens, not spotless.

15. AI-tool bot clicks peaked above 3 million in a single day

Iterable found automated clicks from AI tools surging to over 3 million clicks in a single day by mid-March 2025. As AI assistants increasingly read and act on inboxes for people, the line between human and machine engagement gets blurrier - and rises further up the funnel into clicks.

16. Bots account for roughly 51% of all internet traffic

The email problem is a slice of a bigger one: bots now make up roughly 51% of all internet traffic as of 2025. More than half of what moves online is automated, so any metric that counts raw hits - opens included - inherits that noise by default.

Check whether your engagement is reaching real inboxes →

What broke in automation and segmentation

17. With half of opens machine-made, open-based "active" segments are heavily polluted

Because roughly half of opens are now machine-generated, any segment built on "opened in the last 30 days" is heavily polluted with people who never actually read. You end up treating proxy pre-fetches as interest, keeping genuinely disengaged contacts in your "active" pool and mailing them right up until they complain.

18. Only 24% of senders run a sunset policy, and just 13% test inbox placement

Mailgun found that only 24% of senders operate a sunset policy and just 13% run inbox-placement testing. In a world where open data is unreliable, that leaves the large majority of senders with no trustworthy way to identify dead weight on their list or to confirm mail is actually landing.

19. Open-triggered automations still report a 38% open rate versus 30.7% for campaigns

Omnisend's 2025 data shows automated, open-triggered emails reporting a 38% open rate against 30.7% for one-off campaigns. The gap looks like automation working brilliantly - but when the trigger and the success metric are both "opens," MPP inflates both ends of the loop and the reported lift is partly an artifact.

The shift to clicks and conversions

20. Only 15% of marketers still treat open rate as their primary metric

The market has adjusted: just 15% of email marketers still name open rate as their primary success metric. The overwhelming majority have moved their definition of success downstream, to signals that survive proxies and bots.

21. The real benchmarks now: 2.3% average CTR and 5.3% average click-to-open

HubSpot's current engagement benchmarks put the average click-through rate at 2.3% and click-to-open rate at 5.3%. These are the numbers worth optimizing against - click-through captures action, and click-to-open, imperfect as its denominator is, at least ties back to a deliberate step rather than an automatic pixel load.

22. Email click-to-conversion rates grew 27.6% in 2024, reinforcing conversion-based measurement

Omnisend reports email click-to-conversion rates grew 27.6% in 2024. Conversion is the metric bots and pre-fetchers distort least, because it requires a real outcome - a purchase, a signup, revenue. As opens lose meaning, conversion becomes the anchor that keeps email reporting honest.

What this means: measure what survives

Stop treating the open rate as attention. Apple drives 64.66% of opens, MPP crossed 72% of opens back in 2022, and roughly half of Apple Mail opens are machine-generated. An open no longer implies a human, so a rising open rate is as likely to signal wider Apple adoption as genuine interest. Report opens as a rough directional trend at best, never as your headline number.

Assume a large minority of engagement is non-human. Up to 30-40% of opens can be bots, filtering bots cut one cohort's clicks by 46%, and bots are now over half of all internet traffic. This is not paranoia - it is the base rate. Build your reporting expecting inflation, and be especially skeptical of sudden spikes that clicks and conversions do not corroborate.

Rebuild segmentation and sunsetting on clicks, not opens. Open-based "active" segments are heavily polluted with machine opens, yet only 24% of senders run a sunset policy. Re-anchor your engagement tiers on clicks and site or purchase activity, and sunset contacts who have not clicked or converted in a defined window. Otherwise you keep mailing people the open data falsely marks as alive.

Make clicks and conversions your success metrics. Only 15% of marketers still lead with open rate, benchmark engagement now sits around 2.3% CTR and 5.3% click-to-open, and click-to-conversion grew 27.6% in 2024. These metrics survive proxies and bots because they require a deliberate action. Optimize subject lines and content against clicks and revenue-per-email, and let opens fall to a supporting role.

Verify what open data can no longer tell you: real inbox placement. When opens are unreliable, you lose your informal early warning for deliverability trouble - and only 13% of senders test inbox placement. The fix is to measure placement directly rather than infer it from a polluted metric. Test where your email actually lands across providers, so a soft engagement signal never masks mail quietly slipping into spam.

FAQ

Questions, answered.

Is the email open rate still a reliable metric in 2026?+
No. Apple Mail Privacy Protection pre-fetches images for a huge share of recipients, so every MPP user registers as an open whether or not they read the message. With Apple driving 64.66% of all opens and roughly half of Apple Mail opens being machine-generated, reported open rates are inflated by 15-25 points. Only 15% of marketers still treat open rate as their primary success metric - and for good reason.
What is Apple Mail Privacy Protection (MPP)?+
MPP is a feature Apple launched with iOS 15 in late 2021 that routes email through Apple's proxy servers and pre-loads all remote content, including the tracking pixel senders use to count opens. The effect is that Apple fires an "open" for essentially every MPP-enabled recipient the moment mail arrives, and it hides the recipient's IP address and location. Adoption climbed from 19% of total opens in October 2021 to over 72% by December 2022.
How much do bots inflate email open rates?+
Between MPP pre-fetching and security scanners, up to 30-40% of recorded opens can be bots rather than real recipients. When one cohort filtered bots out, it lost 37% of its opens and 46% of its click-throughs. Bots now account for roughly 51% of all internet traffic, and AI-tool bot clicks peaked above 3 million in a single day in March 2025.
Why did my email open rate suddenly jump?+
If your open rate climbed sharply with no change to your content or list, MPP is the usual cause. Reported open rates rose about 18 points, past 40%, within six months of MPP launching. One newsletter jumped from roughly 28% to 55% overnight, and campaigns tracking ~28% opens now show ~52% - with identical click rates and conversions. The extra opens are machine-generated, not new human attention.
What metrics should I track instead of open rate?+
Move to click and conversion metrics, which bots and pre-fetchers distort far less. Benchmark engagement now runs around a 2.3% click-through rate and a 5.3% click-to-open rate, and email click-to-conversion rates grew 27.6% in 2024. Pair click-based engagement with revenue-per-email and a sunset policy for genuinely inactive contacts, rather than an open-based "active" segment that is heavily polluted by machine opens.
Does Apple MPP affect email deliverability?+
Indirectly, and it can hurt. Because opens are no longer a trustworthy engagement signal, senders who prune inactive contacts using open data will keep mailing recipients who never actually read - inflating volume to unengaged addresses that mailbox providers penalize. Only 24% of senders run a sunset policy and just 13% test inbox placement, so most are flying blind. Click-based engagement and real inbox-placement testing are the reliable replacements.