In the following, we have compiled the most important questions and answers from our experts about the changes in email marketing due to the new Apple Mail Privacy Protection.
On the sidelines of the iPhone 13 keynote, Apple announced that the new operating system versions will be available for download starting September 20, 2021. On the same day, Optimizely Campaign will also roll out an update to detect and exclude auto-generated opens from reports.
The update on 09/20 will detect automatic opens from Apple Mail and exclude them from reports.
In addition, it will then be possible to have automatic openings exported as a file as part of a Response Data Export (RDE). For this purpose, please contact our customer support.
There will be further updates for Optimizely Campaign in the near future, in which we will provide, among other things, a target group functionality for evaluating and segmenting the mailing list using "Automatic Opens". In addition, we plan to include automatic opens as a metric in Deep Analytics, making campaign performance analysis more transparent.
Do you have any other suggestions? Feel free to get in touch with our customer support team.
Our update will filter out automatic opens from Mail Privacy Protection users, which will decrease the overall open rate.
As more users upgrade to the new OS versions, the percentage of automatic opens that are filtered out will increase.
We are currently reviewing and evaluating these options.
Tracking of real opens by affected users is suppressed by Apple and is not possible for us. Therefore, we neither know when nor whether a user with activated Mail Privacy Protection opens incoming mails himself.
However, we do measure openings when a user opens the mail on the web, for example. These openings remain unaffected by Mail Privacy Protection.
Clicks within mailings can be tracked as usual. This allows us to create synthetic opens (once) so that these opens are also included in campaign reporting.
With the Campaign Update, we will be able to identify opens through the Apple Mail Proxy. These opens will be filtered out and can be exported later in the process, allowing us to segment the affected recipients accordingly.
In principle, it is possible to measure in Deep Analytics how high the proportion of openings is through a specific operating system version or a specific browser. Using the "open" metric and grouping by iOS 15 & macOS Monterey, you can approximate the share of Apple Mail users.
With a later update we want to make "automatic openings" selectable as a metric, which can then also be used to determine a good proportion of recipients with Mail Privacy Protection.
iOS penetration in Germany: 25.1% (Statista, 2021)
MacOS penetration in Germany: 18.8% (Statista, 2021)
As of now: No, not clearly. This is because it is not possible to precisely identify Apple Mail usage. However, an approximate analysis via iOS 15 usage is conceivable. In Deep Analytics, the operating system version is selectable as a grouping. In combination with the grouping "recipient list field" and the metric "absolute opens", opens could be assigned on recipient list field level and then matched with the distribution list.
In the future, we will be able to solve this challenge via target groups using the planned "Automatic Opens" condition.
Due to the fact that when Mail Privacy Protection is enabled, non-openers can no longer be cleanly distinguished from openers, we will have to classify all Apple Mail opens as non-openers for the time being. By filtering automatic opens, opens in Apple Mail are determined based on clicks only. This means that an Apple Mail user is not considered a synthetic open in reporting until they click within the mail.
Target groups and triggers that check for non-opens will therefore (incorrectly) classify more recipients as non-opens than before and are no longer reliable. You should therefore adjust them.
In response to the changes in Apple Mail and Optimizely Campaign, you should first review your campaign triggers and audience configurations as soon as possible. Filter conditions that specifically relate to "non-opens" of certain mailings and campaigns should be reconsidered and adjusted. Since openings from Apple Mail users will largely be undetectable in the future, there is otherwise a risk of unwanted email sends to unrecognized openers.
Also, rethink your communication approaches and consider specific customer notices regarding the changing environment in your email communications.
To avoid mixing historical performance data with future data, it may be advisable to swap transactional emails in automated campaigns with a copy of each. This way, you retain old KPIs and benchmarks. This makes it easier to reconcile them with the results of future mailings.
Subject line and A/B testing will still be possible. With slight limitations, random test open rates will also still be comparable.
- Select larger segments (based on the percentage of Apple Mail users in your own mailing list).
- Perform A/B test without detected Apple Mail users
- Use adjusted open rates as a decision criterion
- Use clicks or conversions instead of opens as decision criteria
In A/B testing, can I influence that Apple Mail users are distributed equally among recipients on a pro-rata basis?
As soon as automatic opens are available as a metric for defining target groups, such a setup will basically be possible in Smart Campaigns.
In order to compare future campaigns with historical ones, models must be developed that create comparability with previous campaigns. Here, an adjusted click-top-open rate can help to calculate a synthetic open rate, which can form a comparative value to the previous open rate. Initial empirical values show that the proportion of Apple Mail users in the company's own mailing list must be included as a factor in this calculation.
Separately, the focus on deeper performance metrics such as clicks, page visits, logins and conversions is becoming more important to evaluate the success of sent campaigns. This sometimes requires deeper integration of other channels into your own reporting.
Cross-channel activity scoring models can help to evaluate the activity of a user independently of openings and provide a more comprehensive overall picture of the quality of the distribution.
The application of GDPR-compliant and transparent sign-up processes should remain the top priority in order to maintain a high level of distribution list quality.
In the future, the general activity of recipients will be more important than looking at the open rate alone. It does make sense to focus on the analysis of clicks as an activity characteristic. However, activities of other channels should also be taken into account (website, app, customer portal, etc.). Reactivation measures could then be linked to an "activity score," for example.
To get more clicks in the future and thus continue to recognize active Apple Mail users, interaction-promoting mailings with NPS and similar short surveys can prove successful.
In addition, recognized Apple Mail users could be informed with transparent message texts that your app setting may lead to a reduction in the quality of the newsletters you receive.
Since automatic opens are filtered, Dispatch Time Optimization will only be able to detect activity from affected recipients at the time they click or open in another program.
This means that in the future, dispatch time optimization will not use the opening as the decisive criterion for affected recipients, but will instead optimize for click behavior.
Countdown timers and other dynamic content are loaded at the opening time of the proxy. There is also no way to technically circumvent this. Recipients with activated Mail Privacy Protection will therefore receive the same and therefore outdated countdown timer.
Product recommendations are no longer loaded by Apple Mail users at the actual opening time. Therefore, the back-in-stock must be taken into account in the trigger setup.
The relevance of recommended products may decrease if there is user activity on the page between mail sending and actual opening.
The tracked impressions of your recommendations might be less meaningful due to automatic opens.
So far, the analysis of Apple users of iOS and macOS was only possible based on the openers. In the future, the detected automatic opens will allow for a fairly precise analysis of Apple Mail users with Mail Privacy Protection turned on. Since in principle each of these users generates automatic opens, these opens can be used to calculate a percentage of the total distribution list.
Yes, once automatic opens are available to you as a metric via RDE or in Deep Analytics, this is possible.
The automatic opens will be trackable at the user level, taking into account the respective privacy policies of the senders.