Apple Mail Privacy Protection and open rates: what to trust
Apple's Mail Privacy Protection preloads tracking pixels, so opens include machine activity. What it changes and which metrics still carry signal in B2B.
55 to 60%
On this page
Apple's Mail Privacy Protection, live since September 2021, makes the Mail app download an email's images in the background, tracking pixels included, so an open can be logged when nobody read the message. It applies to people who switch it on. Judge outbound by replies, clicks, unsubscribes and complaints, and use opens only as a rough health check.
What Mail Privacy Protection does
Your sending tool
Sends the email. The tracking pixel is one of its images.
The Mail app
Downloads the images in the background, whether or not anyone opens the message. Litmus says this can happen straight away or a couple of days later.
Apple's two relays
Carry the request. The first knows the device's IP address but not the content; the second knows the content but not the IP address.
Your sending tool
Logs an open, from a proxy address in the reader's general region. Nobody has read the email yet.
The person
Reads it later, or never. Repeat opens inside Apple Mail are not passed back to the sender.
Reading plays no part in whether the pixel fires, so Postmark notes you could see a 100% open rate for Apple Mail recipients, and a click-to-open ratio that drops as a result.
It is a setting. Apple's page names the switch: Protect Mail Activity, under Privacy Protection in the Mail settings. The app decides it: Litmus says it applies to anything opened in the Apple Mail app, Gmail and work accounts included, and not to other apps such as the Gmail app on an iPhone.
How much of your open data is machine activity
Opens by mail client, September 2026 Apple 76.56%
Apple 76.56%, Gmail 16.10%, Outlook 3.48%, Yahoo Mail 2.35%, every other client together 1.51%
Opens Mail Privacy Protection affects Roughly 55 to 60%
Litmus's estimate, which it credits to unnamed studies
Share of all opens
Nobody outside Apple can say how many opens on your list are pixel fetches. The nearest public number is Litmus's client data, which puts the share Mail Privacy Protection affects at roughly 55 to 60% of all opens. That is Litmus's customers, not your list, and we found no published split for B2B senders.
Even on a protected device, a read and a prefetch both arrive as opens. Twilio SendGrid can flag every open from such a recipient as machine activity for that reason, reads included. Any fixed split of machine and human opens is a guess.
In B2B, a prospect on Microsoft 365 who reads in the Apple Mail app on an iPhone is covered; one in Outlook is not. If your sending tool reports the client behind each open, count the Apple share. Corporate mail adds machine activity of its own: HubSpot filters security scanners and corporate screeners such as Mimecast from opens and clicks.
Which metrics still carry signal
| Signal | Why it holds up | Caveat |
|---|---|---|
| Replies | A person has to write one. Belkins stopped tracking opens and reports replies divided by emails sent. | Count auto-replies separately. |
| Clicks | Mailchimp says clicks and purchases are not affected by Mail Privacy Protection. | Security scanners click too. |
| Unsubscribes, bounces, complaints | Postmark names them as more reliable feedback. | Bounces show list quality more than interest. |
| Spam rate | Google says to keep it below 0.10% in Postmaster Tools and never reach 0.30%. | On a small send, one complaint moves it a lot. |
| Meetings and pipeline | Nothing fires them without a person. | Slow, so read them per account. |
How to spot bot clicks
Mailchimp says security filters and privacy tools trigger bot clicks, and lists the signs: dozens of clicks within minutes of sending, clicks on every link, bursts from one IP address within seconds, and email clicks with no matching visits in web analytics. Count a click as a person when a matching visit shows up in your site analytics.
What to change in your reporting
- Move opens to a health check. Litmus says a dramatic drop in open rates can still reveal a deliverability problem. Put replies and meetings at the top.
- Stop triggering on opens. Mailchimp says open-triggered automations go to a larger audience than intended and open-based A/B tests may be inaccurate. In outbound, trigger follow-ups on time and on replies.
- Drop open-based send times, location and device. Litmus says send-time optimization is inaccurate unless the tool excludes these opens, and Mailchimp says location and device cannot be read from them.
- Consider switching open tracking off for cold sends. lemlist advises it and Belkins says pixels were hurting deliverability. Neither publishes its test, so try it on one domain and compare.
Our rule for an engaged contact in B2B: someone who replied, or who clicked and produced a matching site visit. An open alone never qualifies.
If you are buying outbound, ask the agency how it treats machine opens before you read its open rates. If you track clicks instead, the UTM builder tags the links.
Sources
- apple.comApple's Mail Privacy Protection
- litmus.comApple's Mail Privacy Protection, live since September 2021, makes the Mail app download an email's images in the background, tracking pixels included, so an open can be logged when nobody read the message
- postmarkapp.comReading plays no part in whether the pixel fires, so Postmark notes you could see a 100% open rate for Apple Mail recipients, and a click-to-open ratio that drops as a result
- litmus.comThe nearest public number is Litmus's client data, which puts the share Mail Privacy Protection affects at roughly 55 to 60% of all opens
- twilio.comcan flag every open
- knowledge.hubspot.comCorporate mail adds machine activity of its own: HubSpot filters security scanners and corporate screeners such as Mimecast from opens and clicks
- belkins.ioBelkins stopped tracking opens and reports replies divided by emails sent
- mailchimp.comMailchimp says clicks and purchases are not affected by Mail Privacy Protection
- support.google.comGoogle says to keep it below 0.10% in Postmaster Tools and never reach 0.30%
- mailchimp.comMailchimp says security filters and privacy tools trigger bot clicks, and lists the signs: dozens of clicks within minutes of sending, clicks on every link, bursts from one IP address within seconds, and email clicks with no matching visits
- litmus.comLitmus says a dramatic drop in open rates can still reveal a deliverability problem
- lemlist.comlemlist advises it and Belkins says pixels were hurting deliverability