MPP (Apple Mail Privacy). Open rate has been a vanity metric since 2021.

01What MPP actually does

Open tracking has always worked the same way: the sender embeds a transparent 1×1 pixel image at a unique URL in the email body. When the recipient's mail client loads the image, the request hits the sender's server, and the sender logs an open.

This system worked because most mail clients waited for the human to actually open the email before fetching the image. The pixel load was a reasonable proxy for "human looked at this." Not perfect — auto-preview panes always created some noise — but reasonable.

Apple Mail Privacy Protection (MPP), released with iOS 15 in September 2021, broke the proxy. When a user enables MPP (Apple prompts them on first launch with a friendly default that nudges yes), Apple's mail servers fetch every tracking pixel in every incoming email, in advance, before the user has opened anything — and from Apple-owned IP addresses, not the user's device.

The technical mechanism

From the sender's perspective, what was previously visible:

• Whether the pixel was loaded (= open detected)
• The user's approximate location (from their IP)
• The user's device and mail client (from the user-agent string)
• The exact moment of opening

What MPP reveals instead:

• The pixel was loaded by an Apple server, some time after the email was delivered, regardless of whether the user actually opened it
• The Apple server's location (which by design rotates and obscures the user's real location)
• A generic Apple Mail user-agent
• A timestamp that has nothing to do with the user

In other words: MPP doesn't just block tracking. It replaces tracking with a confident lie. Every Apple-Mail recipient appears to have opened the email, opened it quickly, and opened it from a generic location. The open rate goes up. The geolocation data goes wrong. The "best time to send" analysis becomes nonsense. None of this is broken — it's working exactly as Apple designed it.

02Timeline — five years of pretending

The story of MPP in B2B sales tech is mostly a story of denial. Here's how it unfolded:

03The inflation math

Here's the simplest way to feel the size of the problem. Suppose you send 1,000 cold emails to a typical B2B audience.

This isn't a worst case — it's the median case for a B2B audience in 2026. Some audiences are even more skewed (founder / exec audiences over-index heavily on Apple Mail). The inflation is not a fixed offset you can subtract; it depends on audience composition, and you usually can't measure that composition reliably either, because the same MPP that breaks opens also masks the user-agent that would tell you who's using Apple Mail.

04What survived MPP

MPP only breaks one thing: the pixel-based open. Everything else still works, because everything else requires actual user behavior — and Apple's relay servers don't reply, don't click, don't book meetings, and don't unsubscribe. Here's the post-MPP reliability map:

*Click rate has a small MPP-related caveat: some Apple privacy features and some corporate proxies pre-fetch links for security scanning, generating phantom clicks. The rate is much smaller than the open-rate inflation — usually 1–4 percentage points — and is consistent enough that comparisons across campaigns still work.

The pattern: any metric that requires a real human action (reply, click, meeting, unsubscribe) survived. Any metric that requires a passive signal (pixel load, geolocation, open time) is dead or unreliable. This is the framework. Anchor your dashboards on the survivors.

05The four metrics to track instead

If you're removing open rate from the weekly review, here's what should replace it. These four metrics, together, give you a complete operational picture of cold-email performance without depending on anything Apple can fake.

Notice what's missing: open rate. Notice also what's also missing: vanity metrics like "total sends" or "sequences active" — counts of activity rather than outcomes. The four-metric dashboard is small on purpose. Each one is hard to fake and each one drives a different operational decision.

06Warm-up tools and engagement scoring

Email warm-up tools (Mailwarm, Warmup Inbox, Lemwarm, Instantly's warm-up feature, Smartlead's auto-warmup) work by exchanging emails between a network of pooled mailboxes, then opening and replying to those emails to build positive sender reputation. MPP creates a subtle complication here: many warm-up tools historically used "opened the warm-up email" as a signal of healthy sender reputation. Apple opens all of them automatically, so the signal degrades.

The better warm-up tools (Smartlead, Instantly) have adapted by emphasizing reply-based interactions and folder placement (inbox vs spam) over open signals. The cheaper warm-up tools (some white-labeled clones) still report opens as a primary success metric, and their dashboards now show 95%+ "open rates" on warm-up traffic that's mostly Apple pre-fetches. If your warm-up tool's reporting feels too good to be true, MPP is usually why.

Engagement scoring in CRMs

The bigger story is the lead-scoring systems that bake "email opens" into the score. HubSpot, Marketo, Pardot, and most lead-scoring tools historically gave point values for "opened email × N times," and many marketing teams haven't recalibrated. The result: MQL pipelines packed with leads whose only qualifying behavior is Apple's pre-fetch.

If your demand-gen team is celebrating an MQL-volume increase that started in late 2021 and never stopped, the increase is partly real and partly MPP. Audit the lead-scoring rules, drop the open-based scoring, and re-baseline. The MQL number will look worse and the conversion-to-pipeline rate will look better — both are improvements.

07Where sequencers are still misleading you

Most modern sequencers have some MPP-aware reporting, but the implementations vary wildly. Here's the current (2026) state of play:

The pattern is unsurprising: tools built for cold outbound (Smartlead, Instantly) adapted faster than tools built for sales workflow (Outreach, Salesloft, HubSpot), because cold-outbound users are closer to the deliverability layer and felt the inflation more sharply. If your team is on a sales-workflow sequencer, expect to do more work to translate the dashboards into reality — and don't take the open-rate column at face value, regardless of who built it.

08The "filter out Apple" workaround

A common attempted fix: filter all opens from Apple-Mail user-agents out of the reporting, treat the remaining ~40% as "real," and report on that subset. Some sequencers offer this as a setting; some teams build it themselves with SQL on the exported event log.

It's better than nothing. It's still wrong. Here's why:

• Apple's user-agent isn't always identifiable. When MPP pre-fetches, it uses Apple-owned IPs and Apple-tagged user-agents most of the time — but not always. Forwarded mail, mail accessed through non-Apple clients on Apple devices, and edge cases all create slippage.
• The non-Apple subset isn't representative. If you filter out Apple, you're now optimizing for the Outlook/Android/Gmail-web subset of your audience, which may have systematically different behavior than the Apple users you're ignoring. Subject lines that work for the non-Apple sample might underperform on the audience you're not measuring.
• The "real" number is still a pixel load. Even after filtering Apple out, you're still measuring "image was fetched by a client," not "human read the email." Auto-preview panes in Outlook, security scanners at corporate firewalls, and various other middleware all create non-human pixel loads. The number is less wrong, not right.

If you must report opens for political reasons (a VP demands them), then the Apple-filtered open rate is the honest version. But the deeper move is to stop building the political need for opens — replace them in the weekly review, replace them in the comp plan, replace them in the QBR — so the demand goes away.

09The 7-step migration playbook

If your team is still organized around open rate, here's how to migrate. This is the operational sequence — about 4–6 weeks of effort for a typical mid-market outbound team — that we've seen work cleanly:

1. Replace the weekly review's top-line metric. Move reply rate to the headline slot. Move meeting rate immediately under it. Open rate goes into a "legacy / for reference" section, or off the dashboard entirely. This is the single highest-leverage move — once the weekly meeting changes, behavior changes with it.
2. Audit the comp plan. Find every place opens are referenced — SDR commission triggers, AE pipeline bonuses, performance reviews, manager scorecards. Replace each one with reply / meeting / pipeline-sourced. This is harder politically than technically; expect 2 weeks of negotiation.
3. Recalibrate the lead-scoring rules. In HubSpot / Marketo / Pardot / whatever scoring engine, remove or zero-weight any rule that gives points for "opened email × N." Replace with reply-based scoring (any reply = +10, positive reply = +25, meeting booked = +50). Re-baseline MQL volume and acknowledge the number will look different.
4. Update sequence A/B test logic. If your A/B tests declared winners based on open rate, they were testing nothing useful for the last five years. Re-run the relevant tests with reply rate as the success metric — many "proven" subject lines won't actually be better, and some "underperforming" ones were lifting reply rate the whole time.
5. Switch deliverability monitoring on. Tools like GlockApps, Mailtrap inbox-placement, or built-in features in modern sequencers (Smartlead, Instantly) test inbox vs spam placement weekly. This becomes the new "open rate" — a daily-checkable health metric for the funnel.
6. Retrain the team on the new dashboard. A 30-minute team meeting walking through the four metrics, what each measures, and what to do when each one moves. Include the historical context (why opens are dead) so people understand the change rather than feel imposed-upon by it.
7. Run a 30-day baseline. After the new dashboard is live and the comp plan is updated, take 30 days of clean data as the new performance floor. Everything from this point forward gets measured against this baseline, not against the inflated open-rate numbers from before.

The teams that complete this migration generally report that nothing else about their work changes much in week one — the same people send the same emails to the same lists. What changes over months is decision-making: subject lines, sequence length, follow-up cadence, ICP targeting. All of these benefit from being optimized against signals Apple can't fake.

10Common mistakes

11Related terms

MPP sits at the intersection of email deliverability, privacy regulation, and outbound measurement. To go deeper: