Email reports still lead with open rate. It is the easy number. It is also the one that moved first when Apple changed how mail loads. This piece covers what the feature does, what it does to your report, and what to put in its place. For the wider picture, see customer retention strategy for ecommerce.
What does Mail Privacy Protection do?
Apple Mail Privacy Protection (MPP) downloads remote email content in the background by default, whether or not the recipient engages with the email, and routes it through two separate relays so the sender does not get the recipient's IP address2. Apple's support guide says it hides the IP address and prevents senders from seeing whether the email was opened1.
Remote content includes the tiny tracking image most email tools use to record an open. If that image is fetched whether or not a person reads the message, an "open" is recorded either way. Apple does not use the words "open rate"; joining the two is our reading of how open tracking is usually built.
Apple Mail Privacy Protection hides whether a message was opened.
What happens to open rates?
They go up, and they stop meaning much. Klaviyo, an email platform, says in its help centre that MPP prefetches its tracking pixel and that open rates will be inflated3. The same page, written to troubleshoot low flow open rates, still calls open rate a key deliverability metric and tells senders to watch other deliverability metrics too3. That is a vendor describing its own product, so read it as one view.
The number does not only move up. In one thread on the same platform's community forum, a sender's open rates fell overnight on Gmail and Hotmail while other providers were unaffected4. A drop like that points to inbox placement, not to Apple. That is the real mess: some opens are machine-made, some real readers land in spam, and the open rate cannot tell you which is which.
Email deliverability is whether your email reaches the inbox rather than spam. Opens used to be a rough read on it. Now they are a noisy one. Use them as a flag to investigate, not as a result.
Email automation flows are sequences sent when a customer does something: welcome, abandoned cart, post-purchase. They run for months without anyone rereading the report, so a flow report built on opens deserves extra suspicion.
Which metrics still work, and what are their limits?
| Metric | What it tells you | Limit |
|---|---|---|
| Click rate | Someone interacted with the message | Shows interest, not a purchase |
| Placed orders | Orders credited to the email within the tool's window | Counts orders that might have happened anyway |
| Revenue per recipient | Sales credited to a message, divided by people sent to | Revenue, not margin; the same credit problem as orders |
| Margin per recipient | Contribution margin on credited orders, per person sent to | Needs your own cost data; our preferred measure |
| Unsubscribes and spam complaints | Whether the audience wants the programme | Slow signals; they say nothing about sales |
Vendor benchmark pages now report placed order rate and revenue per recipient for campaigns and flows, next to open rate5. We do not quote their numbers. A benchmark describes whoever supplied the data, not your list.
Report margin, not only revenue. Revenue per recipient is easy to pull. Contribution margin, in our convention, is revenue minus the variable costs of an order before marketing. A discount-heavy flow can show healthy revenue per recipient and a thin margin.
What do attribution windows in email tools do to the numbers?
An attribution window is the period after an event, such as an open or a click, during which an order is credited to the message. Tools let you set it. The setting changes the result.
The logic runs as follows. If a tool credits an order to any open or click inside its window, then opens that MPP generates can claim credit for orders the customer was always going to place. The longer the window, the more orders it can claim.
Example (illustrative): A customer buys on Tuesday. An email arrived Monday evening and was fetched in the background. The customer never saw it. A tool with a window of 5 days from open credits Tuesday's order to the email.
So ask three questions of any email report:
- Which events count: opens, clicks, or both?
- What is the window, written out in full, for example "5-day click"?
- Would the order have happened without the email?
The third question is the only one that matters, and no setting answers it. For more on reading reports like this, see how to read an agency report.
How do you test email with a holdout group?
Hold back a random share of recipients from a campaign or flow. They get nothing. Compare the orders and margin per person between the two groups. A holdout test like this measures incrementality: the orders the email caused, over and above those that would have happened anyway.
Example (illustrative): 10,000 customers are eligible. 9,000 receive the flow and 1,000 are held out. The sent group places 450 orders, or 5.0%. The held-out group places 30, or 3.0%. Applying the held-out rate to the sent group predicts 270. The email added about 180 orders. The tool may have credited it with more.
Practical points:
- Pick the held-out group at random, not by engagement. Picking "inactive" people compares different customers.
- Keep the group out for the whole run, not for a week.
- Compare per person and in margin, not only in orders.
- Who gets emails depends on consent. Your lawyer decides that, not the test design.
For ZFD: We report email on clicks, orders and margin per recipient, and we propose a holdout when the number will decide a budget. Our view of what an email report should contain, not an industry standard.
Sources
- Use Mail Privacy Protection on iPhone - Apple SupportPrimary source
- Mail Privacy Protection & Privacy - Apple LegalPrimary source
- Troubleshooting low flow open rates - Klaviyo Help CenterVendor
- Open rates dropped from 40% to 5% overnight on Gmail and Hotmail domains - Klaviyo CommunityForum
- Ecommerce email marketing benchmarks - BranvasVendor