Open rates after Mail Privacy Protection

Open tracking works by embedding an image. The recipient's client requests it, you record an open.

Apple Mail Privacy Protection requests that image on the recipient's behalf — proxied, in advance, whether or not the message is ever read. For every recipient using Apple Mail with the feature on, an open is recorded either way.

Apple Mail is a large share of consumer mail. The open rate is now partly a count of people who use Apple products.

What is still true

Relative movement within one audience still means something, provided the client mix is stable. If your open rate drops from 42% to 28% over a month with no change in audience, something real happened — even if neither number is a count of humans.

Clicks are unaffected. A click requires a deliberate action. Proxies pre-fetch images, not links.

Non-opens still signal. A recipient who has never registered an open across many sends is probably not receiving your mail at all, or has it filtered. Absence remains informative.

What is no longer true

Open rate as a measure of interest. Treating the inflated number as engagement will have you keeping recipients on your list who have not read a message in a year.

Open-based send timing. Pre-fetches happen on the proxy's schedule, not the recipient's. Optimising send time on that data optimises for a cache.

Comparisons across time or audiences. Any change in the proportion of Apple users changes the number independently of behaviour.

What to measure instead

Clicks. The strongest signal available, and the one you probably cared about anyway.

Conversions. Whatever the message asked for — a login, a purchase, a confirmation. Harder to attribute, considerably more meaningful.

Replies, for mail that invites them.

Complaint rate and unsubscribe rate. Both are deliberate recipient actions, both are reliable, and both matter more to your delivery than opens ever did.

Engagement-based suppression

Many senders remove recipients who have not opened in some months. With inflated opens, that filter keeps people it should remove — a recipient whose client pre-fetches every message looks permanently engaged.

Move the criterion to clicks, with a longer window since clicks are rarer. No click in twelve months is a reasonable line for most mail.

Should you keep the pixel

Yes, with caveats. It still provides relative signal and it still identifies total non-openers. Just stop reporting the absolute number as if it meant what it meant in 2020, and make sure whoever reads the dashboard knows why it moved.