Read time: 8 minutes
About this time last year, we all started talking about Apple’s new Mail Privacy Protection, a.k.a. Apple MPP, a.k.a. “that thing Apple is doing to email that’s going to change everything.” There was a lot of speculation about how it would affect email, ranging from laid back to predictions of doomsday. There were frenzied discussions about how to cope (wait and see? Make strategic adjustments? Burn down email and move into a bunker?).
A year later, it’s time to take a look at what actually happened. We saw some of the impacts of these changes in this year’s Benchmarks. Now, let’s dig deeper.
Hold on, what is Apple MPP?
Need a quick refresher? No worries, it’s been a long year — I got you.
- Apple Mail Privacy Protection works by pre-loading email content — including open tracking pixels. (Reminder, email opens are tracked by tiny image pixels embedded in email content. When a user downloads that image, it registers as an “email open.”)
- That means that anyone using Apple MPP will be marked as having “opened” ALL of their emails — whether or not they ever actually looked at any individual message.
- This affects anyone who uses the Apple Mail app to read emails on their iPhone, iPad, or Mac desktop. Since Apple Mail can be set up to work not just with Apple domains, but also with most of the major inbox providers (Gmail, Yahoo, AOL, Hotmail…) this impacts a huge segment of email subscribers.
- Apple MPP was rolled out on September 20, 2021 as a part of Apple’s new iOS 15 operating system. Adoption was relatively slow, but by November 28, more people were using iOS 15 than iOS 14, and by mid-March of this year, 80%+ of Apple users were on the new system — translating to about 40-50% of email users overall.
So what happened to open rates?
Well, first of all, the definition changed. Pre-MPP, we all agreed that open rate generally reflected the rate at which people were interacting with an email. Sure, you’d get the occasional bot open or someone with images blocked, so it wasn’t a perfect data point, but it was a fairly stable, useful number.
Now, however, we’ve got three different open rates to think about. You’ll see them go by different names, but for the sake of this blog I’m going to call them:
- Human opens — what “opens” were before this change rolled out. Meaning, we’re fairly certain a real human person interacted with this email.
- Apple opens — aka “machine” or “robot” opens. Meaning, an algorithm opened this email, and we have NO idea whether a human looked at it or not. (Maybe yes! Maybe no! That’s the “Privacy” part of Mail Privacy Protection.)
- Combined opens — #1 plus #2. Some of these opens are people, some are Apple, and added together you get the sum total of every time your open tracking pixel has been downloaded.
To add to the confusion, different email tools are handling this change in different ways. Some (including our friends at EveryAction and ActionKit) are using the data about when, where, and how the tracking pixel has been downloaded to separate out different types of opens, so you can see two metrics for “human open rate” and “combined open rate.” Others haven’t, so you’re still seeing “open rate” — but now it means “combined open rate.”
Now to the good stuff — what happened??
- Human opens went DOWN. Not because fewer people are opening your emails, but because many of those opens are now hidden behind Apple’s Mail Privacy Protection.
- Apple opens went UP. This type of open didn’t really exist at this time last year. It probably started showing up in your reporting in late September, right after iOS 15 rolled out, and has been increasing ever since.
- Which means, combined opens went UP (for Benchmarks participants, they’d already risen by 17% between September and November 2021 — and they’ve kept rising since). You can think of this number as roughly everyone that was opening your emails before, PLUS everyone on your list that wasn’t opening emails but who is using Apple MPP.
So if you’re in a system that looks primarily at “human opens” your email open rates went down. But if you’re in a system that is still looking just at “combined opens” your email open rates went up — WAY up.
With all these changes, are open rates still useful?
For the most part, no (with one BIG EXCEPTION we’ll get to in a second!).
Even before this change, open rates weren’t a great indicator of how well an email did. Tracking how many people wanted to know more (clicks) or were inspired to act (advocacy actions, survey submissions, or donations) have always been better metrics of success.
But now, open rates are all over the place. There are so many factors other than the quality of your email affecting opens — how many of your subscribers are using Apple MPP? Did they open your email on their iPhone today, or their desktop? How does your email program record and report those opens?
So no, opens are no longer a good way to answer the question “did my email succeed?”
But here’s the big exception: open rates are better than ever at helping identify deliverability problems if you are looking at Apple opens or combined open rates.
Deliverability warning signs often show up in open rates, because poor deliverability means more emails going to spam, and people don’t tend to read the emails in their spam folder very often. (That’s why I’ve been recommending monitoring open rates— particularly open rates by individual recipient domain — since the very beginning!)
That part is still true. But you know who else doesn’t read emails in the spam folder? Apple.
MPP only pre-loads images (including that open tracking pixel!) for messages that land in the inbox. Messages that end up in spam? Left unread. That means the portion of your list signed up for MPP goes from 100% open rate to 0% — meaning any deliverability issues show up loud and clear in your data.
If “open” means something different, do active list definitions need to change too?
Speaking of deliverability — one of the basic requirements to maintain a healthy reputation as a good sender is to make sure you’re only sending emails to recipients that want your emails. That means sending to an “active list” of people who demonstrate interest by engaging with your organization’s emails, whether that’s by donating, filling out advocacy petitions, clicking through links… or opening emails.
Of course, that last one has gotten a bit tricky, since an open has gone from indicating “at least mild interest” to “has a pulse and/or an iPhone.”
So the question is, do you change your list definition to exclude opens? Or keep emailing every subscriber who shows signs of life?
Option 1: Segment out Apple opens
Some organizations have taken the proactive approach: identifying people who are only “opening” via Apple, and taking them out of the active list. Anyone who’s clicking, donating, taking action, or registering a human open (in systems where you can tell the difference) stays in.
Pulling out those “Apple opens only” subscribers usually puts quite a dent in list size. To compensate, I’d suggest two additional steps:
One, go a little deeper into those other indicators. For example, if your previous audience was anyone who opened in the past 3 months, clicked in the past 6 months, or donated in the past 12 months, bump that up to clicked in the past 9 months or donated in the past 18.
Two, keep reaching out to that Apple opens only segment, treating them like a warm reactivation list. After all, some of those people probably are opening and reading your emails — you just can’t tell because Apple is protecting their privacy by obscuring that information. And even if they aren’t, Apple only opens messages for email addresses that are in use — so you can be sure that you’re not emailing any abandoned accounts or spam traps.
Sending these folks an email once a month, or when your organization has something truly urgent to say (like after a natural disaster, war, or Supreme Court decision), gives you an opportunity to stay in touch and potentially reactivate some lapsed subscribers.
Option 2: Wait and see
Other organizations are choosing to hold off on making any changes until they see negative impacts from mailing potentially inactive contacts — those people who aren’t interacting with mail at all other than Apple downloading open tracking pixels on their behalf. (Realistically, some organizations are doing this less as a deliberate strategy and more because they aren’t paying attention, but since you’re reading this blog post I’m sure that doesn’t describe you!)
So far, for organizations we work with, neither approach seems to be better than the other.
The proactive approach of segmenting out people with only an Apple open on file reduces both risk and list size, and doesn’t seem to have much impact on the bottom line. Organizations that have taken this path see very little activity from that “Apple opens only” segment, with 95%+ of actions and revenue coming from the rest of the list. And occasional sends to the Apple openers are only maintaining a small but steady trickle of people back onto the active list by enticing them to click, take action, or donate.
On the flip side, the wait-and-see groups haven’t seen any deliverability problems and are continuing much as they have before, just with much higher open rates.
Of course, this could change over time. Maybe the Option 1-ers will start to see key metrics sag over longer periods of time. Maybe the end-of-year season will be more dramatically impacted. And maybe the Option 2-ers will start to see deliverability problems crop up more often as we get months and even years further from Apple MPP’s deployment, pushing them towards the Option 1 camp. We’ll have to wait and see — and when we have anything to report, we’ll be sure to let you know!
Is this the type of thing YOU geek out about too? We’re hiring! I’ve worked here for 10 years and love getting to work with some of the smartest people in the biz — and working for some of the best causes out there. Check out all our openings here.