The short answer
Since Apple's Mail Privacy Protection (introduced in 2021 with iOS 15), the open rate is no longer a reliable measure of success. Apple Mail loads an email's content on receipt, whatever the recipient then does with it [1][2].
Instead, measure what requires a real action: clicks, replies and, in the end, enquiries, meetings or purchases. Add the warning signals bounce rate, unsubscribes and spam complaints. For cold email campaigns, a 3–6 % reply rate counts as a good result [6].
Why has the open rate been unreliable since iOS 15?
Opens are traditionally measured with an invisible image, a so-called tracking pixel. If the inbox loads this image, the newsletter tool counts an open. This is exactly where Apple comes in.
Apple announced Mail Privacy Protection in June 2021. The feature is meant to stop senders from learning when someone opens an email, and it hides the IP address [1]. Technically, Apple Mail loads remote content immediately in the background on receipt by default. According to Apple, this happens regardless of what you do with the email [2]. The feature applies to Mail on iPhone, iPad, Mac and Apple Vision Pro as well as Mail on iCloud.com [2][3].
To your newsletter tool this looks like an open, even if the email is deleted unread. The more recipients use Apple Mail with the protection switched on, the more the open rate is inflated. You can see how high this share is for you if your tool breaks down opens by email client.
What is the open rate still good for?
As an individual value for one campaign, hardly anything. As a rough trend across many sends, it can be a signal if your list stays similar. A sudden drop then points to a delivery problem rather than to declining interest. In that case, check deliverability first.
You should no longer use the open rate for three things:
- A/B tests of subject lines: if part of the "opens" happens automatically, chance may decide the winner. Decide by clicks.
- Triggers in automations: "has opened" as a condition sends Apple users follow-up emails that don't fit.
- Inactivity rules: anyone who only sorts by opens keeps dead addresses that never read.
Which metrics count instead?
Calculate rates based on delivered emails where possible, not sent ones. Otherwise bounces distort all other values.
| Metric | Calculation | What it tells you |
|---|---|---|
| Bounce rate | undeliverable ÷ sent emails | List quality, technical problems |
| Click rate | recipients with a click ÷ delivered emails | Whether content and button convince |
| Reply rate | replies ÷ delivered emails | Real interest, especially in sales |
| Unsubscribe rate | unsubscribes ÷ delivered emails | Whether frequency and content fit |
| Spam complaints | spam reports ÷ delivered emails | Risk for delivery, at Gmail the target is below 0.1 % [5] |
| Conversion rate | conversions ÷ delivered emails | Whether the channel brings business |
The click-to-open rate (clicks divided by opens) was popular for a long time. Because the denominator is distorted, today it's just as unreliable as the open rate itself.
A note on the spam rate: Google calculates it itself in Postmaster Tools and expects it to stay permanently below 0.1 %; from 0.3 % it gets critical [5]. Your newsletter tool only sees the complaints providers report back. That's why a look at Postmaster Tools pays off.
Are clicks always genuine?
Not quite. Security filters in companies check links before a person clicks them. Microsoft describes this for Safe Links in Defender for Office 365: URLs are checked before delivery, and links without a trusted reputation are opened and tested in the background [4]. Other filters work similarly.
Typical patterns of an automatic click are clicks a few seconds after delivery or clicks on all links in an email at the same time. Check whether your tool offers a bot filter and whether it's active. If not, look at the individual values when click rates are suspiciously high, before you celebrate.
How do you measure enquiries and revenue from emails?
Clicks are an intermediate step. Whether email pays off is only shown by what happens next. For that you need two things.
First, UTM parameters on every link. Google Analytics always requires utm_source, utm_medium and utm_campaign [7]. Example (fictional):
https://www.example.com/offer?utm_source=newsletter&utm_medium=email&utm_campaign=2026-10-maintenance
Define a fixed scheme, for example always "email" as the medium and year-month-topic as the campaign. Otherwise the evaluation falls apart into spelling variants after a few months.
Second, a connection to the CRM. In B2B the deal rarely closes on the website. Record in the CRM which campaign or flow a contact came from. Then after three months you can say how many meetings and orders came from email. How visitors on the website turn into enquiries is described in the article on conversion funnels.
Which metric for which type of email?
- Newsletter: click rate, unsubscribe rate and replies. Long term: enquiries per 1,000 delivered emails.
- Automation flows: conversion of the flow, i.e. how many contacts reach the goal, such as a meeting. More on this under automation & sequences.
- Sales emails and sequences: reply rate and the meetings that come from it. Apollo names 3–6 % as a good result for broad B2B campaigns; below 3 % points to problems with the target group or deliverability [6]. What applies legally is under cold email in B2B.
- Transactional emails such as order confirmations or invoices: delivery rate and bounces. Here it's only about them arriving.
How do you test subject lines without the open rate?
A/B tests remain useful; you just need a different success criterion. Split the list randomly into two groups of equal size, change exactly one thing and decide by clicks or replies instead of opens.
- One variable per test: subject or sender name or button text, never everything at once. Otherwise you won't know afterwards what worked.
- Enough recipients: with small lists, a single click decides the winner. As a rule of thumb from our project experience, a test with fewer than a few hundred recipients per group is more of a hint than proof. Then collect results across several sends.
- Wait long enough: some recipients only read on the next working day. Anyone who automatically sends the winner after two hours mainly measures the early risers.
- Note the result: a simple table with date, variant, click rate and insight. After six months you have your own rules instead of gut feeling.
Typical mistakes in reporting
- The open rate as the headline figure in the report to management.
- Comparing click rates without a bot filter with previous-year values measured without a filter.
- Campaigns without UTM parameters that show up as "direct" in Google Analytics.
- Treating unsubscribes as failure. Someone who unsubscribes might otherwise have clicked "Spam".
What does honest reporting look like?
Example of a monthly report that fits on one screen:
- Result: enquiries, meetings and revenue from email, with the previous month.
- Engagement: click rate and replies per send, without the open rate as the headline figure.
- Health: bounce rate, unsubscribe rate, spam rate from Postmaster Tools.
- List: new confirmed sign-ups minus unsubscribes and removed addresses.
- One insight, one measure: what you'll do differently next month and why.
If a report shows the open rate at the top and enquiries not at all, it's measuring technology, not success.
Conclusion: from opening to acting
Saying goodbye to the open rate isn't a fashion. The basis of measurement has changed technically. Anyone who puts clicks, replies and conversions at the centre and keeps an eye on the warning signals measures what counts: not whether an email arrived, but whether it achieved something.