Apple and Gmail Privacy Updates: What's Changed for Email Marketers

For most of email marketing's history, one number ran the show: the open rate. It decided which subject lines "won," which segments were "engaged," and whether a campaign was a success. Then Apple broke it, Google rewrote the rules of who even reaches the inbox, and AI quietly took over deciding who sees your email once it lands. If your email marketing playbook still runs on opens and volume, you're playing a game that ended a couple of years ago — and the scoreboard has been lying to you ever since.

This is a plain-English rundown of what Apple and Gmail actually changed, why it matters, and what to do about it now. The individual updates arrived piecemeal, but together they point in one clear direction, and adapting to it is one of the highest-return moves an email programme can make.

What Apple changed

Apple's shifts have been about privacy first, and their side effects have quietly reshaped how marketers work.

The headline consequence is simple: the open rate is no longer a reliable signal, and your carefully written preheader may never be seen as written. Apple's AI also renders summaries far better from live text than from image-heavy emails — a design consideration we'll come back to.

What Google changed

Google's changes have been about enforcement and control. Since February 2024, high-volume senders to Gmail (and Yahoo) have had to meet firm requirements or pay the price in the inbox.

• Authentication is mandatory. SPF, DKIM, and DMARC must be set up and aligned. Unauthenticated bulk mail is increasingly rejected outright, not merely filtered.
• One-click unsubscribe. Marketing emails must include a working List-Unsubscribe header and honour opt-outs within about two days.
• Spam complaints capped. Keep your complaint rate below 0.3% — Google recommends staying under 0.1% — or watch deliverability fall off a cliff.
• Manage subscriptions. Gmail now gives users a central place to review and drop subscriptions, sorted by sending volume — so your highest-frequency senders are the most exposed to being culled.
• Gemini summaries. Like Apple, Gmail uses AI to summarise incoming mail, often from the first hundred or two hundred characters of the body.

And this is no longer just Google's game: Microsoft has rolled out its own bulk-sender requirements for Outlook, so the authentication bar is becoming universal across every major inbox.

The one-line summary Apple made your metrics lie, Google made the rules mandatory, and AI decided your visibility. Volume and opens are out; trust and engagement are in.

Shift one: the open rate is dead

Start here, because everything downstream depends on it. With MPP inflating around half of opens, an "open" no longer means a human read your email — it might just mean Apple's servers loaded a pixel. Optimising subject lines by open rate is now measuring noise, and any automation built on opens — "resend to non-openers," open-based lead scoring, engagement segments defined by opens — is quietly making decisions on fiction.

The fix is to move your source of truth to signals you can still trust: clicks, replies, conversions, and on-site behaviour. That's the same reckoning with unreliable measurement we cover in why attribution is getting harder, and it's exactly the vanity-versus-real-metric distinction from rethinking success metrics. For the fuller picture across channels, modelling approaches like marketing mix modelling help fill the gap the open rate left behind.

Shift two: authenticate, or you don't arrive

This one isn't optional and isn't subtle. If you send at any volume and haven't set up SPF, DKIM, and DMARC properly, a growing share of your mail simply won't be delivered. Getting the records right — and moving DMARC toward an enforcement policy rather than leaving it in monitoring mode — is now table stakes, alongside a genuine one-click unsubscribe and disciplined attention to your spam-complaint rate. Most modern email platforms will help you configure this; the mistake is assuming it's handled when it isn't. Check your setup in a postmaster tool before you assume you're safe. There's an upside to getting it right, too: once DMARC is at enforcement, you become eligible for BIMI, which can show your verified logo — and, with the right certificate, a checkmark — beside your sender name in supporting inboxes, a small but growing trust signal that lifts recognition.

Shift three: engagement is the new deliverability

Here's the deeper change hiding behind the headlines. Deliverability used to be binary — inbox or spam. Now there's a whole spectrum of visibility inside the inbox, and AI decides where you land on it based on how your recipients behave over time. If most of your list ignores you, the inbox learns to bury you, and even perfectly authenticated mail goes unseen. Engagement is no longer just a marketing metric; it's a deliverability input.

That flips the old volume mindset on its head. Sending more to everyone now actively hurts you. The winning move is to send less, to people who actually want it: align frequency to engagement, reduce or pause sends to dormant subscribers, and run genuine re-engagement before sunsetting the silent ones. Gmail's volume-sorted subscription manager only sharpens the point — over-mailing puts you at the top of the list people scroll to unsubscribe from. A smaller, engaged list beats a big, indifferent one on nearly every measure, which is why permission and quality collection, covered in building an email list from scratch, matters more than raw size. Ultimately, the most reliable way to stay engaging — and therefore deliverable — is to send genuinely useful, well-made email content people are glad to receive, rather than filler that pads the calendar. Relevance is the deliverability strategy.

Shift four: write for the AI reader, too

You now have a second audience for every email: the AI that summarises it. When Apple Intelligence or Gemini condenses your message from its opening characters, a first line of "We hope this email finds you well" produces a summary that says precisely nothing — or worse, something misleading. Front-load the actual point in your first sentence, and favour live, selectable text over emails built entirely from images, because the AI reads text far better than pictures. It's a small change in craft with an outsized effect on whether your message survives the summary intact — and it dovetails with the broader move toward AI-shaped communication we explore in how AI is changing email personalisation.

The data shift: privacy-first by default

Underneath all of this is a simple reality: you can track less than you used to. Pixels are unreliable, link tracking is increasingly protected, and hide-my-email addresses break the old identity links. The durable answer is to rely on data people give you willingly — zero-party data from preference centres and surveys, and first-party behaviour on your own site. Ask subscribers what they want and how often, then honour it. That both improves relevance and sidesteps the tracking that privacy features are designed to block, and it makes every click that does land on a well-built landing page count for more.

Your 2026 email playbook

Pulling it together into what to actually do.

Two more habits round it out: apply the same "shift from volume to value" discipline to SMS, given Apple's unknown-sender filter is heading the same way as email tabs; and keep an eye on downstream conversion, since the whole point of protecting deliverability is to get more of the right people onto a page that converts — the end-to-end view a funnel audit gives you, and the same reason product experiences like a strong product page matter as much as the email itself.

The bottom line

None of this means email is in decline — it remains one of the highest-return channels in marketing, still delivering tens of dollars back for every dollar spent. What's changed is that the easy version stopped working. Apple made your favourite metric unreliable, Google (and now Microsoft) made authentication and permission non-negotiable, and AI turned the inbox into a place where visibility is earned through genuine engagement rather than bought with volume. Stop optimising for opens, authenticate properly, send less to people who want it more, write for the AI as well as the human, and build on data people freely give you. Do that, and the privacy era isn't a threat to your email programme — it's the thing that forces it to get good.