Gmail and Apple MPP: How to Track Newsletter Ads Accurately in 2026

Two infrastructure decisions — one from Apple, one from Google — have made open-rate based newsletter ad tracking unreliable. Apple Mail Privacy Protection pre-fetches every image in every email before the subscriber reads it. Gmail caches all images on Google's servers, masking subscriber IP addresses and device signals. Together, these systems mean that naive impression tracking based on image-fetch events will both overcounting and misattribute a significant fraction of your newsletter ad impressions. This guide explains exactly what each system does, why it breaks standard tracking, and how MailAdx's architecture handles both accurately.

Apple Mail Privacy Protection: What It Does and What It Breaks

Apple introduced Mail Privacy Protection in iOS 15 and macOS Monterey in September 2021. When a user enables MPP (it is offered as an opt-in during setup and now on by default for many Apple Mail users), Apple's servers fetch all remote images in incoming emails — including tracking pixels and ad serving images — before the email is delivered to the subscriber's device.

This prefetch happens on Apple's infrastructure, not on the subscriber's device. From the ad server's perspective, the image request comes from an Apple data center IP address, using an Apple User-Agent string, immediately upon email delivery — not when the subscriber actually opens the email.

The consequences for ad tracking are severe:

Open events fire for every subscriber regardless of whether they opened. Every MPP-enabled Apple Mail subscriber's tracking pixel is fetched the moment the email arrives. A newsletter sent to 10,000 subscribers with 40% MPP penetration will see approximately 4,000 "opens" fire immediately upon delivery — before any human has opened the email. This inflates apparent open rates dramatically.

Timing data becomes meaningless. With MPP, opens appear to happen in bulk clusters immediately after send, regardless of when subscribers actually read the email. Time-of-day open analytics are unreliable for the MPP-affected fraction of your list.

Device and location signals are wrong. The prefetch request comes from Apple's infrastructure. IP-based geolocation shows an Apple data center, not the subscriber's location. Device User-Agent strings reflect Apple's fetch infrastructure, not the subscriber's actual device.

For send-time ad platforms that track impressions via image pixel fetch, MPP means they are logging impressions for subscribers who may never have opened the email. For open-time ad serving platformslike MailAdx, the challenge is more nuanced: the platform must distinguish between an MPP prefetch (not a real human open) and a genuine subscriber open (a real impression).

Gmail Image Proxy: The Older, Quieter Problem

Google began routing email images through its own image proxy in 2013. When a subscriber reads an email in Gmail, images are served from Google's caching servers (googleusercontent.com) rather than from the original image URL. The original server (your ad server) receives only Google's proxy request — not the subscriber's device request.

Unlike MPP, Gmail's proxy does not pre-fetch images. The cache is typically checked and populated on first open, then served from cache on subsequent opens. This means:

The first open fires an image request to the ad server. Subsequent opens from the same email serve from Google's cache and do not reach the ad server. This means ad-server impression counts for Gmail subscribers may undercount actual multiple-open events, but do generally reflect at least one genuine open.

IP-based subscriber identification doesn't work for Gmail. The request comes from Google's proxy IP, not the subscriber's IP. Any tracking that relies on IP addresses to identify or locate subscribers will misattribute Gmail traffic.

The Gmail proxy problem is less severe than MPP because it doesn't create phantom opens for unopened emails. But it does affect IP-based frequency capping and location targeting — both of which need to be handled through subscriber hash methods rather than IP resolution.

Specifically What Breaks in Newsletter Ad Tracking

The combined effect of Apple MPP and Gmail proxy creates these specific problems for newsletter ad platforms that use standard tracking approaches:

Impression overcounting: MPP fires impression events for every Apple Mail subscriber regardless of whether they opened. A newsletter with 30% MPP penetration will show approximately 30% more "impressions" than actually occurred. Advertisers paying per impression in this model are paying for events that include a significant fraction of prefetched, unread emails.

CTR deflation: If impressions are overcounted by 30% but clicks are correctly counted (clicks require a real human action and cannot be prefetched), the calculated CTR will be proportionally lower. An actual CTR of 2.0% in a newsletter with 30% MPP penetration might appear as 1.4% if impressions include all MPP prefetches. Advertisers evaluating campaign performance against CTR benchmarks will draw incorrect conclusions.

Open-rate unreliability: Publishers using open-rate data to report campaign performance to direct sponsors will overstate engaged readership. Sponsors who use open rates to evaluate whether a newsletter campaign was "seen" by subscribers are working with inflated numbers.

Location targeting errors: IP-based location targeting routes Apple prefetch traffic to the data center location (commonly Virginia, California, or Ireland depending on Apple's infrastructure) rather than the subscriber's actual location. A geo-targeted campaign for a US-only advertiser will show delivery in Apple data center locations rather than subscriber locations.

How MailAdx Detects and Handles Both Systems

MailAdx's ad server uses a multi-signal detection approach to distinguish MPP prefetch events from genuine human opens, and to handle Gmail proxy traffic correctly.

MPP detection via User-Agent analysis: Apple's image prefetch infrastructure uses distinctive User-Agent strings that identify the request as coming from Apple's mail privacy proxy rather than a real email client. MailAdx's ad server checks incoming requests against a continuously updated list of these identifiers. Requests identified as MPP prefetches are handled differently than genuine opens.

Timing pattern analysis: Genuine subscriber opens are distributed over time — a cluster immediately after send, then a long tail over hours and days. MPP prefetches cluster in a very tight window immediately after delivery. When the timing pattern of image requests indicates a bulk, near-simultaneous cluster that is inconsistent with human reading behaviour, these events are flagged as probable MPP traffic.

Subscriber hash for identity: Rather than relying on IP addresses for subscriber identification, MailAdx uses the SHA-256 hash of the subscriber's email address (passed via merge tag in the ad URL) as the primary identifier. This works correctly through Gmail's proxy (the hash is in the URL, not derived from the requesting IP) and through Apple's prefetch (though MPP requests are flagged and handled separately). See how subscriber email hashing worksfor the full technical detail.

The result: Impression counts in MailAdx reflect verified human opens rather than a mix of human opens and machine prefetches. Advertisers viewing MailAdx reporting see impressions that represent actual reader attention, not inflated raw counts.

Why Clicks Are Now the Primary Newsletter Ad Metric

In a world where open events are contaminated by MPP prefetches, clicks have become the most reliable signal in newsletter advertising. A click requires a real human action — a subscriber seeing an ad, finding it relevant enough to act on, and clicking through. Neither Apple's prefetch infrastructure nor Gmail's image proxy can simulate a click.

MailAdx counts clicks as the primary engagement metric. CTR (clicks divided by verified impressions) is the performance denominator that advertisers use to evaluate campaign quality and publishers use to substantiate their rate card.

For direct-sold campaigns, publishers should present click data — not open data — in their post-campaign reports. A direct sponsor who received 12,400 clicks on a campaign can verify those clicks in their own analytics. A direct sponsor who received "42,000 opens" cannot verify that number and may correctly suspect it is inflated by MPP. See newsletter ad reportingfor how to present campaign performance to sponsors and advertisers.

The shift to click-primary metrics also affects how publishers talk about their newsletters to potential sponsors. "We average 2.2% CTR on header placements" is more credible and more useful than "we average 44% open rate" in a post-MPP world. Sophisticated advertisers already know this.

What Advertisers Need to Know

Advertisers running newsletter campaigns through MailAdx should understand how impression counting works to set appropriate performance expectations:

MailAdx impressions are verified opens, not sends or raw image fetches. When your campaign reports 15,000 impressions, that is the platform's best estimate of 15,000 genuine subscriber opens — not 15,000 emails delivered or 15,000 pixel fires including MPP prefetches. This is a higher quality impression than most newsletter ad platforms report.

CTR is calculated against verified impressions. A 2.1% CTR in MailAdx represents clicks divided by verified human opens, not clicks divided by inflated MPP-inclusive open counts. Comparing MailAdx CTR to send-time platforms or MPP-unhandled platforms is an apples-to-oranges comparison — MailAdx CTRs will typically appear lower but represent more accurate engagement against a cleaner denominator.

Conversion attribution works through click tracking. MailAdx provides click-through URLs that advertisers can use for conversion attribution in their own analytics. A subscriber who clicks an ad and converts on the advertiser's site generates a trackable event that cannot be confused with an MPP prefetch.

How Widespread MPP Adoption Actually Is

Estimates of MPP prevalence vary depending on the newsletter's audience profile. In general:

Apple Mail accounts for approximately 55–60% of email opens globally, though this varies significantly by region and demographic. In the United States, Apple Mail is dominant among iOS users — who represent over 55% of smartphone users. MPP is enabled by default for many users who accepted Apple's iOS 15 update without changing privacy settings.

B2B newsletters targeting enterprise professionals in the United States typically see 35–50% MPP penetration. Consumer newsletters targeting broad demographics see 40–60%. Newsletters with large international audiences, particularly in markets with lower iPhone penetration, see lower MPP rates.

The practical implication: in almost all newsletter contexts, a significant fraction of your subscriber list is affected by MPP. Platforms that don't handle it are producing unreliable impression data for the majority of their publishers' audiences.

For a complete picture of newsletter ad performance metrics in context, see newsletter CPM benchmarks and performance data for 2026.

Frequently Asked Questions

Does Apple MPP affect all subscribers who use Apple Mail?

MPP affects subscribers who use Apple Mail (on iPhone, iPad, or Mac) and have either opted into MPP or accepted the default privacy settings in iOS 15 and later. Subscribers who use Apple devices but read email through the Gmail app, Outlook app, or other non-Apple clients are not affected by MPP. The proportion of your Apple Mail subscribers with MPP enabled varies, but most publishers assume 70–85% of Apple Mail subscribers have it active.

Does MailAdx's open-time serving still work with MPP?

Yes, with modifications. When MailAdx detects that an image request is an MPP prefetch (via User-Agent and timing signals), it does not serve a paid ad to that request — the MPP fetch receives a 1×1 transparent pixel response. The actual ad is served on the genuine open event when the subscriber reads the email. This means MailAdx only charges advertisers for real human opens, not for MPP machine fetches.

How do I explain MPP to my direct sponsors?

The simplest explanation: "Open rates are inflated by an Apple infrastructure change that pre-fetches email images. We use click-based metrics, which are unaffected by this system. Our click data is fully verified." Most sophisticated advertisers already understand MPP. For those who don't, presenting click data prominently in your post-campaign reports and de-emphasising open rates is the right approach.

Does Gmail's proxy affect frequency capping?

It would if frequency caps were implemented via IP-based identification. MailAdx implements frequency capping via subscriber email hash, not IP address — so Gmail's proxy doesn't affect cap accuracy. A subscriber who has been served 2 impressions this week is correctly capped regardless of whether those impressions were requested through Gmail's proxy or directly. See: audience targeting with email hashes.