Skip to content
Updated 11 min readByBob Thordarson

Email attribution is broken: why your revenue numbers are probably wrong

Your ESP says email generated $X this month. That number is almost certainly too high. Last-click attribution gives email 100% credit for purchases that involved multiple touchpoints. This post covers why your attribution data is less reliable than it looks, how incrementality testing reveals your real email impact, and what privacy regulation means for the tracking signals attribution depends on.

Email attribution 2026: last-click gives the final touch 100% of credit; GDPR fines reached €1.2B in the year to January 2025

CMOs estimate that 45% of the data their teams use for decisions is incomplete, inaccurate, or outdated. Here's where email attribution goes wrong.


Key takeaways

  • 35% of US marketing and finance leaders say more than 20% of their marketing budget is inefficiently allocated (Haus, 2026)
  • CMOs estimate that 45% of their teams' decision data is incomplete, inaccurate, or outdated (Adverity, 2025)
  • Last-click attribution gives email 100% credit for multi-touch journeys
  • Incrementality testing shows how much of the revenue your ESP reports would have happened anyway
  • France's CNIL now requires consent for open-tracking pixels used to measure campaigns; GDPR fines totalled about €1.2 billion in 2025 (DLA Piper)

Every email marketer has a revenue number they report upward. "Email generated $X this month." The number comes from their Email Service Provider (ESP); it looks precise, and it's almost certainly too high.

The problem isn't that email doesn't work. Email works. The problem is that last-click attribution gives email 100% credit for purchases that involved multiple touchpoints. A customer sees an Instagram ad, reads a blog post, gets retargeted, receives an abandoned cart email, and buys. Email gets all the credit. The other three touches get nothing.

In Haus's January 2026 survey of 500 US marketing and finance leaders, 35% said more than 20% of their marketing budget is inefficiently allocated, and one in three reported conflicting data sources. And expanding privacy regulation is making the tracking signals that attribution depends on less reliable every quarter.

This post is part of our 2026 Email Marketing Benchmarks series.


The scale of the problem

MetricData point
Leaders who say more than 20% of budget is inefficiently allocated35%
CMOs' estimate of decision data that is incomplete, inaccurate, or outdated45%
Senders who don't measure promotional email ROI or are unsure54%

In Sinch Mailgun's 2026 survey of 1,234 email senders, only 46% said they actively measure promotional email ROI. The 200 CMOs in Adverity's 2025 survey estimated, on average, that 45% of the data their teams use for decisions is incomplete, inaccurate, or outdated. That's not a measurement optimization problem. That's a credibility problem.


How last-click attribution warps email metrics

A realistic purchase path

  1. Sees Instagram ad, visits site, leaves
  2. Reads blog post from Google search, browses products, leaves
  3. Returns via retargeting ad, adds to cart, abandons
  4. Receives abandoned cart email, clicks, purchases

Last-click attribution: email gets 100% credit.

What actually happened: four touchpoints contributed. The Instagram ad created awareness. The blog built interest. The retargeting ad drove the cart addition. The email was the final nudge — important, but not the whole story.

What last-click overvalues

ChannelWhy
Branded searchCaptures existing demand, doesn't create it
Retargeting adsWarm audience, already aware and interested
Promotional emailsFinal nudge gets full credit for a multi-step journey
Coupon/deal sitesIntercepts purchases already in progress

What last-click undervalues

ChannelWhy
Prospecting campaignsCreates the awareness that other channels capture
Video advertisingBuilds consideration, rarely the last click
Influencer partnershipsDrives discovery, conversion happens elsewhere
Organic social/contentEducates and builds trust over time
Podcast advertisingHigh influence, hard to measure via clicks

The uncomfortable question

"If I hadn't sent this email, would the purchase still have happened?"

For an abandoned cart email sent 30 minutes after abandonment to a customer already in your checkout flow, the honest answer is often yes. At least partially.

Alex Greifeld (author of the "No Best Practices" newsletter, ecommerce operator since 2011) makes this point regularly: some SaaS and martech providers "abuse attribution to make their products look more successful" than they are. When your ESP reports a flow "generated" $50,000, how much was truly incremental?

Many email agencies report inflated ROI by picking whichever attribution model makes email look best. Measure incrementality even when the numbers are smaller. Honest data leads to better decisions.


Incrementality testing: the closest thing to truth

How it works

Hold back a random percentage of qualified recipients from receiving the email. Compare their purchase rate to the group that got it. The difference is your real incremental impact.

GroupTreatmentMeasurement
Test group (90%)Receives the emailTrack purchases
Holdout group (10%)Does not receive the emailTrack purchases
Incremental impactTest conversion minus holdout conversionTrue email effect

If the test group converts at 8% and the holdout at 5%, the email's incremental impact is 3 percentage points. Not the full 8% that last-click would claim.

What you get from it

Incrementality testing shows you which emails drive truly incremental revenue (spend more there), which emails claim credit for purchases that would have happened anyway (reallocate that spend), and where additional sends stop producing incremental value.

How to run it

  1. Start with your highest-volume flow (usually abandoned cart).
  2. Hold out 5-10% of qualified recipients randomly.
  3. Measure purchase rates for both groups over 7-14 days.
  4. Calculate lift — that's your true incremental contribution.
  5. Rotate holdout groups so you're not permanently excluding anyone.
  6. Test quarterly. Incrementality changes as your audience and flows evolve.

Other attribution approaches

Multi-touch attribution (MTA)

ModelHow it worksUse case
LinearEqual credit to every touchpointSimple, fair baseline
Time-decayMore credit to recent touchesFavors conversion-proximate channels
Position-based40% first, 40% last, 20% middleValues discovery and conversion
Algorithmic/data-drivenML-assigned weights based on patternsMost accurate, most complex

Marketing mix modeling (MMM)

MMM uses aggregated, anonymized data to estimate channel contribution — no user-level tracking required. This makes it privacy-safe and unaffected by the tracking signal loss from iOS App Tracking Transparency, cookie restrictions, Apple MPP, and state privacy laws.

For brands that need honest measurement in a privacy-first environment, MMM is increasingly the answer.

Profit attribution

A framework that cuts through vanity ROAS:

CampaignROASActual profit marginVerdict
Campaign A5x0%Scales loss
Campaign B3.5x41%Actually profitable

Last-click and ROAS-based optimization would scale Campaign A. Profit attribution correctly identifies Campaign B as the one worth scaling. This distinction matters more than most teams realize.


Privacy regulation and the end of individual tracking

The regulatory landscape

Europe:

  • France's CNIL requires consent for open-tracking pixels used to measure and optimize campaigns, under a recommendation adopted in March 2026

  • The CNIL exempts only pixels used for login security, or used solely to spot inactive recipients for list cleaning on transactional or consented emails

  • GDPR fines: about €1.2 billion in 2025, broadly matching 2024 (DLA Piper)

  • France's CNIL fined Google €325 million on 1 September 2025 over ads shown between Gmail emails and cookies set without valid consent

  • Violations: up to 20 million euros or 4% of global revenue

US state privacy laws:

  • At least ten states, including California, Colorado, Connecticut, Texas and New Jersey, require businesses to honour browser opt-out signals such as Global Privacy Control

  • Indiana, Kentucky and Rhode Island privacy laws took effect on 1 January 2026

  • California CCPA/CPRA: behavioral profiling is regulated activity

Technical signal loss:

What email tracking can still collect (and what's regulated)

Data typeStatus
Opening timestampsNeed consent in France for campaign measurement (CNIL 2026)
IP addresses / locationMasked by proxies, regulated
Device types and OSIncreasingly anonymized
Screen resolutionRegulated under privacy laws
Open counts/patternsUnreliable (MPP) and regulated
Click dataBest remaining signal (but bot-compromised)
Purchase data (first-party)Most reliable, consent-required

Dark pattern prohibitions

Regulators are going after manipulative consent practices. Countdown timers creating false urgency invalidate consent. Pre-checked opt-in boxes are prohibited. Easier-accept-than-reject paths are prohibited. Obfuscated opt-out options violate regulations. Accept and reject must have equal prominence.

Where this is heading

Individual-level email tracking is ending. What comes next is less granular but more defensible: cohort-level aggregate measurement, first-party data (declared preferences, purchase history), direct response signals (purchases, replies, forwards), marketing mix modeling, and incrementality testing. You'll know less about each individual subscriber and more about whether your program is actually working.


What to change in your email program

Stop doing

  • Trusting platform-reported revenue at face value
  • Using last-click attribution as your sole performance measure
  • Citing "email generated $X revenue" without an incrementality context
  • Optimizing for ROAS without considering profit

Start doing

  • Running incrementality holdout tests on major flows quarterly
  • Implementing multi-touch or MMM for channel budget allocation
  • Tracking profit attribution alongside revenue attribution
  • Building first-party data collection into every email touchpoint
  • Reporting email's incremental contribution honestly, even if the numbers are smaller

The numbers will be smaller. They'll also be real, which matters more than most people think — especially when someone in finance eventually asks where the revenue figures come from.

Related: Revenue per email: the metric that should have replaced open rates | The email metrics that actually matter


Frequently asked questions

How much marketing budget is wasted on poor measurement?

In Haus's January 2026 survey of 500 US marketing and finance leaders, 35% said more than 20% of their marketing budget is inefficiently allocated. Fragmented data and reliance on last-click models are the main drivers. In Adverity's 2025 survey, CMOs estimated that 45% of the data their teams use for decisions is incomplete, inaccurate, or outdated.

What is incrementality testing for email?

A holdout experiment where 5-10% of qualified recipients don't get the email. The purchase rate gap between the email group and the holdout group is your true incremental impact. Start with abandoned cart.

Why does last-click attribution overvalue email?

It gives 100% credit to the final touchpoint. Email — especially abandoned cart and promotional emails — is often the last nudge in a journey that started elsewhere. The awareness and consideration built by earlier touches get zero credit.

How are privacy laws affecting email tracking?

France's CNIL requires consent for pixels that measure campaign open rates. At least ten US states require businesses to honour Global Privacy Control, and about half of iOS users decline app tracking (AppsFlyer, 2025). Individual-level tracking is being replaced by aggregate and first-party measurement.

What is marketing mix modeling?

MMM uses aggregated data (not individual tracking) to estimate each channel's contribution. It's privacy-safe, unaffected by cookie restrictions and ATT opt-outs, and gives a more honest view of email's role without relying on click-based attribution.


Related reading:

Sources


Part of the 2026 Email Marketing Benchmarks series by Geysera

Bob Thordarson

Co-Founder and CEO

Bob Thordarson is CEO and Co-Founder of Geysera, a serial entrepreneur with 25+ years and five co-founded ventures, including Cequint (acquired by TNS in 2010 for $112.5M) and Consumerware (acquired by ParkerVision). A graduate of the University of Washington and MIT Entrepreneurial Masters Program, based in Seattle, he serves on the boards of DRY Soda Co. and the Entrepreneurs' Organization Seattle chapter. He is an expert in retention marketing email systems and methodology for ecommerce and B2B brands — measured by incremental revenue, not vanity metrics.