News
tech

An Email Open Is an Image Load, Not a Read

An email open gets logged when a tracking image loads, which is why Apple Mail fetching your message in the background counts as an open while a resident who reads it four times counts once. Here is what the number measures and which signals to move your decisions onto.

Open rate is the share of delivered emails where a tracking image loaded. That is the whole definition. It is not the share of people who read the message, and on most lists it is no longer close.

The mechanism is small. Your email tool puts a one pixel transparent image in every message, with a unique address for each recipient. When a mail client fetches that image, the tool logs an open. No fetch, no open. A resident who reads your renewal notice with images turned off counts as never having opened it. A mail server that loads the image with no human present counts as a reader.

Why the number climbed while the replies did not

Apple shipped Mail Privacy Protection with iOS 15 in 2021, and it is on by default for people who turn the protection on. Apple's own description says the feature downloads remote content in the background through two relays, where the first relay knows the recipient's IP address but not the content and the second knows the content but not the address. Apple's support guide states the consequence for senders plainly: senders cannot tell whether you opened the email.

Two effects run at once. Images load for messages nobody looked at, so your open count now includes mail that was deleted from a notification shade. And repeat views are served from Apple's cache, so a resident who reads your newsletter four times registers once at most. The number is inflated at one end and deflated at the other, which is why no fixed discount repairs it.

Email platforms have adjusted. Twilio SendGrid's documentation marks suspected background fetches with a machine open flag and tells senders to treat clicks as the stronger signal of intent. If your platform carries a similar flag, that field is the one to pull. The headline open rate on the dashboard is the one to stop quoting.

Three decisions this breaks

Subject line tests. A subject line test is scored on opens by design, which worked when an open meant a person. A share of each variant's opens are now background fetches that happened before any human saw the subject line. The interference is not even across your two splits either, because it follows the mix of mail clients that landed in each one. On a list of a few hundred addresses you cannot read a winner out of that.

Resending to non-openers. The usual move is to mail everyone who did not open. That segment now holds people who read the message in Apple Mail and decided against it. You end up sending a duplicate to engaged residents, and sending twice in one week to the prospects most likely to be annoyed by it.

Lead scoring and automation triggers. When an open advances a prospect or fires a follow-up, the trigger runs on a server fetch. A prospect who signed somewhere else three weeks ago keeps scoring as warm, because their phone keeps fetching images in the background. Every report built on that score inherits the error.

What to measure instead

Choose signals that required a person to decide something.

  • Clicks on a named link. A click is a request a human made. The number will be far lower than the open rate you are used to, so rebuild your benchmarks around it rather than reading it against the old figure.
  • Replies. Count them by hand if your platform will not. On a leasing enquiry or a renewal offer, the reply is the outcome you wanted.
  • The action the email asked for. Showings booked, applications started, renewal forms returned, payment details updated. Tie each one to the send date and you have a number that connects to rent.
  • Unsubscribes and spam complaints per send. Both are deliberate acts, which makes them trustworthy, and both tell you when a send went wrong.

The number mailbox providers keep about you

Open rate lives inside your own email tool. Gmail never sees it. Your spam complaint rate is the reverse: Gmail measures it, you do not, and it decides whether your arrears reminders reach an inbox. Google's sender guidelines tell every sender to keep the spam rate reported in Postmaster Tools below 0.30 per cent, and the same page sets a tighter goal of staying under 0.10 per cent and never reaching 0.30 per cent at all.

If you mail more than a few hundred addresses, verify your sending domain in Postmaster Tools and read the spam rate graph once a month. It takes a few minutes, and it governs delivery in a way your open rate never has.

What to change this week

  1. Open your email platform's reporting and find out whether it separates machine opens from human ones. Note the field name so your whole team reads the same column.
  2. List every automation that triggers on an open. Move each one to a click, a reply, or no trigger at all.
  3. Score subject line tests on clicks, or stop testing wording and test the offer instead.
  4. Turn your non-opener resend rule into a non-clicker rule, and cap how often one address can receive a duplicate.
  5. Add your sending domain to Google Postmaster Tools and write down today's spam rate as your baseline.

Keep the metric, demote it

Opens are not worthless. A list whose open count falls to near zero overnight usually has a delivery fault, and that deserves an alert. One campaign sitting far below the rest of your sends is a reason to check the from address and the authentication records. Treat the number as a smoke detector rather than a scoreboard, and put your decisions on signals a person had to create.


Back to all articles