Why Email Attribution Breaks — And How Surveys Fix It

Email marketing is the channel that always seems to win. Open any ecommerce dashboard and email sits near the top of the revenue column — often claiming 20%, 30%, or more of attributed sales. ESPs display these numbers prominently. Marketing teams cite them in budget reviews.

But beneath that confident reporting sit three compounding measurement problems: a broken metric that inflates open-rate data for nearly two-thirds of all sends, a tracking fragility that silently disconnects clicks from conversions, and a last-click convention that assigns revenue based on proximity rather than actual influence.

Here's what the data actually shows — and why post-purchase surveys are the signal that reconciles everything else.

Open Rates Have Been Unreliable Since September 2021

When Apple launched Mail Privacy Protection (MPP) with iOS 15 in September 2021, it changed what the word "open" means in email. Under MPP, Apple's servers pre-fetch email content — including tracking pixels — when a message arrives in the inbox. The open registers whether or not the recipient ever looks at the message.

The scale matters. Apple's ecosystem — iPhone Mail, iPad Mail, macOS Mail, and all opens classified under MPP — now accounts for approximately 64.66% of all email opens, according to Litmus, calculated across more than one billion tracked opens in May 2026 (source). Litmus estimates 55–60% of all tracked opens are now affected by MPP pre-fetching.

When your ESP celebrates a 42% open rate, an unknowable slice of that figure is phantom opens — emails Apple loaded on behalf of subscribers who may have never glanced at the subject line.

The downstream effects are concrete. Re-engagement campaigns trigger for subscribers who aren't actually engaging. Suppression lists fail to suppress the right people. The "most engaged" segments used to justify increased sending frequency may be largely artifacts of pre-fetching rather than real interest.

Open rates have become a vanity metric for most email programs. Treating them as behavioral signals — and making frequency, segmentation, or content decisions based on them — is one of the most expensive mistakes in email marketing today.

Click Tracking Has Its Own Fractures

Clicks, at least, require a human to act. But click data has its own vulnerabilities.

iOS 17 introduced Link Tracking Protection, which strips certain tracking parameters — including gclid, fbclid, and platform-specific click IDs — from links opened in Mail and Messages (source). For brands that reconcile email click data with ad platform performance, this creates gaps at the critical last handoff between channels.

There's also mobile in-app browsing. When a subscriber taps an email link and lands in an in-app browser rather than their default browser, cookies set on your store don't persist into later sessions in Safari or Chrome. A customer who clicks a promotional email Thursday, browses but doesn't buy, then returns via direct navigation Friday — that conversion records as direct, not email.

At the device level, only about 35% of iOS users opt into tracking under Apple's App Tracking Transparency (ATT) framework as of Q2 2025, according to Adjust (source). The other 65% are invisible to device-level attribution models. Their multi-session path from email click to purchase exists in a blind spot no pixel can illuminate.

The Last-Click Trap: Email as Accidental Hero and Forgotten Nurturer

Attribution models that give 100% credit to the last tracked touchpoint before conversion were always a simplification. For email, this creates two distinct and opposite distortions.

Distortion 1: Email gets credit it didn't earn.

Picture a brand launching a 15%-off flash sale via email on a Friday. By Monday, their ESP attributes 35% of weekend revenue to email. What it doesn't capture: many of those buyers had already visited the site multiple times after a paid social ad, had products in their cart, and were simply waiting for a reason to act. The email was a trigger — not the persuasion that built purchase intent.

When teams see "email drove 35% of revenue," they read it as email being a powerful awareness and persuasion channel. The real signal is narrower: email is an efficient conversion nudge for a pre-warmed audience. Those are different things with very different strategic consequences.

Distortion 2: Email gets no credit when it earned some.

Picture a brand with a 10-email welcome series that educates new subscribers about their product category over 30 days. A subscriber opens six of those emails, learns the vocabulary of the space, and then buys after clicking a Google Shopping ad on day 26. Under last-click attribution, search gets 100% of that conversion. Email gets nothing.

For nurture-heavy brands, this systematically understates email's contribution to acquisition. Over time, it produces under-investment in list growth and email content quality — because the numbers make email look like it doesn't drive new customers.

Both distortions are real. They don't average out — they compound in ways that vary by brand, promotion cadence, and channel mix.

What Post-Purchase Surveys Actually Reveal

The solution isn't a more sophisticated attribution model. It's asking customers what happened.

A post-purchase survey on the order confirmation page — when attention is high and the purchase is fresh — can ask two questions that no pixel can answer:

"How did you first hear about us?" — Discovery attribution. Which channel introduced the brand. For new customers, this is where your awareness investment either shows up or goes missing.
"What made you decide to buy today?" — Decision trigger. The final nudge. Email legitimately appears here — a promotional offer arriving at the right moment, a recommendation that finally landed.

The gap between these two responses is where the real strategic insight lives.

When 25% of new buyers cite email as their purchase trigger but fewer than 3% say it's how they discovered the brand, you have direct evidence that email is a conversion engine, not an acquisition channel. Budget decisions should reflect the difference.

Picture a DTC skincare brand whose ESP attributes 28% of revenue to email. Post-purchase surveys of first-time buyers tell a different story: 55% discovered the brand through organic social or word of mouth, 30% through paid social, 10% through search, and fewer than 5% through email. Email is efficiently converting people who were already going to buy — but it isn't building the customer base. Without survey data, this distinction is invisible and budget stays misallocated.

For more on structuring the question itself — phrasing, response options, single vs. multi-select — see our guide to running a "how did you hear about us" survey.

Reconciling Three Signals at Once

The goal isn't to find the one "correct" attribution source and ignore the others. Each signal captures something real that the others miss:

ESP-claimed attribution: reflects click behavior within email sessions — real, but last-touch-biased and inflated by phantom MPP opens
Pixel + UTM attribution: reflects cross-session digital journeys — useful directionally, but degraded by MPP, ATT, and in-app browser fragmentation
Survey-stated attribution: captures self-reported channel influence — the channels tracking can't reach, including word-of-mouth, offline exposure, dark social, and brand memory from months ago

When all three point the same direction, you have strong confidence. When they diverge — when your ESP claims email drove 38% of new customer revenue but surveys show only 5% of first-time buyers cite email as their discovery channel — that divergence is the signal worth investigating, not averaging away.

This is the operating logic of multi-signal attribution: surveys don't replace pixel data; they give it context it can't generate on its own.

A Practical Starting Point

You don't need to rebuild your analytics infrastructure to start getting better email attribution signal. Here's the minimum viable approach:

Add two questions to your order confirmation page. One for discovery ("how did you first hear about us?"), one for decision trigger ("what made you decide to buy today?"). More than two questions and completion rates drop sharply.
Run for at least 60 days before drawing conclusions. You need enough volume to separate new buyers from returning customers — email performs very differently for each group.
Compare survey-stated email rates against your ESP's attributed number. A gap above 10 percentage points in either direction is worth investigating. Over-credit is common with promotional programs running 30-day attribution windows. Under-credit is common for brands with strong nurture sequences.
Report discovery and decision trigger as separate metrics. When you present email performance in a budget review, these two numbers are strategically different. Treating them as one is where misallocation starts.

The brands that allocate budget most accurately don't just have strong email programs — they know exactly what role email plays at each stage of the customer journey. A two-question post-purchase survey is the fastest path to that clarity.

Ready to capture the attribution signal your pixels can't reach? Try Rauxdata for free.